“I WANT TO TRY AND TRY”: INCREASING ACHIEVEMENT MOTIVATION IN YOUNG CHILDREN A DISSERTATION SUBMITTED TO THE DEPARTMENT OF PSYCHOLOGY AND THE COMMITTEE ON GRADUATE STUDIES OF STANFORD UNIVERSITY IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY Allison Master May 2011
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
“I WANT TO TRY AND TRY”:
INCREASING ACHIEVEMENT MOTIVATION IN YOUNG CHILDREN
A DISSERTATION
SUBMITTED TO THE DEPARTMENT OF PSYCHOLOGY
AND THE COMMITTEE ON GRADUATE STUDIES
OF STANFORD UNIVERSITY
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
Allison Master
May 2011
http://creativecommons.org/licenses/by-nc/3.0/us/
This dissertation is online at: http://purl.stanford.edu/hx314zv7838
I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.
Carol Dweck, Primary Adviser
I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.
Ellen Markman
I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.
Gregory Walton
Approved for the Stanford University Committee on Graduate Studies.
Patricia J. Gumport, Vice Provost Graduate Education
This signature page was generated electronically upon submission of this dissertation in electronic format. An original signed hard copy of the signature page is on file inUniversity Archives.
iii
iv
ABSTRACT
Motivation to learn plays a critical role in students’ academic success. This
dissertation reports five experiments (N = 250) that increase children’s motivation
(specifically, challenge seeking and persistence) through storybooks. The first two
studies examined how manipulating the similarity between the main character of a story
and the participating child affected preschoolers’ (Study 1) and kindergarteners’ (Study
2) motivation as assessed by the choice of and persistence on challenging puzzles.
Study 2 also compared effects for struggling versus non-struggling students. Study 3
examined whether persuasion would increase challenge seeking, when children
convince someone else of the value of taking on challenges and persisting. Study 4
examined effects over time, and found that children showed a robust boost in challenge
seeking two weeks later, especially those who were reminded of the original book.
Study 5 examined whether the effects could be due to a particular aspect of the books,
process praise, which sends a message that effort and persistence are effective and
valued. Increasing children’s motivation at a young age may set the stage for future
academic achievement, creating a cycle of positive motivation and academic success.
v
ACKNOWLEDGEMENTS
This work would not have been possible without the support and encouragement
of a number of people. First, I would like to thank Carol Dweck for her tremendous
support, wisdom, and mentoring over the past few years. Her work and ideas have
forever changed my life. I also consider myself fortunate to have had the opportunity to
collaborate with and learn from Ellen Markman, Greg Walton, Geoff Cohen, Mark
Lepper, and James Gross. They have all had a profound effect on the way I think about
the world and the psychological processes that shape motivation and well-being.
I would also like to thank the members of the Dweck/Walton and Markman
Labs, and all the other DevoStuds, for all of their friendship, advice, and game nights
through the years, especially Quin Yow, Carissa Romero, and Luke Butler.
The majority of this research was conducted at the Bing Nursery School, so I
would like to thank the children who participated and the teachers who encouraged
them to participate. I am particularly grateful to Chia-wa Yeh, Jennifer Winters, and
Beth Wise for all of their help and valuable feedback. Bing is an amazing place, and I
have had far more fun there over the past few years as “Teacher Game Room” than a
respectable doctoral candidate should. Thanks as well to the staff and students of the
Palo Alto Unified School District.
Many outstanding research assistants helped me collect data for this and other
projects. Special thanks to Emily Campbell, Lauren Hay, Amy Ho, Armando Lopez,
Charishma Chotalia, Isa Dillingham, Carly Janiga, Jessica Snyder, Whitney Worthen,
intervention also incorporates concrete examples of motivation (e.g., a child choosing
a harder puzzle over an easier puzzle), and utilizes storybooks to communicate the
intervention. However, many motivational factors have not been tested systematically
with young children, and a solid understanding of the ways to increase challenge
seeking in preschoolers is still lacking from the literature. Will this intervention be
more effective if the storybooks are relevant to children’s personal identity? Will
having children persuade someone else of the value of challenges increase the
intervention’s effectiveness?
Current Research
The current dissertation has several main aims:
1) The primary aim of Study 1 is to extend research in motivation while
beginning to incorporate findings from cognitive development to increase young
children’s challenge seeking and persistence through storybooks. Do these storybooks
become more effective as the main character in the story increases in similarity to the
child? Independently of this manipulation, Study 1 also looks at how children’s
beliefs about goodness and badness relate to their motivation. Are children who
believe that mistakes imply badness more likely to avoid challenge?
2) A second aim, in Study 2, includes the examination of the effectiveness of
this manipulation in older children (kindergarteners). Are older children better able to
generalize from a story to themselves, regardless of how similar they are to the main
14
character? Further, Study 2 compares the effectiveness of the manipulation in
populations of struggling and non-struggling students.
3) A third aim, in Study 3, is to explore whether persuading another might
increase the book’s effectiveness, by having children transmit the ideas of the
storybook to others, compared to children who actively explain what happened in the
story. Do children who teach a puppet to choose hard tasks show more challenge
seeking themselves? Furthermore, do the storybooks create robust challenge seeking,
even after children have experienced a real challenge?
4) The fourth aim, in Study 4, is to explore whether effects persist over time.
Do the effects last at least two weeks?
5) A fifth aim, in Study 5, is to examine more closely one particular element in
the story (praise for effort), and how it might affect motivation. Does hearing a story
containing effort/process praise increase challenge seeking and persistence relative to
a story that does not?
STUDY 1: MOTIVATION IN PRESCHOOLERS
The research cited in the Introduction was used to create storybooks to
encourage children to choose hard tasks and persist at them. The storybooks used
processes shown to be effective in younger (4-5-year-old) and/or older (9-12-year-old)
children, such as praise for effort, strategies to resist the temptation to give up, and a
role model who struggles before succeeding (Kamins & Dweck, 1999; Patterson &
Mischel, 1975; Schunk, Hanson, & Cox, 1987). Previous interventions have also
utilized this process of combining several factors to maximize effects on children’s
motivation (Miller, Brickman, & Bolen, 1975). Thus, the story was designed to teach
children that they can and should choose challenges and persist in the face of
difficulty. Many children are familiar with the story of “The Little Engine that Could”
(Piper, 1930), but the story of a train trying to get over a mountain may not be
personally meaningful to children. But what if the main character were the child
herself? What if the task were more relevant to her life?
The story included a concrete situation (choosing easy or hard puzzles) that
was highly similar to the dependent measures, since the primary purposes of this study
15
were: 1) to establish the effectiveness of the storybooks in increasing children’s
challenge seeking and persistence; and 2) to examine whether similarity to the main
character increased this effectiveness. In Study 1, 4- and 5-year-old preschool
children were read one of four books by a researcher. Three were treatment stories
and one was a control story. The treatment books involved a main character learning
to try hard and persist, and varied in the main character’s similarity to the participating
child. Children then responded to questions assessing their preference for easy or
challenging tasks (including puzzles, as in the story, and also other activities, to assess
children’s generalization to other types of tasks). Children also gave explanations for
their choice of tasks, to see whether they explicitly mentioned seeking or avoiding
challenge. They were then given the chance to persist on an actual puzzle.
In addition to these measures, a secondary goal of this study (and of Study 2)
was to explore how children’s beliefs relate to motivation. Some children hold beliefs
that have been related to helplessness and lack of motivation in the face of difficulty.
These children have a tendency to overgeneralize trait inferences on the basis of
limited information, and perceive a single instance of making a mistake or
misbehaving as diagnostic of overall badness. That is, they see mistakes and
misbehavior as conveying implications about global goodness and badness. For these
children, a single mistake takes on broader significance, and can have negative
motivational consequences. In this dissertation, these types of beliefs will be referred
to as “global badness beliefs,” as children who hold these beliefs make global
assumptions about goodness and badness based on limited evidence.
Previous research has found that children who endorse these global beliefs are
more likely to show helpless responses to challenge, such as negative affect, self-
blame, and lack of persistence (Heyman, Dweck, & Cain, 1992). While those results
are correlational, children’s beliefs were measured before they displayed helplessness,
suggesting that children who hold these beliefs are more likely to then overgeneralize
following negative experiences and thus are more likely to become helpless (Heyman,
Dweck, & Cain, 1992). Much previous research suggests that 4- and 5-year-old
children are remarkably resilient in general and show high motivation when they are
16
not currently faced with an overly challenging task (Cimpian, 2010). However, it is
possible that children’s beliefs may also predict their behavior even in the absence of
experienced difficulty, that is, when they are simply contemplating challenges.
Children who believe that making a mistake is bad may attempt to avoid taking on
challenges, because failing at a challenging task may have more global implications
for their self-worth. Thus, an additional goal of this study (and Study 2) was to
examine whether children’s beliefs are related to challenge seeking even in the
absence of current difficulty.
While previous studies had asked children to make these global badness
judgments about other children (e.g., Heyman, Dweck, & Cain, 1992; Kamins &
Dweck, 1999), this study added a new question tapping into children’s beliefs about
their own mistakes, which may be an even more powerful method of assessing beliefs
critical for their motivation (Dweck, 1999).
Method
Participants
Participants were 53 preschoolers (mean age = 4 years, 7 months; range: 4
years, 0 months to 5 years, 7 months; 18 male, 35 female; 32 Caucasian, nine Asian
American, four Latino/a, three African American, and five multiple ethnicities) at a
research nursery school. The majority of them came from middle- to upper-middle-
class backgrounds.
Materials
Participants were randomly assigned to hear one of four storybooks. (See
Appendix A for full texts.) Three treatment books involved a character that learned to
“keep trying” on a hard puzzle. The main character was either an animal (“Little
Bird”), another child (“Taylor”), or the participating child him/herself. “Animal” and
“another child” books were matched to the participant’s gender (McArthur & Eisen,
1976). A control book involved the participating child playing with a bouncing ball
(to control for potential effects of hearing a self-relevant story), with a storyline that
did not involve motivation. The books in which the participating child was the main
character were individually printed with the child’s name.
17
The researcher frequently made comments directly to the child to connect the
child in the story to him/her. This helped to ensure that the participating child knew
the story was about him/her and not a child with the same name. For example, after
reading “Carol’s parents gave her two new puzzles,” the researcher turned to the child
and said directly to her, “So YOUR parents gave YOU two new puzzles.”
In the treatment books, the main character was given a choice between an easy
puzzle and a hard puzzle, and chose the hard puzzle. After the character had struggled
on the puzzle, his or her parents gave encouragement and praise for trying (including
the phrases, “Just keep trying! You’ll see what happens,” and the child telling him or
herself, “I’ll just keep trying, and I’ll see what happens”), and he or she succeeded at
getting the pieces. The character then went to school and was given another choice
between an easy puzzle and a hard puzzle. Again, he or she chose the hard puzzle so
he or she could learn. Finally, the character’s classmates were impressed by the hard
choice and praised him or her (see Child, 1946).
Procedure
Children were brought to a research room individually by a researcher who
was familiar to them. First, children listened to the story read aloud by a researcher.
While listening and looking at the pictures, participants were asked several questions
to ensure they stayed engaged in the story. (See Appendix A.)
Challenge Seeking
After the story, children were given a series of eight choices between easy and
hard tasks to assess the effects on challenge seeking. (These questions were selected
based on the results of several pilot studies.) Four questions involved puzzle tasks
(similar to the story), and four questions involved other activities common in
preschool (e.g., building a tower with blocks) to assess whether the effect would
generalize to other tasks (Child & Adelsheim, 1944). (See Table 1.) After each
choice, children were told “Good choice!” and asked to explain why they chose that
one (Smiley & Dweck, 1994).
Persistence
18
Next, children worked on a moderately challenging puzzle but were told they
could switch at any time to another, obviously easier puzzle (to assess persistence).
Every minute and a half, children were reminded, “Remember, you can switch to the
other puzzle if you want.” There was no time limit.
Global Badness Beliefs
Finally, children answered two questions about their global beliefs about
badness: 1) in terms of others: “Imagine there’s a new boy/girl in your class. You
look over at his/her drawing and see that s/he did lots and lots of things wrong. Does
this mean that s/he is bad? [Yes or no?]” and 2) in terms of the self: “Now imagine
that you made a drawing and did lots and lots of things wrong. How would this make
you feel? Like you were not bad at all, a little bit bad, or really bad?” The “other
child” question was always matched to the participant’s gender.
Debriefing
Children were thanked for their hard work on the puzzle. Children who had
switched to the easy puzzle were asked if they wanted the researcher to help them
finish the challenging puzzle.
Results
The dependent measures (challenge seeking and persistence) were examined in
terms of three main predictors: condition, age, and global beliefs. Analyses also
looked at children’s explanations for their choices, and a motivational composite
combining challenge seeking and persistence. The effect of researcher was also tested
in all studies to ensure that the effect of researcher did not interact with condition
(which was the case in all analyses). Some children did not respond to all questions,
so the number of participants varied for different measures.
Overall, all three treatment books increased children’s challenge seeking, but
only the treatment story about the participating child increased persistence. For an
overall summary of results, see Table 2.
Challenge Seeking
The eight challenge-seeking items formed a reliable scale, α = .86 (and none of
the items increased the reliability of the scale if deleted). Effects on challenge seeking
19
were analyzed in two complementary ways: first, the main effects were examined
individually using Analysis of Variance (ANOVA). Second, the challenge seeking
data were also analyzed using a regression model, in which all the independent
variables were analyzed simultaneously (along with random effects of participant) to
gain a more complete picture of how these variables might work together.
Effect of Condition
As predicted, all three treatment books taken together significantly increased
overall challenge seeking, compared to the control condition. Across all eight
challenge-seeking choices, an overall ANOVA indicated that the conditions were
significantly different from each other, F(3,49) = 3.31, p = .028. (See Figure 1.) Post-
hoc contrasts revealed that this overall difference was driven by the difference
between the three treatment conditions combined compared to the control condition.
Children in the three treatment conditions combined chose significantly more
challenging items than children in the control condition, t(49) = 3.14, p = .003, which
supported the primary hypothesis that these books would increase challenge seeking.1
However, the three treatment books did not differ from one another (ps > .82) and
hence, for this measure, did not support the hypothesis that young children would be
more influenced by a book that featured themselves as the main character than a book
that featured someone else as the main character. (For results analyzed in terms of
puzzle choices and activity choices separately, see Table 4.) Thus, as predicted,
hearing storybooks that incorporated praise for effort, strategies for persisting, and
concrete examples of a role model who coped with failure significantly increased
preschool children’s challenge seeking.
Age
There was a marginal interaction between age and condition (treatment vs.
control). The treatment books were marginally more effective for children who were
four and a half or older (younger control: M = 2.67, SD = 2.60; older control: M =
1 Individually, each treatment condition was also significantly higher than the control condition; animal vs. control: t(49) = 2.68, p = .010; another child vs. control: t(49) = 2.46, p = .018; participating child vs. control: t(49) = 2.59, p = .013.
20
1.25, SD = 1.89; younger treatment: M = 4.00, SD = 2.66; older treatment: M = 5.75,
SD = 2.36), F(1,49) = 3.50, p = .067. The main effect of condition remained
significant overall (p = .001). Thus, unsurprisingly, while the intervention was
effective for all children, it was especially effective for older children, who were better
able to generalize from a story to themselves.
Global Beliefs
Children’s answers to the global belief questions (e.g., “If a child makes
mistakes, does that mean she is bad?”) were coded as one point for a “yes” or “little
bad,” and two points for “really bad.” The two questions were marginally correlated,
r(51) = .25, p = .08.
As predicted, children’s overall global beliefs were related to one aspect of
motivation. Beliefs were significantly negatively correlated with total number of
challenge-seeking choices, r(51) = -.44, p = .001. (See Figure 2.) Children who
endorsed more global beliefs (i.e., believed that making mistakes implies global
badness) were less willing to take on challenging tasks.
If each global belief question was considered separately, both questions
showed the same pattern: the question about the self was significantly correlated with
challenge-seeking choices, r(53) = -.38, p = .005, while the question about others was
marginally significant, r(51) = -.27, p = .057 (the two overlapping correlations were
not significantly different from each other, z = .67, p = .25). This suggests that global
belief questions regarding both the self and others may be useful predictors of
children’s challenge seeking, although combining both questions may be even more
useful.
Regression Model
Because challenge seeking involved repeated measures and a binary outcome
variable, the challenge-seeking choice data was also analyzed using a logistic mixed-
effects model that analyzed the effects of all independent variables simultaneously
(Baayen, Davidson, & Bates, 2008; Bates & Sarkar, 2007). This type of analysis is an
alternative to the use of analysis of variance (ANOVA) models for categorical
dependent measures, and has been argued to be less problematic (e.g., when
21
confidence intervals exceed interpretable results) and less likely to lead to spurious
null results or spurious significances (Jaeger, 2008; see also Perry, Samuelson,
Malloy, & Schiffer, 2010). Condition was coded using three orthogonal Helmert style
contrasts (see Table 3). Contrast 1 tested whether the three treatment conditions were
different from the control. Contrast 2 tested whether the participating child condition
was different from the other two treatment conditions. Contrast 3 tested whether the
animal and another child conditions differed from each other.
The results of the regression model confirmed the findings of the previous
analyses with significant effects of both condition and global beliefs. The initial
model was a completely specified model with condition, item type (puzzles vs.
activities), global beliefs, and age (older or younger than 4 years, 6 months) as fixed
effects, and participant and item as random effects. Uninformative effects were
systematically removed by comparing models (see Perry, Samuelson, Malloy, &
Schiffer, 2010). For children’s global beliefs, missing scores were replaced by the
overall mean, and beliefs were centered.
The final model included condition, item type,2 and global beliefs as predictors
of challenge seeking choices. Most importantly, children in the treatment conditions
chose significantly more challenging options compared to children in the control
condition, z = 2.59, p = .001. In addition, children who endorsed more global beliefs
(about both self and others) chose significantly fewer challenging options, z = -2.40, p
= .017. (While age was not a significant predictor in the final model, it did show a
marginal interaction with condition when only condition and age were included as
fixed effects, z = 1.91, p = .056.) Thus, as predicted and supporting prior analyses,
the best overall model included condition and global beliefs as predictors of challenge
seeking choices.
Explanations
2 Across all conditions, children were more likely to choose hard puzzles than hard activities, z = 4.41, p < .001. This suggests that children were better able to generalize from the story to real life in terms of tasks that were similar to those explicitly mentioned in the story.
22
In addition to the effects on challenge seeking, children’s explanations of their
choices were also coded, to give further insight into children’s thought processes.
Children’s explanations of their choices were coded as to whether they were related to
seeking challenge (e.g., “I want to try and try,” “I like doing hard things,” “I want to
learn”) or avoiding challenge (e.g., “because I like easy ones,” “because I can’t do the
non-easy puzzle”). An overall ANOVA revealed that the conditions were marginally
different in the number of challenge-seeking explanations that children gave, F(3,49)
= 2.75, p = .053 (control: M = 0.77, SD = 1.79; animal: M = 2.46, SD = 2.33, another
child: M = 3.46, SD = 3.05, participating child: M = 2.50, SD = 2.41). Post-hoc
comparisons showed that children who heard the three treatment books gave
significantly more challenge-seeking explanations than children in the control
condition, t(49) = 2.62, p = .012. In terms of challenge-avoiding explanations, post-
hoc comparisons showed that children who heard the three treatment books were
marginally less likely to give challenge-avoiding explanations, compared to the
control condition, t(49) = -1.81, p = .077. Thus, in addition to being more likely to
take on challenging tasks, children in the treatment conditions were more likely to
give explanations for those choices that referenced wanting or valuing challenge, and
less likely to give explanations involving a desire to avoid challenge. (See Table 5.)
Children were indeed choosing harder tasks specifically because they wanted to learn
or wanted a challenge, and not for other reasons.
Persistence
As predicted, there was an effect of condition on persistence. Persistence was
examined in terms of whether children finished the puzzle or quit to work on the easy
puzzle. (For further analyses of how long children persisted before quitting the puzzle
and the number of puzzle pieces correctly placed, see Table 6.) An overall chi-square
revealed that the conditions were marginally different from each other in persistence,
χ2(3, N = 53) = 6.63, p = .085. Specifically, the treatment book about the participating
child increased persistence on the challenging puzzle compared to the other three
23
conditions, χ2(1, N = 53) = 5.81, p = .016.3 (See Figure 3.)4 Thus, in line with the
original hypothesis, hearing a story in which their personal identity was connected to
Significantly more older children persisted on the puzzle than younger
children, χ2(1, N = 53) = 6.94, p = .008 (older children: 71% persisted; younger
children: 34%). However, there was no interaction with condition (p = .98).
Global Beliefs
Unlike the choice measure, there was no relationship between beliefs and
persistence (p = .81).
Motivation Composite
Challenge seeking and persistence are related theoretically and both carry
predictive importance for motivation and achievement-related outcomes (as a potential
marker of academic trajectories). For example, Dweck and Leggett (1988) suggest that
both seeking challenge and showing high persistence are part of a mastery-oriented
approach to learning. Thus, both challenge seeking and persistence were combined in
a motivation composite to examine whether the composite varied as similarity to the
main character increased. (In this sample, persistence was not correlated with
challenge seeking, r(53) = -.03, p = .84.)
Effect of Condition
The overall motivation composite revealed a significant effect of condition.
Challenge-seeking choices (how many hard options chosen out of 8) and a dummy
3 Children in the participating child condition were marginally more likely to complete the puzzle than children in the control condition, χ2(1, N = 27) = 3.04, p = .08, and than children in the another child condition, χ2(1, N = 27) = 3.04, p = .08, and significantly more likely than children in the animal condition, χ2(1, N = 27) = 6.24, p = .013. 4 Some children were given an 8-piece puzzle while others were given a 14-piece puzzle to examine effects at different levels of difficulty, although the proportion of children given each puzzle did not differ by condition, p = .99. The pattern of results was identical for both puzzles, and there was no interaction between condition and puzzle type on persistence, p = .67.
24
coding of persistence (whether or not children finished the challenging puzzle) were
equally weighted and combined into an overall motivation composite ranging from 0
to 1. An overall ANOVA revealed that the conditions were significantly different
from each other, F(3,49) = 3.12, p = .034. In this case, the linear trend was also
examined to determine the effects of varying similarity to the main character. The
linear trend was significant, F(1,49) = 8.89, p = .004. (See Figure 4.) Post-hoc
comparisons showed that the participating child condition was significantly higher in
composite motivation than the animal condition, t(49) = 2.09, p = .042. The
participating child condition also showed a trend toward being higher than the another
child condition, t(49) = 1.52, p = .14. These results indicated that, as predicted,
overall motivation increased as similarity to the main character increased—the
“participating child” condition was slightly more effective than the “another child”
condition, and significantly more effective than the “animal” condition.
Discussion
These results support the hypothesis that storybooks about a character learning
to try hard things and persist can lead young children to choose harder tasks and
persist at them.
All three treatment books were effective in increasing the number of
challenging tasks that children chose. This effect held for hard tasks related to puzzles
(as mentioned in the story) and to a lesser extent for tasks related to other activities
such as singing and playing games, suggesting that children in all three treatment
conditions were able to generalize the idea of the story to tasks other than puzzles.
Although children were more likely to select the challenging choice for puzzles than
other activities, the treatment books led to significantly more choices of challenge both
for puzzles and for other activities.
Interestingly, the book about themselves learning to try hard was particularly
effective in making children persist on an actual puzzle. Thus, personal identity was
not critical for children’s challenge seeking choices, but had a significant effect on
their motivated behavior on a real puzzle. Perhaps this book was more effective in
convincing these children that they were the kind of children who would not give up.
25
They heard a story in which they tried hard and refused to give up, and this changed
their behavior when they encountered a challenge. Perhaps the other treatment books
helped children have a positive attitude toward challenges, but that effect was not
robust enough to help children persist. Overall, the motivation composite did vary as
a function of how similar the child was to the main character in the story, indicating
that making the main character more similar to the child increased his or her overall
motivation.
As hypothesized, children’s beliefs also played an important role in one aspect
of motivation, as children with more global beliefs about badness were less likely to
choose challenging tasks. Independently of condition, children’s beliefs were
significantly related to the number of hard items they selected. Perhaps these children,
who believed that mistakes carried global implications, were more concerned about
the potential consequences of failure if the task proved too challenging. Previous
studies have shown that children’s beliefs are related to their responses when they are
experiencing difficulty or challenge. The current findings suggest that children’s
beliefs have important consequences even when children are not currently
experiencing challenge. Indeed, Study 1 suggests that children who hold more global
beliefs may be less likely to choose to put themselves into challenging situations.
This study was only the first step toward gaining a richer understanding of
motivation in early childhood. One potential concern about this study is that it
allowed for the possibility of social desirability effects. In essence, the treatment
storybooks told children that what the researcher valued was motivation and the
choice of hard tasks. However, teaching children to value challenges is a theoretically
important component of the effect, and social desirability cannot entirely explain the
effects. Only children who heard a book about themselves were more likely to persist.
Moreover, Study 4 suggests that the effects on challenge seeking last over a longer
time period, indicating that the storybooks may lead to real and lasting motivational
change.
Finally, several results indicated the importance of age in the effects of the
storybooks on children’s motivation. In particular, the effects on challenge seeking
26
were stronger in older children than younger children. This suggests that there may be
interesting developmental differences in the effectiveness of this manipulation.
To gain a better understanding of the effectiveness of this manipulation across
different groups, Study 2 examined effects in older children. More specifically, Study
2 examined effects in two samples of kindergarten children: a group who was
struggling academically and a non-struggling group.
STUDY 2: MOTIVATION IN OLDER CHILDREN
The results of Study 1 suggest that it is possible to increase the challenge
seeking and persistence of young children with storybooks. What about the challenge
seeking and persistence of slightly older students? In terms of generalization, older
children should have less trouble than younger children in transferring what they have
learned to a new situation. In terms of the effect of the storybooks, this improved
generalization ability of older children indicates that the similarity between the main
character and the participant may be less critical for older children. Further, older
children are in a more formal school setting, and some are beginning to experience
academic challenge and struggle.
Study 2 tested two primary questions. First, does the difference between
treatment books matter for older children? Second, how effective are the treatment
books for a group of students struggling academically, compared to a non-struggling
group? To test these questions, two groups of 5- and 6-year-old children participated
in the same procedure as in Study 1. The non-struggling group was comprised of
middle-to upper-middle-class students in a suburban public school, from a population
highly similar to the children in Study 1, and the other group was comprised of
students in a literacy summer program, who had been nominated by their kindergarten
teachers as needing extra instruction.
Method
Participants
Participants were 70 children in kindergarten (N = 37) or in a summer school
program immediately following kindergarten (N = 33); mean age = 6 years, 1 month;
27
range: 5 years, 5 months to 7 years, 6 months; 32 male, 38 female. The two school
samples did not differ in age or gender (ps > .86).
According to school demographic information, there were 34 Caucasian, 11
Asian American, 16 Latino/a, one African American, and eight “Other Ethnicity”
participants. The summer school sample had a higher proportion of Latino/a children
(39%) and a lower proportion of Caucasian children (33%), compared to the public
school population, 8% Latino/a and 62% Caucasian (p = .018). (Thus, the struggling
and non-struggling samples also differed in demographics and school characteristics;
however, the struggling/non-struggling distinction is most relevant for the purposes of
this study.)
Some children did not respond to all questions, so the number of participants
varied for different measures.
Materials
Participants heard the same books as Study 1. Children were randomly
assigned to the control condition or one of three (animal, another child, or
participating child) treatment conditions.
Procedure
The procedure was identical to Study 1. Children responded to questions
assessing: 1) challenge seeking choices and explanations, 2) persistence on a
challenging puzzle, and 3) global beliefs about badness.
Results
See Table 2 for a summary of overall results. In addition to condition, age, and
global beliefs, analyses also examined differences between the two school groups.
As expected, there were no differences among the three treatment conditions
for either challenge seeking or persistence. However, there were intriguing
differences between the two groups of students in the effectiveness of the treatment
conditions compared to the control condition. The intervention had a greater effect on
challenge seeking for the non-struggling students (particularly for the puzzle items),
but had a greater effect on persistence for the struggling students. While it may be
more difficult to encourage challenge seeking among children who have struggled
28
with academic challenges, the results are promising in suggesting that struggling
students may be responsive to intervention in terms of behavioral persistence.
Challenge Seeking
The eight challenge-seeking items formed a reliable scale, α = .87 (and none of
the items increased the reliability of the scale if deleted). For the results of children’s
explanations for their choices, see Table 5. Again, analyses are presented twice: first,
main effects were examined individually using ANOVAs, and second, a regression
model examined all independent variables simultaneously.
Effect of Condition
First, differences among the three treatment conditions were examined. There
were no significant differences among treatment conditions in either school sample (ps
> .20), so further analyses collapsed across treatment conditions.
Among children in the public school, there was a trend for children in the
treatment conditions to show greater challenge seeking compared to children in the
control condition (control: M = 4.56, SD = 3.05; treatment: M = 5.92, SD = 2.48),
t(32) = -1.33, p = .19. Among children at the literacy summer school, there was no
effect on challenge seeking (control: M = 3.22, SD = 3.19; treatment: M = 3.05, SD =
2.21), t(29) = 0.18, n.s.
Global Beliefs
Beliefs were examined combining both groups of students. As in Study 1, the
two global belief questions (self and other) were significantly correlated with each
other, r(67) = .24, p = .048.
As predicted, children’s beliefs were again related to motivational outcomes.
Beliefs were significantly negatively correlated with total number of challenging
choices, r(62) = -.28, p = .026.5 (See Figure 5.) As in Study 1, children who endorsed
more global beliefs about badness were less willing to take on challenging tasks.
School Site
There was a main effect of school site. Children at the public school were
5 Considered separately, each global belief question was also significantly correlated with challenge seeking; self: r(63) = -.25, p = .048; other: r(64) = -.26, p = .035.
29
considerably higher in challenge seeking than children at the literacy summer program
(public school: M = 5.56, SD = 2.67; summer program: M = 3.10, SD = 2.48), t(63) = -
3.84, p < .001. This effect did not interact with condition, suggesting that there were
large differences between these two groups in how willing they were to take on
challenges. It is unclear whether this difference may be a cause or consequence of the
differences in academic experiences of these children. See further discussion below of
differential effects of school setting.
Regression Model
The results of the regression supported the findings of the ANOVA analyses,
in terms of a main effect of school site and a slight tendency for students in the public
school in the treatment conditions to show an increase in challenge seeking. As in
Study 1, challenge-seeking choices were analyzed using a logistic mixed-effects
model (Bates & Sarkar, 2007). Condition was again coded using three orthogonal
Helmert style contrasts (see Table 3).
I began with a completely specified model with condition, item type (puzzles
vs. activities), global beliefs, school, and age (centered around the mean) as fixed
effects, and participant and item as random effects. As in Study 1, models were
compared systematically to remove uninformative effects. For children’s global
beliefs, missing scores were replaced by the overall mean, and beliefs were centered.
Missing challenge-seeking responses were also replaced by the overall mean.
The final model included school, age,6 and a three-way interaction involving
condition as predictors of challenge seeking choices. First, school was a significant
predictor. Children at the literacy summer school were less likely to choose
challenging tasks compared to children at the public school, z = 4.19, p < .001.
Interestingly, there was a significant three-way interaction between school, item type,
and condition (treatment versus control), z = 2.63, p = .009. The treatment conditions
increased challenge seeking primarily in terms of puzzle items for children at the
6 Age (continuous) was a marginally significant predictor, z = 1.72, p = .086, but did not interact with other variables. As in Study 1, older children were more likely to choose challenging tasks than younger children.
30
public school. This suggests that the treatment condition did have an effect on
challenge seeking, but chiefly for students at the public school, and chiefly in terms of
challenging puzzles.
In summary, Study 2 addressed two primary questions: are there differences
among the treatment conditions for older children, and are there differences for
struggling and non-struggling students? First, there was no difference between
treatment conditions in terms of challenge seeking for older children. Regardless of
the main character of the story, children showed equivalent challenge seeking.
Second, it appears that the intervention was not enough to increase the challenge
seeking of struggling students, although it did have some effect on students at the
public school.
Persistence
The analyses of persistence revealed that the three treatment books
significantly increased persistence compared to the control condition.
Again, persistence was examined in terms of whether children finished the
puzzle or quit to work on the easy puzzle. (See Table 6 for additional analyses of
persistence.) First, differences among the three treatment conditions were examined.
There were no significant differences at either school site, so analyses collapsed across
treatment conditions (ps > .30).
Although there were no significant differences in persistence between children
in the treatment and control conditions at the public school, the results were in the
predicted direction (control: 50% of children persisted; treatment: 70% of children
persisted), χ2(1, N = 37) = 1.33, p = .25.
In contrast, there was a highly significant difference between the treatment and
control conditions at the literacy summer school (control: 33% of children persisted;
treatment: 88% of children persisted), χ2(1, N = 33) = 9.68, p = .002.
Across both school groups, the overall effect of condition was significant, χ2(1,
N = 70) = 8.48, p = .004.7 (See Figure 6.) Thus, these older children were more likely
7 Some children were given an 11-piece puzzle while others were given a 14-piece puzzle to examine effects at different levels of difficulty, although the proportion of
31
to persist at a moderately challenging puzzle after hearing a motivational book,
regardless of whether the main character was an animal, another child, or the
participating child.
School Site
There was no difference between sites in the overall percentage of children who
persisted (literacy summer school: 73% persisted; public school: 65%), n.s. However,
logistic regression revealed a significant interaction between school and condition, z =
-3.66, p < .001, indicating that the effect of condition on persistence was larger for
summer school students than for public school students. Effect size was calculated
(using proportion of children who finished the puzzle) for each school site and
compared the effect of the three treatment books to the control book. The effect size
was very high at the literacy summer school, d = -1.13, and moderate at the public
school, d = -0.40.
In summary, recall the primary questions of Study 2: are there differences
among the treatment conditions for older children, and are there differences for
struggling vs. non-struggling students? In terms of persistence, just as for challenge
seeking, there were no differences among treatment conditions for older children.
Children were just as likely to persist after hearing a book about an animal, another
child, or themselves persisting. However, in contrast to the challenge seeking
measure, the effect of the intervention on persistence was larger for struggling
students than for non-struggling students.
Motivation Composite
Correlation between Challenge Seeking and Persistence
Persistence was positively but not significantly correlated with hard choices,
r(65) = .15, p = .23.
Effect of Condition
children given each puzzle did not differ by condition, p = .92. There was no difference in how likely children were to persist on either puzzle, and there was no interaction between condition and puzzle type on the proportion of children who finished the puzzle, p = .40.
32
Again, an overall motivation composite was created to explore effects on
overall motivation. In a motivation composite including all eight challenge-seeking
items and a dummy coding of persistence, once again, there were no differences
among the three treatment conditions at either school (ps > .29), so the treatment
conditions were combined. Overall, across both schools, the effect of condition was
significant, t(63) = -2.69, p = .009. (See Figure 7.) The difference between the
treatment and control conditions was significant for students in the summer program
(control: M = .37, SD = .31; treatment: M = .62, SD = .24), t(29) = -2.43, p = .021, and
showed a trend toward being significant for students in the public school (control: M =
.51, SD = .40; treatment: M = .71, SD = .33), t(32) = -1.51, p = .14.
In summary, when considering overall motivation, again, there were no
differences among treatment conditions for older children. However, across both
samples of students, the treatment storybooks lead to significantly greater overall
motivation.
Discussion
These results suggest that the intervention is also effective for older students,
and has real potential to make a difference for struggling students. These results also
support findings from Study 1 that children who hold global beliefs about badness are
less willing to take on challenge tasks.
Across both challenge seeking and persistence, there were no differences among
treatment conditions. This finding supports results from the literature suggesting that
older children are better able to generalize than younger children. In the case of the
storybooks, kindergarten children may have been able to connect the story equally
well to their own lives, regardless of who the main character in the story was. This
was particularly true for the persistence measure, indicating that kindergarteners were
better able to generalize what they had learned to change their own behavior.
Although the effects for challenge seeking were small (perhaps because
kindergarten children are in a more formal and evaluative academic setting compared
to preschool children), the intervention showed a trend among non-struggling students
for increased challenge seeking. However, there was no effect for struggling students.
33
Struggling students may have had many experiences with academic challenges, and
may be more likely to consider “hard” tasks beyond their ability. If students have
different understandings of what the word “hard” means based on their skill level, then
“hard” for struggling children may signify “too hard” to them. Future research could
examine this question by asking children (at the end of the research procedure) to
identify which of various puzzles they consider “hard” and which the researcher meant
by “hard.” The context of the story may also be interpreted differently by struggling
and non-struggling students—perhaps the non-struggling students were more familiar
with a classroom context (as in the story) that encouraged challenge seeking and
supported effort. In terms of the current results, while it may be more difficult to
encourage challenge seeking among children who have struggled with academic
challenges, the results were highly promising in suggesting that struggling students
may be responsive in terms of behavioral persistence.
For kindergarten children, all three treatment books increased persistence, and
this effect was even stronger for the struggling students. Although they may have
avoided choosing challenging tasks, the intervention helped them find the motivation
to keep working when they encountered a moderately challenging puzzle. They may
have been even more likely than the non-struggling students to identify with the
struggling role model in the story, and been encouraged to keep working when they
encountered a challenge. As discussed in the Introduction, the achievement gap in this
country is evident from the time that students begin elementary school, and widens
over time (Phillips, Crouse, & Ralph, 1998). If this manipulation could be used to
increase the persistence of children who are slightly academically behind, that
persistence could translate into slightly better learning, which could then create a
recursive cycle of improved motivation and enhanced learning. Even a small change
in students’ academic trajectories at this point has the potential to create larger effects
in the long run (see Cohen et al., 2009).
Overall, Studies 1 and 2 indicate that it is possible to use storybooks to increase
young children’s motivation. However, several concerns remain. For example, one
concern with Studies 1 and 2 is that children had no frame of reference when choosing
34
between “easy” and “hard” tasks. When the storybook had an effect on challenge
seeking and students desired hard tasks, did they really know what they were
choosing? How strong was their desire for challenge? Thus, subsequent studies, with
preschoolers, focused on challenge seeking and gave children the opportunity to
express a desire for challenge both before and after they experienced genuine
difficulty. Does the storybook intervention help children remain challenge seeking,
even after they have experienced an extremely challenging task? Thus, for Studies 3
and 4, challenge seeking was measured both before and after children experienced a
highly challenging puzzle. Moreover, the next study also explored additional ways to
increase the effectiveness of the storybooks: explanation of the value of challenges to
another person, either with or without the intention to persuade that other person to
take on a challenging task.
STUDY 3: PERSUASION AND CHALLENGE SEEKING
In Study 3, two treatment conditions were compared to a control condition.
The first treatment condition involved children attempting to persuade another person
(in this case, a puppet) to take on a hard task. As discussed in the Introduction,
persuading another person often leads to greater attitude change for the persuader as
well. In terms of challenge seeking, dissonance could potentially occur if children
have told someone else to choose challenges, yet choose an easier task themselves. To
avoid this dissonance, they may choose harder tasks for themselves in order to be
consistent with their previous behavior of telling someone else that challenges are
good. Further, in this case, the persuasion involved an explanation to the puppet of
why challenges might be beneficial. Generating explanations (e.g., explicitly stating
why hard tasks are good) may help children in generalizing what they learned in a
previous situation to a new situation (Brown & Kane, 1988; Williams & Lombrozo,
2010). That is, explaining why challenges are beneficial may help children generalize
from the story to show greater challenge seeking in their own lives. This is similar to
a tactic frequently used by highly effective tutors, namely, to ask the student to reflect
on a solution or principle and consider how it might apply to other contexts (Lepper &
Woolverton, 2002). To control for the effects of this explanation, that treatment
35
(“explain + persuade”) condition was compared to the effect of a treatment condition
in which children explained to the researcher why hard tasks are good without any
element of persuasion (“explain”).
These two treatment conditions (in which the participating child was the main
character of the motivational story) were compared to a control condition in which
children answered questions about the control storybook.
Method
Participants
Participants were 36 children in at a research nursery school (mean = 4 years, 9
months; range: 4 years, 6 months to 5 years, 7 months; 14 male, 22 female; 13
Caucasian, 10 Asian American, nine Latino/a, two Middle Eastern, one African
American, and one multiple ethnicity). Because this was potentially a more subtle
manipulation (with a small distinction between treatment conditions), only children
who were at least four and a half years old participated.
Materials
Participants heard the same books as children in the participating child and
control conditions in Study 1, with slight modifications to ensure children were more
involved in the story by answering questions about it. While children in Studies 1 and
2 responded to two questions during the story (e.g., “What do YOU think you/he/she
should do?” and “Which one did you/he/she pick?”), in Study 3 children answered
additional questions during the story to keep them engaged.
For the puzzle measure, a different puzzle was used to determine the hardiness
of the effect of the storybook on challenge seeking—an extremely challenging 20-
piece jigsaw puzzle.
Children were randomly assigned to one of three conditions: control, treatment
book with explanation questions only (“explain”), or treatment book with explanation
of the value of challenges and inducement to persuade a puppet (“explain +
persuade”).
Procedure
Storybook and Questions
36
Control condition. After hearing the story, children in the control condition
answered three questions about the story they had heard (“What happened in that
story?” “Which ball did you pick?” and “Why that one was a good one to pick?”).
Explain condition. After hearing the story, children in the explain condition
answered three questions about the story they had heard (“What happened in that
story?” “Which puzzle did you pick?” and “Why that one was a good one to pick?”).
The third question elicited explanations from children about the advantages of taking
on challenges.
Explain + persuade condition. After hearing the story, children in the explain
+ persuade condition were introduced to a gender-matched puppet who could choose
an easy or a hard puzzle, and were asked to teach the puppet what they had learned
from the book in choosing: “Which one should Sam pick?” The puppet then asked
which one s/he should pick (“Which one do you think I should pick?”), why s/he
should pick that one (“Why should I pick that one?”), and why it was good for him/her
to choose hard things (“Why is it good for me to learn hard things?”). Thus, children
had the opportunity to make a recommendation, attempt to persuade the puppet, and
explain why taking on challenges is beneficial. If children suggested that the puppet
choose the easy puzzle, they were reminded to think about what happened in the book,
and asked again.
Emotion
To make sure that mood was not responsible for condition effects, children
were asked to rate how happy they felt (on a 6-point scale). This question was
repeated immediately after children worked on the puzzle as well.
Challenge Seeking
Next, children responded to four of the challenge-seeking items (“pre-
challenge”) and gave explanations for their choices. The remaining four challenge-
seeking questions were asked after children worked on the challenging puzzle (“post-
challenge”), which allowed an examination of how the intervening challenge
experience affected children’s responses. The pre- and post-challenge sets of four
questions each included two puzzle items and two activity items.
37
Puzzle
Children were given a difficult puzzle and, so as not to induce an overly high
degree of frustration, told they could switch to another (visibly easier) puzzle at any
point. This puzzle had 20 pieces and each piece was jigsaw shaped, which made it
extremely challenging. A five-minute time limit was established so that children
would not be disrupted from the classroom for an unlimited period of time.
Debriefing
All children who had not previously finished the challenging puzzle were then
told, “Now let’s finish this puzzle together,” and the researcher scaffolded children at
the appropriate level to help them successfully complete the puzzle. Finally, children
were thanked for their hard work on the puzzle and returned to the classroom.
Results
There was a failure of random assignment to condition, such that children in
the control condition were significantly older than children in the two treatment
conditions, F(1,34) = 5.49, p = .025 (control: M = 4 years, 10 months; treatment: M =
4 years, 8 months). However, controlling for age, the results replicated the effects of
the treatment books on overall challenge seeking. (See Table 2.) Children in both
treatment conditions were significantly more likely to choose challenging tasks than
children in the control condition, but this effect was not robust enough to withstand
children’s subsequent experience of an extremely challenging task.
Emotions
There were no significant differences by condition in how happy children
reported that they felt, either immediately after the story (control: M = 5.00 out of 6,
SD = 1.61; explain: M = 5.85, SD = 0.38; explain + persuade: M = 5.58, SD = 0.79),
F(2,33) = 2.09, p = .14, or immediately after working on the puzzle (control: M = 5.27
out of 6, SD = 1.27; explain: M = 5.31, SD = 1.11; explain + persuade: M = 5.42, SD =
1.00), F(2,33) = 0.05, p = .95.
Challenge Seeking
The eight challenge-seeking items formed a reliable scale, α = .87. For the
results of children’s explanations for their overall choices, see Table 5. Again,
38
analyses are presented twice: first, main effects were examined individually using
ANOVAs, and second, a regression model examined all independent variables
simultaneously.
Effect of Condition
Both treatment conditions led to greater challenge seeking than the control
condition. In terms of number of overall hard choices selected, there was no
significant difference between the two treatment groups, t(33) = -0.10, p = .92.
Looking at pre-challenge and post-challenge choices separately, there were also no
differences between the two treatment groups, ps > .73. Thus, further analyses
collapsed across treatment condition. Controlling for age in months, condition
(combined treatment conditions vs. control) significantly predicted challenge seeking,
F(1,33) = 6.74, p = .014. (See Figures 8 and 9 .) Older children also showed greater
challenge seeking, F(1,33) = 5.85, p = .021. Results remained the same controlling
for the emotion that children reported after hearing the story.
Thus, although there was a tendency for older children to choose more
challenging tasks (as in Study 1), the treatment conditions overall led children to be
more likely to take on challenges.
Pre- and Post-Challenge
When challenge-seeking choices were divided into pre- and post-challenge
items, the effect of condition was significant for items assessed pre-challenge, and
marginally significant for items assessed post-challenge. Controlling for age, children
in the treatment conditions were significantly more likely to choose challenging tasks
before they experienced a challenge, F(1,33) = 7.34, p = .011, d = .82. After they
experienced a challenge, children in the treatment conditions were marginally more
likely to choose challenging task, F(1,33) = 3.33, p = .077, d = .31. The interaction
between assessment (pre- or post-challenge) and condition was marginally significant,
F(1,34) = 3.40, p = .074. Thus, the effect of the intervention was marginally weaker
after children directly experienced an extremely challenging situation.
Regression Model
39
The results of the regression model supported the previous analyses, indicating
a main effect of condition and an interaction between condition and time (pre- or post-
challenge). As in Studies 1 and 2, overall challenge-seeking choices were analyzed
using a logistic mixed-effects model (Bates & Sarkar, 2007), and looked in particular
at effects of condition and item time (pre- and post-challenge). Condition was coded
using two orthogonal Helmert style contrasts. Contrast 1 tested whether the two
treatment conditions were different from the control condition. Contrast 2 tested
whether the two treatment conditions differed from each other. The initial model was
a completely specified model with condition, item type (puzzles vs. activities), item
time (pre- or post-challenge), and age (older or younger than 4 years, 8 months) as
fixed effects, and participant and item as random effects. As before, models were
compared systematically to remove uninformative effects.
The final model included only condition and item time8 as predictors of
challenge seeking choices. Condition was a marginally significant predictor in terms
of the two treatment conditions versus the control condition, z = 1.66, p = .097. As
discussed previously, children in the treatment conditions were more likely to choose
challenging tasks than children in the control condition.
Importantly, there was a marginally significant interaction between condition
and item time, z = 1.76, p = .078. (See Table 4 for effects of condition on challenge
seeking pre- and post-challenge.) Children in the treatment conditions were
significantly different from those in the control condition before working on the
challenging puzzle, but not after working on the puzzle. This suggests that the effect
of the intervention was not robust enough to fully withstand children’s experience of
an extremely challenging puzzle.
Discussion
Persuasion, in the form of instructing a puppet of the value of choosing hard
tasks and persisting, did not appear to increase challenge seeking above a condition in
8 Children were more likely to choose challenging tasks pre-challenge than post-challenge, z = 3.46, p < .001.
40
which children explained the value of choosing hard tasks. There are several potential
explanations for these results.
First, it is possible that young children need more explicit emphasis and
reminders of their behavior (e.g., that they have said that hard tasks are good) in order
to show motivational change (e.g., to demonstrate a positive attitude toward hard tasks
by choosing more hard tasks) in accordance with that behavior. Perhaps the
persuasion experience was not salient enough, as the puppet went away after the child
explained the value of challenges, and children were not reminded of what they had
told the puppet. Furthermore, previous research suggests that attitudes and behaviors
show little consistency in young children unless linked in some explicit way, such as a
group identity (Master & Walton, 2011; see also Patterson & Bigler, 2006). Future
research should examine the effects of inducing children to advocate for the choice of
hard tasks using minimal and unobtrusive coercion (see Lepper, Nisbett, & Greene,
1973), giving children repeated reminders of that behavior, and then examining
attitude change (in terms of whether children show a greater preference for hard
tasks).
Second, it is also possible that the explain condition was equally effective for
independent reasons. Perhaps thinking about why the hard puzzle was good to choose
and constructing an explanation for the researcher was just as effective in boosting
challenge seeking as telling the puppet to choose the hard one and constructing an
explanation for the puppet. When children generate their own explanations or
examples, they may generalize knowledge more easily (Brown & Kane, 1988). For
example, in one study, 3-year-old children were given several sets of analogous
problems and were trained in one of several conditions. One group was asked to say
how the first two analogous problems were similar, and another group was told to
teach a puppet the two solutions to the problems. Both groups were significantly
better than a control group at solving the third analogous problem (Brown & Kane,
1988). Indeed, there were no differences in that study between children who
explained the link between two solutions to a problem and children who taught those
solutions to a puppet. Although in the current study the explain condition was
41
included as a strong control for the explain + persuade condition, future research
should examine whether each of these conditions increases challenge seeking above a
situation in which children are exposed to the story but do not answer questions about
it.
The other primary result of this study was the interaction between condition (in
terms of treatment vs. control) and the time point at which children responded (either
before or after the challenging experience). Children in the treatment conditions were
significantly more likely to choose challenging tasks, but only before they experienced
the challenging puzzle. The effect was weaker after they experienced the challenge.
This suggests that the effects of the treatment books may not be robust enough to fully
withstand children’s experience of challenges in the real world. However, it is also
possible that the treatment storybooks may become more effective over a longer
period of time. Perhaps children need real-world experience with a variety of
challenging tasks to connect the message of the storybooks to the value of challenges
in their own lives. This possibility is examined in Study 4.
Indeed, in evaluating the findings of Studies 1, 2, and 3, an important issue to
consider is how applicable these results (from laboratory studies) are to long-lasting
motivation in actual classrooms. Previous short-term studies that have been conducted
in classrooms or that examined long-term academic effects have found promising
results with long-term effects on students’ motivation and achievement (Miller,
Brickman, & Bolen, 1975; Walton & Cohen, 2007; Wilson & Linville, 1982). A first
step toward examining whether the effects of this manipulation lead to long-lasting
motivation is to look at whether the effects last for a longer period of time than the 15-
minute duration of the research session. As mentioned previously, the effects may
become even stronger over time, as children experience the positive consequences of
working hard at challenging tasks.
STUDY 4: LONGITUDINAL EFFECTS ON CHALLENGE SEEKING
Can this manipulation (hearing a story about themselves taking on a challenge,
trying different strategies, being praised for effort, and finally succeeding) change the
way children approach challenges not just in the immediate aftermath of the
42
storybook? Could children be more likely to choose challenging tasks and persist,
even up to several weeks later? Furthermore, would the effect be more likely to last if
children were given a “booster” reminder of the manipulation? Study 4 explored these
questions.
This study also offered a methodological improvement on previous studies:
because some measures were taken long after the manipulation, it may help rule out
social desirability as an alternate explanation for the results of Studies 1-3. Children
participated at multiple time points: Time 1 involved the storybook and measures of
challenge seeking, Time 2 involved a “booster” reminder for some children of the
story, and Time 3 involved measures of challenge seeking several weeks later.
Method
Participants
Participants were 4- and 5-year-old preschoolers at a research nursery school.
Forty children participated throughout all three time points (mean age = 4 years, 6
months; range = 4 years, 0 months to 5 years, 3 months; 15 male, 25 female).
Twenty-one were Caucasian, seven were Latino/a, six were African American, four
were Asian American, and two were multiple ethnicities.
Due to absences or disinterest in going to the “Game Room” (the research
room), one child was run only at Time 1 and an additional child was run only at Times
1 and 2.
Materials
The materials were the same as in Study 3. Children were randomly assigned
to the treatment or control condition. Within each condition, half the children were
randomly assigned to receive a “booster” reminder of the storybook.
Procedure
Time 1
The procedure during Time 1 was very similar to the procedure in Study 3.
Children 1) heard a story (either the treatment or control story containing the
participating child as the main character); 2) answered four challenge-seeking
43
questions; 3) worked on a highly challenging puzzle; 4) answered four more
challenge-seeking questions; and 5) were debriefed.
Time 2
To see whether the effects lasted longer if children were given a reminder of
the intervention during a “booster” session, half the children from each condition were
randomly assigned to return to the Game Room approximately a week after Time 1
with the same researcher. They were shown the storybook again. Rather than reading
the book word for word, the researcher reminded children what happened in the story,
and talked them through it, e.g., “Remember how you got the easy puzzle and the hard
puzzle? You had a choice. Which one did you pick? Why did you pick the hard one?
When you try hard things, you learn more, right?” or “Remember how you got the
green ball and the red ball? You had a choice. Which one did you pick? Why did
you pick the green one? When you played with the green ball, you had fun, right?”
Children’s responses to the questions at Time 2 were coded on a scale from 0-2 (0 =
clearly incorrect, 1 = other, 2 = clearly correct) to assess how well they processed the
story.
If children were absent or chose not to go to the “Game Room” that day, the
researcher continued to invite them on the next available school day(s).
Time 3
Approximately two weeks after children participated at Time 1, all children
returned to the “Game Room” with the same researcher as before. The procedure was
the same as Time 1, except that children did not hear a storybook or any reminder of
the storybook that they heard. A new (and similarly difficult puzzle) was used for the
challenging experience. If children were absent or chose not to go to the Game Room
that day, the researcher continued to invite them on the next available school day(s).
Results
Overall, the results suggest that the effect of condition on challenge seeking
became stronger over time, as well as better able to withstand the experience of an
extreme challenge. (See Table 2 for a summary of overall results.) In addition,
challenge seeking showed high stability over time.
44
Time 1
Challenge Seeking
The eight challenge-seeking items formed a reliable scale, α = .84 (and none of
the items increased the reliability of the scale if deleted). For the results of children’s
explanations for their choices, see Table 5.
Effect of condition. Overall, there was a trend for children in the treatment
condition to choose more hard options than children in the control condition (control:
M = 2.50, SD = 0.48; treatment: M = 3.80, SD = 0.66), t(38) = -1.59, p = .12. (See
Figure 10.)
There was a significant interaction with age—among older children (4 years, 6
months and older), children in the treatment condition showed more challenge seeking
than children in the control condition (younger control: M = 3.75, SD = 2.38; older
control: M = 1.67, SD = 1.61; younger treatment: M = 2.67, SD = 2.69; older
treatment: M = 4.73, SD = 2.94), F(1,36) = 7.10, p = .011. As in Study 1, the
treatment storybook was most effective for older children.
Regression model. The results of the regression model supported the previous
analyses. The initial model for overall challenge-seeking choices included fixed
effects for condition, age (older or younger than 4 years, 6 months), item type (puzzles
vs. activities),9 and item time (pre- or post-challenge), and random effects of item and
participant. There was a significant interaction between condition and age, z = 2.63, p
= .009. Older children in the treatment condition were most likely to choose
challenging tasks (see above).
Time 2
Children participated at Time 2 (the “booster”) an average of eight days after
participating at Time 1, and this did not differ by condition, p = .42.
Children in the treatment and control conditions seemed to remember and
respond to questions about the books equally well (average scores: control: M = 1.60
out of 2, SD = 0.38; treatment: M = 1.67, SD = 0.47), t(18) = -0.38, p > .70.
9 On average, children chose more challenging puzzles than activities, z = 3.80, p < .001.
45
Time 3
Children participated at Time 3 an average of 17 days after participating at
Time 1, and this did not differ by condition, p = .54.
Challenge Seeking
The eight challenge-seeking items formed a reliable scale, α = .89 (and none of
the items increased the reliability of the scale if deleted). There was a significant
correlation between challenge seeking at Time 1 and Time 3, r(40) = .82, p < .001.
The effect of condition became stronger over time, particularly for older
children in the treatment condition, and for younger children in the treatment condition
who received a booster at Time 2. Furthermore, the effect of condition on challenge
seeking was significant both pre- and post-challenge experience.
Effect of condition. Considering the four groups separately (treatment/booster,
treatment/no booster, control/booster, control/no booster), there was an overall
marginal effect of condition on hard choices, F(3,36) = 2.56, p = .07. (See Figure 11.)
A planned post-hoc contrast between the treatment and control conditions was
significant, t(36) = 2.32, p = .026, indicating that children in the treatment condition
were significantly more likely to choose challenging tasks overall. There was no
difference between the two control conditions, p = .75. There was a trend for children
in the treatment/booster condition to choose more challenging tasks than children in
the treatment/no booster condition, t(36) = 1.45, p = .15. Comparing this effect to the
effect at Time 1, there was a marginal interaction between condition and time point,
indicating that the effect of condition was marginally larger at Time 3, F(3,36) = 2.41,
p = .08.
Pre- and post-challenge. When challenge-seeking choices were divided into
pre- and post-challenge items, the effect of condition was significant for items
assessed pre-challenge, and remained significant even for items assessed post-
challenge. (See Table 4.) Children in the treatment conditions were significantly
more likely to choose challenging tasks before they experienced a challenge, t(38) = -
2.17, p = .036. They were also significantly more likely to choose challenging tasks
after they experienced a challenge, t(38) = -2.23, p = .032. The interaction between
46
assessment (pre- or post-challenge) and condition (treatment or control) was not
significant, indicating that the size of the condition effect did not change, F(1,38) =
0.12, p = .73. Thus, two weeks after hearing the storybooks, children who heard the
treatment books remained challenge seeking even after working on a very challenging
task.
Regression model. The regression model supported the previous analyses,
indicating a significant effect of treatment condition. The initial model for challenge-
seeking choices included fixed effects of the child’s choice for that item at Time 1,10
condition (treatment vs. control), booster status,11 age (older or younger than 4 years, 6
months), item type (puzzles vs. activities), item time (pre- or post-challenge), and
random effects of participant and item.
The main effect of treatment was a trend, z = 1.54, p = .12, and there was a
significant interaction between treatment and age, z = 2.57, p = .010. Older children in
the treatment condition were most likely to choose the challenging option overall.
Interestingly, there was a marginally significant three-way interaction between
treatment, booster, and age, z = -1.89, p = .059. Older children in the treatment
condition showed more challenge seeking regardless of the booster, but younger
children in the treatment condition showed more challenge seeking only if they
received the booster (see Figure 12). For older children, hearing the intervention
storybook once may be enough to produce lasting change. However, younger children
may need more explicit cues and reminders of the value of choosing challenging tasks
in order to show long-term effects.
Discussion
The results of Study 4 suggest that even brief manipulations in a laboratory
have the potential to create changes for young children that last beyond the laboratory
session, with effects that may become even stronger over time. The effect size of 10 Children who chose the challenging option for a particular item at Time 1 were also more likely to choose the challenging option for that item at Time 3, z = 3.49, p < .001. 11 Across both conditions, children who received a booster were more likely to choose the challenging option, z = 2.19, p = .028 (but see below for more details about the combined effect of treatment and booster).
47
condition (treatment vs. control) on challenge seeking at Time 1 was 0.52, but at Time
3, it was 0.77. Thus, this manipulation has the potential to affect children’s choices
and persistence in the classroom over longer periods of time.
Furthermore, the ability of the intervention to create resilient motivation, the
kind of challenge seeking that can withstand an immediate challenging experience,
increased over time. At Time 1, there was not a significant difference between
conditions post-challenge, but there was a significant different post-challenge two
weeks later. Why did the effect get stronger? And why did the effect remain robust at
Time 3 in Study 4, but not in Study 3? These findings illustrate both the fragility of
this effect (post-challenge in Study 3) and its robustness (Studies 1, 2, and 4 and pre-
challenge in Study 3). Perhaps, following Time 1, children attempted a variety of
challenging tasks, succeeded at many of them, and paid greater attention to the
challenges they experienced. This then gave them a new way of perceiving and
construing those challenges (for similar effects of time in a very different population
and for a very different psychological process, see Marigold, Holmes, & Ross, 2007).
This could create a recursive cycle in which children begin to take on more challenges
and see the benefits of those challenging experiences. Similarly, children may also
have been able to weight that challenging experience (at Time 3) in light of all the
other challenges that they had succeeded at in the past two weeks. Thus, that
particular experience with difficulty may not have been as salient as it was in Study 3.
Previous studies suggest that identity- and motivation-related interventions
frequently have effects that strengthen over time, as they tap into the recursive cycles
of motivation and learning and achievement (Cohen et al., 2009). Children may go
away, think about these ideas, put them into practice, and fully embrace them. One of
the most important directions for future research may be to better understand these
recursive cycles and their power at different stages of development. Can children who
experience successful motivational interventions in early childhood better withstand
the decline in motivation that is typically seen in early adolescence?
In addition, this study suggests that this aspect of children’s motivation shows
a great deal of stability over time. Challenge seeking showed high test-retest
48
reliability between Time 1 and Time 3. This suggests that children by this age may
already be on particular motivational trajectories, yet those trajectories can be shifted
by interventions that target relevant psychological processes.
Study 4 was only the first step in longitudinally examining how long these
effects may last and how much motivation children may show in a subsequent
situation. The situations in Time 1 and Time 3 were extremely similar, in terms of an
identical researcher asking identical questions. Further research, beyond the scope of
this dissertation, should examine a new situation with a new adult in addition to a
different time point. Furthermore, another possibility for increasing real-world
validity might be to get teachers’ ratings of children’s academic motivation pre- and
post-intervention.
Despite these limitations, anecdotal evidence suggests that this manipulation
has the potential to affect children’s lives outside the laboratory and inside the
classroom. As I was conducting this study, I was in the classroom preparing to bring a
child to the research “Game Room” for Time 3. He was in the treatment condition and
had received the booster reminder the week before. We were planning to go to the
Game Room as soon as he finished the project he was working on. Another child
nearby was trying to get a basketball through the basketball hoop. She wasn’t
succeeding, so she decided to quit. He called out to her, “You have to keep trying!”
Then he turned to me and said, “Remember the book? If you keep trying the same
thing, you learn that thing and if you keep trying the same thing over and over, you get
better at that same thing. Then you can do whatever you want better and better.”
While this is anecdotal evidence, it suggests that some children were able to retain the
message of the storybook and apply it to their lives in the classroom. Moreover, some
children may even be able to communicate that message to other children. Given the
power of social norms to affect behavior, this raises the possibility that a future
intervention that could change classroom norms related to challenge seeking could be
extremely powerful. Could we create a classroom where the group identity is one that
embraces challenge (see Master & Walton, 2011; Woodhead, 1988)?
49
While the results of Studies 1-4 suggest that children’s motivation can indeed
be increased through the use of storybooks, the storybooks included several different
types of factors predicted to increase motivation in children, including praise for
effort, strategies for persisting (such as, “Just keep trying!”), and role models who
coped with failure. Thus, these studies offer no way of isolating which aspects of the
storybook might have been most critical in creating the effects on challenge-seeking
and persistence. As a beginning, Study 5 examined one aspect of the story that may
have been particularly important: whether or not children were praised for effort. In
all of the treatment books in Studies 1-4, but none of the control books, children heard
parents and peers praise the main character for trying. Could the fact that children
heard themselves (or another character) praised for effort be at least partially
responsible for the effects in Studies 1-4?
STUDY 5: EFFECTS OF PRAISE ON MOTIVATION
Based on previous research, one element of the storybook in particular may
have been particularly effective in increasing children’s motivation. In all the
treatment condition stories, after struggling with a challenging puzzle, the main
character received praise and positive feedback from an adult and peers for persisting.
Previous research with young children shows that process praise (also referred to as
effort praise) is beneficial for young children’s motivation (Kamins & Dweck, 1999;
Zentall & Morris, 2010). As described in the Introduction, in a study by Kamins and
Dweck (1999), children role-played their reactions to an unsuccessful activity, and
those who had been given process praise were less likely than those given person
praise to show helpless behaviors (such as negative self-evaluations and a lack of
persistence—i.e., choosing to give up and do something else rather than continue).
They were also more likely to construct a solution to the role-played problem (e.g., “I
can do it again better if I take my time”). Similar effects were seen when preschool
children were given generic person praise (“You are a good drawer”) or non-generic
process praise (“You did a good job drawing”) (Cimpian, Arce, Markman, & Dweck,
2007; Zentall & Morris, 2010). Children given process praise were more likely to
50
show mastery-oriented behaviors, that is, to evaluate themselves positively and report
that they would persist following a failure scenario.
However, those studies differ from the current studies in several respects.
First, in those studies, the praise was administered during a role-playing scenario,
while the current studies used a storybook in which the child received praise. Second,
those studies used only self-report measures for persistence, in addition to the self-
evaluation measures. In contrast, the current research measured behavioral persistence
on a challenging task. Finally, those studies did not directly measure challenge
seeking. Will process praise administered through a storybook (without the example
of a role model who copes with failure or who has strategies for persisting) increase
challenge seeking and behavioral persistence, in addition to creating more positive
self-evaluations after an imagined failure?
In addition, on an exploratory basis, this study examined how children
responded to a combination of person- and process-praise: praise for being a “hard
worker.” Such praise might have positive effects on children’s motivation due to the
emphasis on hard work, and help children attribute their success to consistent hard
work and effort. Indeed, previous work has found that allowing children to take on a
noun identity (such as being a “helper”) can motivate preschool children’s prosocial
behavior (Bryan & Master, 2011). However, it is also possible that such praise might
have negative effects on children’s motivation due to the emphasis on an aspect of
personality that they do not perceive control over. Will children be more sensitive to
the evaluative “person” aspect of this type of praise, or more sensitive to the
“process”-based emphasis on hard work?
In addition to challenge seeking and persistence, children’s self-evaluations
following failure were also measured to give a more complete picture of children’s
helpless or mastery-oriented behaviors (as another index of children’s responding to
difficulty), and to allow a more direct comparison of the results to previous studies in
the literature. This comparison was important given that the current study changed the
method of delivering praise from role-played scenarios to storybooks. Study 5 also
included a neutral control condition, in which children were not praised.
51
Method
Participants
Participants were 51 4- and 5-year-old children at a research nursery school
(25 were male; 26 were female). Twenty-six were Caucasian, 14 were Asian
American, five were African American, three were Latina, and three were Middle
Eastern. As in Study 3, only children older than 4 years, 6 months participated (mean
age = 4 years, 10 months; range = 4 years, 6 months to 5 years, 6 months).
Materials
Children were randomly assigned to one of four conditions. The storybooks
were simplified so that the element of praise was highly salient for children, and there
was no mention of struggle. All children were read a storybook in which they were
the main character and completed three puzzles (see Appendix B). After each puzzle,
children in the control condition were told, “You got all the pieces in the puzzle!”
Children in the person praise condition were told, “You got all the pieces in the
puzzle! You must be really good at puzzles.” Children in the person/process praise
condition were told, “You got all the pieces in the puzzle! You must be a really hard
worker at puzzles.” Children in the process praise condition were told, “You got all
the pieces in the puzzle! You must have worked really hard at puzzles.” Thus, all
children heard a book in which they successfully completed several puzzles and
received positive feedback, but different conditions attributed the child’s success to
different factors.
Procedure
Introduction
Once children entered the laboratory room, the researcher told them that they
would play with toys and act out stories. The researcher spread out a row of small
diverse-looking dolls, and asked the child to select a doll to represent him or her. The
researcher then brought out another doll named “Teacher Debbie.” To get used to
interacting with the dolls, the “Teacher Debbie” doll asked the child’s doll a series of
questions about his/her favorite color and flavor of ice cream and whether s/he had
siblings.
52
Next, the researcher read children the story in which they successfully
completed several puzzles and received a particular type of praise. After that, the
researcher brought the dolls back out and had the child role-play two situations in
which Teacher Debbie asked the child to complete a puzzle and then criticized the
puzzle. For example, in the second scenario the researcher said, “Let’s see what
happens in the next story. Here’s another puzzle piece. Another day you were sitting
at the puzzle table. After a little while Teacher Debbie says, ‘(Child’s name), will you
do this horse puzzle?’ and you say, ‘Okay, teacher.’ And so you start doing the
puzzle. You start putting pieces in, but you aren’t sure where they go. You look at
what you did and think to yourself, ‘Uh-oh, those pieces are in the wrong place,’ but
you want to show the teacher the puzzle you did so you say, ‘Teacher, I did a puzzle
for you.’ The teacher looks at the puzzle you did and says, ‘THAT doesn’t look like a
horse. The pieces are in the wrong place and you didn’t finish it.’”
Note that there were two primary advantages to using a role-played scenario to
measure children’s reactions to failure: 1) This controlled for individual differences in
children’s actual ability to complete puzzles; and 2) The role-played scenario
prevented the researcher from criticizing the child’s actual performance (Burhans &
Dweck, 1995).
Self-Evaluations
A new measure (not assessed in Studies 1-4) involved children’s self-
evaluations following this failure experience. As in Kamins and Dweck (1999) and
Cimpian et al. (2007), children responded to four questions measuring their self-
evaluations following criticism to assess an important aspect of helpless vs. mastery-
oriented responses. First, they rated how much they liked the puzzle they had made in
the last scenario on a 6-point scale (“product evaluation”). Second, they rated how
happy they felt on a 6-point scale, after not getting the pieces in the second scenario
(“affect”). Third, they were asked whether not getting the pieces made them feel like
they were good or not good at puzzles (“assessment of goodness”). Fourth, they were
asked whether not getting the pieces made them feel like they were a good girl/boy or
not a good girl/boy (“assessment of goodness”). Measuring children’s self-evaluations
53
provided another index of children’s motivation following difficulty and allowed more
direct comparison of the results to previous studies in the literature.
Challenge Seeking
Next, children responded to four of the challenge-seeking questions used in
Studies 1-4 (easy vs. hard puzzle; few vs. lots of pieces; easy vs. hard tower of blocks;
and easy vs. hard game) and gave explanations for their choices.
Persistence
Children were given a puzzle and told they could switch to a visibly easier
puzzle at any time. They worked for up to five minutes, and were reminded every 90
seconds that they could switch to the other puzzle if they wanted.
Results
Overall, children in the process praise condition showed higher motivation
than children in the other conditions. More specifically, they showed higher
motivation than children in the person praise condition in terms of self-evaluations,
and higher motivation than the control condition in terms of challenge seeking and
persistence.
Self-Evaluations
Correlations
The two assessment of goodness questions (good vs. not good at puzzles, and
good vs. not good child) were significantly correlated, r(51) = .34, p = .015. The
affect question (“how happy did you feel?”) was marginally correlated with a
composite of the two assessment of goodness questions, r(50) = .26, p = .072. The
product evaluation question (“how much did you like the puzzle you made?”) was not
significantly correlated with the other self-evaluation questions.
Effect of Condition
There was a significant effect of condition for the self-evaluation composite.
Children in the person and person/process praise conditions showed particularly low
self-evaluations. Post-hoc comparisons compared the process praise condition to the
other conditions as a group and individually.
54
Self-evaluation composite. The four items were standardized and averaged to
form a self-evaluation composite. Overall, there was a significant effect of condition,
F(3,46) = 2.78, p = .05. (See Figure 13.) Post-hoc contrasts indicated a trend for
children in the process praise condition to report higher self-evaluations than the other
three conditions combined, t(46) = 1.54, p = .13, and indicated significantly higher
self-evaluations compared to the person and person/process conditions combined,
t(46) = 2.16, p = .036. Recall that the self-evaluation measures allowed a comparison
to previous studies to ensure that the change in the method of giving praise (from role-
played scenario to storybooks) did not change the pattern of results. Indeed,
replicating previous findings, children in the process praise condition reported
significantly higher self-evaluations than children in the person praise condition, in
particular, t(46) = 2.36, p = .022.12
Challenge Seeking
The four challenge-seeking items formed a less reliable scale (compared to the
full eight item scale in Studies 1-4), α = .67. For the results of children’s explanations
for their choices, see Table 5. Overall, the results indicate that process praise
increased children’s challenge seeking compared to other types of praise or no praise.
Effect of Condition
There was an overall trend in the effect of condition on challenge-seeking
choices, F(3,47) = 2.16, p = .11. (See Figure 14.) Post-hoc contrasts revealed that
children in the process praise condition were marginally more likely to choose
challenges than children in the other three conditions, t(47) = 1.98, p = .054, and
significantly more likely to choose challenges than children in the control condition,
t(47) = 2.52, p = .015. Consistent with the previous studies reported in this
dissertation, children who heard a story in which they were praised for their effort
showed increased challenge seeking. While the effects in previous studies could be
due to other factors as well (e.g., strategies for persisting or role models who cope
12 Children in the control condition also showed significantly higher self-evaluations compared to children in the person praise condition, t(46) = -2.47, p = .017.
55
with failure), these results indicate that praise for effort without those other factors can
increase children’s challenge seeking.
Age
There was a significant effect of age on challenge-seeking choices, F(1,43) =
9.42, p = .004, and a marginal interaction of age with condition, F(3,43) = 2.43, p =
.079. After controlling for age, condition became marginally significant, F(3,43) =
2.65, p = .061, instead of showing a weaker trend toward significance. Younger
children (less than 4 years, 10 months) in the control condition were lowest in
challenge-seeking choices (M = 0.56, SD = 0.53), while older children in the process
praise condition were highest (M = 3.80, SD = 0.45). Thus, older preschool children
who heard a story involving process praise showed a greater increase in challenge
seeking (similar to the age-related findings in Studies 1 and 3, although those studies
involved storybooks with other motivational factors as well).
Regression Model
The regression model supported the previous analyses, indicating a main effect
of process praise (compared to the control condition) and an interaction between
condition and age. The initial model included fixed effects of item type (puzzle vs.
activities),13 condition (dummy-coded to contrast each condition with the control
condition), and age (older or younger than 4 years, 10 months),14 and random effects
of participant and item.
There was a significant effect of the process praise contrast, z = 2.68, p = .007
(process praise vs. control). Children in the process praise condition were more likely
to choose challenging tasks than children in the control condition. There were
marginal interactions between age and the process praise contrast, z = 1.79, p = .073,
and between age and the person praise contrast, z = -1.72, p = .086. As in the previous
analysis, older children in the process praise condition were the most challenge
13 Children were more likely to choose challenging puzzles than challenging activities, z = 2.13, p = .034. 14 Overall, older children were more likely to choose challenging tasks, z = 3.25, p = .001.
56
seeking, while younger children in the control condition were the least challenge
seeking.
Persistence
The main effect of condition on whether children persisted on the puzzle was
not significant (control: 21% of children persisted; person praise: 42%; person/process
praise: 23%; process praise: 50%), χ2 (1, N = 51) = 3.38, p = .34. However, there was
a trend for children in the process praise condition to be more likely than children in
the control condition to persist, χ2(1, N = 26) = 2.33, p = .13. (See Table 6 for further
persistence analyses.)
Motivation Composite
Correlation between Challenge Seeking and Persistence
Persistence was significantly positively correlated with the number of hard
choices, r(51) = .47, p < .001.
Effect of Condition
As in Studies 1 and 2, a motivational composite of challenge seeking and
persistence was examined, and revealed higher motivation in the process praise
condition than the control condition. There was a trend indicating an effect of
condition on the overall motivation composite of challenge seeking and persistence,
F(3,47) = 1.95, p = .13. (See Figure 15.) Children in the process praise condition
showed marginally higher motivation on the composite measure than children in the
other three conditions, t(47) = 1.90, p = .06, and significantly higher motivation than
children in the control condition, t(47) = 2.29, p = .027. Thus, children who heard a
story about themselves that involved effort praise showed greater overall motivation in
terms of challenge seeking and persistence combined than did children who heard a
story not involving effort praise (similar to the motivation composite results in Studies
1 and 2, although those studies involved storybooks with other motivational factors as
well).
Mastery-Orientation Composite
I also created a broader mastery-oriented motivation composite that included
self-evaluation in addition to challenge seeking and persistence, by standardizing and
57
averaging the three measures. There was a slight trend for an overall condition
difference, F(3,46) = 1.93, p = .14, and post-hoc contrasts revealed that children in the
process praise condition showed significantly higher mastery orientation overall
compared to the other three conditions, t(46) = 2.30, p = .026. When compared to
each other condition individually, children in the process praise condition showed
significantly higher mastery orientation compared to the control condition, t(46) =
2.24, p = .030, marginally higher compared to the person/process praise condition,
t(46) = 1.97, p = .055, and a trend toward higher mastery orientation compared to the
person praise condition, t(46) = 1.56, p = .13. (See Figure 16.) Overall, children who
heard a story about themselves being praised for effort showed more mastery-oriented
behaviors, in terms of more positive reactions to failure and greater challenge seeking
and persistence, than children who received other types of praise or no praise.
Discussion
These results suggest that the praise element of the storybooks in Studies 1-4
was at least partially responsible for the effects on challenge seeking and persistence.
Praising children for their process and effort through stories led to greater challenge
seeking and a trend toward greater persistence.
The conditions showed different patterns of effects for different measures. In
particular, the pattern of results for self-evaluations differed from that of challenge
seeking and persistence. Children in the process praise and control conditions showed
the highest motivation in terms of positive self-evaluations, while children in the
process praise condition showed higher motivation than the control condition in
challenge seeking and persistence. The first set of results (self-evaluations) may be
because person-related praise makes children particularly vulnerable to evaluating
themselves negatively following a failure experience. They may conclude that if they
are good at puzzles when they succeed, then they must not be good at puzzles when
they fail. In terms of the second set of results (challenge seeking and persistence), the
finding that person praise did not lead to less challenge seeking and persistence is
surprising. However, previous studies with young children have found that person
praise in young children can sometimes lead to low self-evaluations following failure
58
yet relatively high persistence (Zentall & Morris, 2010). Overall, when all aspects of
motivation and mastery orientation were considered, children who had been given
process praise (“You must have worked really hard”) showed the highest total
motivation.
In contrast to the effects of process praise, children in the person/process
condition (“hard worker”) frequently showed effects intermediate in terms of
motivation. These children did not differ in negative self-evaluations compared to
children in the person praise condition, and did not differ in challenge seeking and
persistence compared to the control condition. These results suggest that this type of
praise may carry some of the negative effects of person praise without all of the
benefits of process praise. Any type of praise that evaluates children for personal
qualities, even qualities that place value on effort, may hinder children when they
encounter challenges (see also Zentall & Morris, 2010).
Although the current study added additional conditions, these results replicate
previous findings in the literature and in this dissertation. First, children in the person
praise condition were more negatively affected by the role-played failure scenario in
terms of their self-evaluations, compared to children in the process praise condition
(Cimpian et al., 2007; Kamins & Dweck, 1999; Zentall & Morris, 2010). As in
previous research, it seems that children are extremely vulnerable to the sense that
they are being evaluated based on their traits or ability. Rather than giving them
enduring confidence, praising children for ability makes them more likely to doubt
their ability when they encounter difficulty.
Second, children in the process praise condition showed higher levels of
challenge seeking and tended to show higher levels of persistence, compared to
children in the control condition. This supports the findings of other studies in this
dissertation, and suggests that children who receive positive feedback from adults
about taking on challenges and not giving up show motivational benefits compared to
children in a control condition. Stories in which the participating child is praised can
send the message that effort is valued and important. Although the other studies in
59
this dissertation contained other elements besides praise, the praise element of the
story seems to be at least partially responsible for the effects in those studies.
However, this does not rule out the possibility that the results of Studies 1-4
were also due to other elements in the story. For example, hearing about a character
who struggled to cope with failure may also have contributed to greater challenge
seeking in the previous studies. Indeed, the results of a pilot study (N = 17) indicated
that older preschool children who heard about a character who struggled before
learning how to do a task showed marginally more challenge seeking (p = .098) than
children who heard about a character who was immediately successful at a task. It
would be very interesting for future research to vary both effort praise (vs. no effort
praise) and struggle (vs. immediate success) within the same study to more directly
compare the effects. Thinking about the process of learning as involving struggle may
be critical for children, such that thinking about success (without this element of the
learning process) may not instill the same motivation to persevere.
Overall, these results indicate that process praise has beneficial consequences
for young children’s challenge seeking and persistence, even following situations in
which children do not succeed.
GENERAL DISCUSSION
Challenge seeking and persistence are aspects of motivation that are
particularly critical for students’ learning. The studies reported here lay the
foundation for a better understanding of the factors that affect challenge seeking and
persistence in young children, with the ultimate goal of using that understanding to
create educational interventions that set children on more successful academic
trajectories.
Discussion of Findings
The results from Study 1 suggest that challenge seeking and persistence can be
increased in children using storybooks that include praise for effort, strategies for
persisting, and a role model who copes with failure. While these books increased
children’s challenge seeking regardless of who the main character was, the books were
most effective for preschoolers’ behavioral persistence when the main character was
60
the child himself or herself. Thus, personal identity seemed to matter more for
preschool children’s persistence than for challenge seeking.
In contrast, in Study 2, the books were equally effective in increasing
persistence for kindergarten students whether the main character was the participating
child, another child, or an animal. Moreover, the books were highly effective in
increasing persistence in a group of academically struggling students. Furthermore,
both Studies 1 and 2 also found that children who endorsed global beliefs about
badness (that making a mistake implies global badness) were less likely to take on
challenging tasks.
The results from Study 3 showed that the treatment book was equally effective
regardless of whether children explained the value of challenges or attempted to
persuade someone else of the value of challenges. Study 3 also found that the effect of
the treatment book on challenge seeking was weakened after children experienced a
real challenge.
The results from Study 4 are particularly exciting, as these results suggest that
these effects may become stronger over time, and that children may become more
likely to show robust challenge seeking even after experiencing real challenges. Older
preschool children showed these long-term effects whether or not they were reminded
of the value of taking on challenges, whereas younger children remained challenge
seeking only if they had been previously reminded of the story.
The results from Study 5 indicate that praise may be a critical component of
this process. Children who heard a story about themselves being praised for effort
showed more positive self-evaluations following failure, compared to children who
heard a story about themselves praised for person-based traits such as ability.
Children who heard a story in which they were praised for effort also showed greater
challenge seeking and slightly greater persistence than children who heard a control
story in which they were not praised for effort.
In Studies 1, 3, 4, and 5, the treatment books showed reliable effects on
challenge seeking (see Table 2). This suggests that children are sensitive to the
messages they hear in stories, and that stories can be used to teach children about the
61
value of taking on challenges. Studies 1, 2, and 5 also suggested that these books
could be use to increase young children’s persistence at challenging tasks. Hearing a
story about themselves persisting (“I’ll just keep trying, and I’ll see what happens”) in
which they are praised for persistence and working hard may be an important way to
encourage young children to persist when they encounter challenges.
Challenge Seeking and Persistence
Another interesting finding across studies was the relationship between
challenge seeking and persistence. Although challenge seeking and persistence were
not correlated in Study 1, there was a moderately high correlation in Studies 2 and 5
between the number of challenging items that children chose and their persistence on a
challenging puzzle. When data from the three studies were combined, the overall
correlation between challenge seeking and persistence was significant, r(169) = .23, p
= .002. This suggests that these may be related components of motivation. Indeed, in
the real world, children’s behavior in both types of situations may create a feedback
loop, in which children choose a challenging task, persist at it, and realize the benefits
of having persisted, which then increases their continuing preference for challenging
tasks that they can learn from. Children who show motivational deficits may also
create feedback loops: children who choose easy tasks may enjoy the sensation of
easily completing tasks and dislike the sensation of frustration while trying to persist
on challenging tasks, which may make them more likely to continue to choose easy
options in the future. In this way, children with different motivational tendencies may
place themselves on very different academic trajectories in school.
However, there could be several reasons for the overall correlation between
challenge seeking and persistence. The manipulation incorporated several different
motivational factors, so different aspects of the manipulation may have had separate
effects on challenge seeking and persistence. Furthermore, challenge seeking and
persistence did not always show similar patterns of effects. In Study 1, all the
treatment books increased preschoolers’ challenge seeking, but only the book about
the participating child increased persistence. In contrast, in Study 2, the treatment
books led to significant increases in kindergarteners’ persistence but not challenge
62
seeking. Perhaps challenge seeking may be more strongly related to goal-related
processes (e.g., whether children are more concerned with learning or performing
well), while persistence may be more strongly tied to self-regulation processes. Few
studies have measured both challenge seeking and persistence in children, and many
of those previous studies used self-report measures of persistence rather than
behavioral persistence at a task. In line with the current findings, some studies also
demonstrate that challenge seeking and persistence do not always correlate or show
identical effects. For example, Nichols, Whelan, and Meyer (1991) found that 5th
grade students who had been given learning goals chose harder tasks following failure
feedback compared to those who had been given performance goals, but there was no
difference in their persistence (in terms of the number of additional problems they
wanted to work on). In terms of persistence, recent work found that preschoolers’
persistence on some tasks was correlated with motivational measures such as valuing a
task, and on other tasks was correlated with self-regulation measures such as working
Marum, E. (1996). Reading more than print. In E. Marum (Ed.), Children and books in
the modern world: Contemporary perspectives on literacy (pp. 141-145).
London: The Falmer Press.
Master, A., & Walton, G. M. (2011). Perceived group membership increases task
motivation and learning in young children. Manuscript in preparation,
Stanford University.
McArthur, L. Z., & Eisen, S.V. (1976). Achievement of male and female storybook
characters as determinants of achievement behavior by boys and girls. Journal
of Personality and Social Psychology, 33, 467-473.
McClelland, D. C. (1961). The achieving society. Princeton, NJ: D. Van Nostrand
Company, Inc.
Meece, J. L., Blumenfield, P. C., & Hoyle, R. H. (1988). Students’ goal orientations
and cognitive engagement in classroom activities. Journal of Educational
Psychology, 80, 514-523.
Miller, R. L., Brickman, P., & Bolen, D. (1975). Attribution versus persuasion as a
means for modifying behavior. Journal of Personality and Social Psychology,
31, 430-441.
Miller, P. J., Wiley, A. R., Fung, H., & Liang, C.-H. (1997). Personal storytelling as a
medium of socialization in Chinese and American families. Child
Development, 68, 557-568.
Mischel, W., & Patterson, C. J. (1976). Substantive and structural elements of
effective plans for self-control. Journal of Personality and Social Psychology,
34, 942-950.
Morgan, G. A., Harmon, R. J., & Maslin-Cole, C. A. (1990). Mastery motivation:
Definition and measurement. Early Education and Development, 1, 318-339.
Mueller, C. M., & Dweck, C. S. (1998). Intelligence praise can undermine motivation
and performance. Journal of Personality and Social Psychology, 75, 33-52.
87
Nicholls, J. G. (1978). The development of the concepts of effort and ability,
perception of academic attainment, and the understanding that difficult tasks
require more ability. Child Development, 49, 800-814.
Nicholls, J. G. (1979). Quality and equality in intellectual development: The role of
motivation in education. American Psychologist, 34, 1071-1084.
Nicholls, J. G., Patashnick, M., & Nolen, S. B. (1985). Adolescents’ theories of
education. Journal of Educational Psychology, 77, 683-692.
Nichols, A. L., Whelan, J. P., & Meyers, A. W. (1991). The effects of children’s goal
structures and performance feedback on mood, task choice, and task
persistence. Behavior Therapy, 22, 491-503.
Patterson, M. M., & Bigler, R. S. (2006). Preschool children’s attention to
environmental messages about groups: Social categorization and the origins of
intergroup bias. Child Development, 77, 847-860.
Patterson, C. J., & Mischel, W. (1975). Plans to resist distraction. Developmental
Psychology, 11, 369-378.
Patterson, C. J., & Mischel, W. (1976). Effects of temptation-inhibiting and task-
facilitating plans on self-control. Journal of Personality and Social
Psychology, 33, 209-217.
Perry, L. K., Samuelson, L. K., Malloy, L. M., & Schiffer, R. N. (2010). Learn locally,
think globally: Exemplar variability supports higher-order generalization and
word learning. Psychological Science, 21, 1894-1902.
Phillips, M., Crouse, J., & Ralph, J. (1998). Does the black-white test score gap widen
after children enter school? In C. Jencks & M. Phillips, The black-white test
score gap (pp. 229-272). Washington, D. C.: Brookings Institution Press.
Piaget, J. (1952). The origins of intelligence in children. New York: W. W. Norton &
Company, Inc.
Piaget, J. (1977). Problems of equilibration. In M. H. Appel & L. S. Goldberg (Eds.),
Topics in cognitive development, Volume 1: Equilibration: Theory, research,
and application (pp. 3-13). New York: Plenum Press.
Pintrich, P. (2003). A motivational science perspective on the role of student
88
motivation in learning and teaching contexts. Journal of Educational
Psychology, 95, 667-686.
Piper, W. (1930). The little engine that could. New York: Platt & Munk.
Pomerantz, E. M., & Saxon, J. L. (2001). Conceptions of ability as stable and self-
evaluative processes: A longitudinal examination. Child Development, 72, 152-
173.
Pugh, K. J., & Bergin, D. A. (2006). Motivational influences on transfer. Educational
Psychologist, 41, 147-160.
Quattrone, G. A. (1985). On the congruity between internal states and action.
Psychological Bulletin, 98, 3-40.
Scarr, S., & McCartney, K. (1983). How people make their own environments: A
theory of genotype to environment effects. Child Development, 54, 424-435.
Schunk, D. H. (1981). Modeling and attributional effects on children’s achievement: A
self-efficacy analysis. Journal of Educational Psychology, 73, 93-105.
Schunk, D. H., Hanson, A. R., & Cox, P. D. (1987). Peer-model attributes and
children’s achievement behaviors. Journal of Educational Psychology, 79, 54-
61.
Sigman, M., Cohen, S. E., Beckwith, L., & Topinka, C. (1987). Task persistence in 2-
year-old preterm infants in relation to subsequent attentiveness and
intelligence. Infant Behavior and Development, 10, 295-305.
Smiley, P. A., & Dweck, C. S. (1994). Individual differences in achievement goals
among young children. Child Development, 65, 1723-1743.
Stipek, D. J. (2001). Pathways to constructive behavior: Importance of academic
achievement in early elementary grades. In A. Bohart & D. Stipek (Eds.),
Constructive and destructive behavior: Implications for family, school, and
society (pp. 291-315). Washington, DC: American Psychological Association.
Stipek, D. J. (2002). Motivation to learn: Integrating theory and practice. Boston:
Allyn & Bacon.
Stipek, D. J., & Greene, J. K. (2001). Achievement motivation in early childhood:
Cause for concern or celebration? In S. Golbeck (Ed.), Psychological
89
perspectives on early childhood education: Reframing dilemmas in research
and practice (pp. 64-91). Mahwah, NJ: Erlbaum.
Thomas, A., & Pashley, B. (1982). Effects of classroom training on LD students’ task
persistence and attributions. Learning Disability Quarterly, 5, 133-144.
Tomasello, M., Carpenter, M., Call, J., Behne, T., & Moll, H. (2005). Understanding
and sharing intentions: The origins of cultural cognition. Behavioral and Brain
Sciences, 28, 675-735.
Tsai, J. L., Louie, J. Y., Chen, E. E., & Uchida, Y. (2007). Learning what feelings to
desire: Socialization of ideal affect through children’s storybooks. Personality
and Social Psychology Bulletin, 33, 17-30.
Van Dulken, S. (2000). Inventing the 20th century: 100 inventions that shaped the
world. New York, NY: New York University Press.
Vasquez, N. A., & Buehler, R. (2007). Seeing future success: Does imagery
perspective influence achievement motivation? Personality and Social
Psychology Bulletin, 33, 1392-1405.
Walton, G. M., & Cohen, G. L. (2007). A question of belonging: Race, social fit, and
achievement. Journal of Personality and Social Psychology, 92, 82-96.
Wigfield, A., & Eccles, J. S. (2002). Students’ motivation during the middle school
years. In J. Aronson (Ed.), Improving academic achievement: Impact of
psychological factors on education (pp. 159-184). San Diego, CA: Academic
Press.
Williams, J. J., & Lombrozo, T. (2010). The role of explanation in discovery and
generalization: Evidence from category learning. Cognitive Science, 34, 776-
806.
Wilson, T. D. (1990). Self-persuasion via self-reflection. In J. M. Olson & M. P.
Zanna (Eds.), Self-inference processes: The Ontario symposium, Volume 6, pp.
43-67. Hillsdale, NJ: Lawrence Erlbaum.
Wilson, T. D., & Linville, P. W. (1982). Improving the academic performance of
college freshmen: Attribution therapy revisited. Journal of Personality and
Social Psychology, 42, 367-376.
90
Woodhead, M. (1988). When psychology informs public policy: The case of early
childhood intervention. American Psychologist, 43, 443-454.
Zentall, S. R., & Morris, B. J. (2010). “Good job, you’re so smart”: The effects of
inconsistency of praise type on young children’s motivation. Journal of
Experimental Child Psychology, 107, 155-163.
91
Table 1. Challenge seeking scale, Studies 1 and 2.
Puzzle Choices
1. This puzzle is hard and this puzzle is easy. I might give you this one or this one. Which one do you want to do?
2. This puzzle you won’t have to work so much to finish, and this puzzle you’ll have to really work a lot. I might give you this one or this one. Which one do you want to do?
3. This puzzle will take a long time to finish, and this one you can do quickly. I might give you this one or this one. Which one do you want to do?
4. This puzzle has just a few pieces and this puzzle has lots of pieces. I might give you this one or this one. Which one do you want to do?
Activity Choices
1. Now, instead of puzzles, there are two songs I could teach you. One song is hard to learn and one song is easy. Which one do you want to learn?
2. I have one picture that’s easy to color and one picture that’s hard to color. Which one do you want to color?
3. I have some blocks that make a tower that’s hard to build and some blocks that make a tower that’s easy to build. Which one would you rather build?
4. I have one game that’s easy to play and one game that’s hard to play. Which one would you rather play?
92
Table 2. Summary of results.
CHALLENGE SEEKING
PERSISTENCE MOTIVATION composite
Study 1: 3 treatment vs. control
p = .003
p = .085
p = .039
Study 2: 3 treatment vs. control
n.s.
p = .004
p = .009
Study 3: 2 treatment vs. control (controlling for age)
p = .021
Study 4: Time 1 treatment vs. control
p = .12
Time 3 treatment vs. control p = .026
Study 5: process praise vs. control
p = .015
p = .13
p = .027
93
Table 3. Helmert style contrasts for condition, Studies 1 and 2. Contrast 1 Contrast 2 Contrast 3 Control -3 0 0 Animal 1 -1 -1 Another child 1 -1 1 Participating child 1 2 0
94
Table 4. Effects of condition on puzzle and activity choices separately, and pre-challenge and post-challenge choices.
PUZZLE Choices
ACTIVITY Choices
PRE- Challenge
POST- Challenge
Study 1: 3 treatment vs. control
p = .003
p = .014
N/A
N/A
Study 2: 3 treatment vs. control
n.s.
n.s.
N/A
N/A
Study 3: 2 treatment vs. control (controlling for age)
p = .031
p = .012
p = .011
p = .077
Study 4: Time 1 treatment vs. control
p = .20
p = .11
p = .20
p = .14
Time 3 treatment vs. control p = .065 p = .023 p = .036 p = .032
Study 5: process praise vs. control
p = .012
p = .072
N/A
N/A
95
Table 5. Effects on children’s open-ended explanations for their choices.
Challenge-SEEKING Challenge-AVOIDING Study 1: 3 treatment vs. control
p = .012
p = .077
Study 2: 3 treatment vs. control
n.s.
n.s.
Study 3: 2 treatment vs. control (controlling for age)
p = .033
p = .001
Study 4: Time 1 treatment vs. control
p = .022
p = .15
Time 3 treatment vs. control p = .032 p = .018
Study 5: process praise vs. control
n.s.
n.s.
96
Table 6. Effects of condition on length of time persisted (for children who quit) and number of puzzle pieces correctly placed.
Length of Time before Quit
Correct Puzzle Pieces
Study 1: Participating child vs. other conditions
n.s.
p = .020 Study 2: 3 treatment vs. control
n.s.
p = .13 Study 5: process praise vs. control
p = .07
p = .07
97
Figure 1. Challenge seeking, Study 1. The control condition was significantly different from the other conditions. Error bars represent +1 standard error.
98
Figure 2. Children’s global badness beliefs and challenge seeking, Study 1.
99
Figure 3. Persistence, Study 1. The participating child condition was significantly different from the other conditions.
100
Figure 4. Motivation composite, Study 1. The participating child condition was significantly different from the control and animal conditions.
101
Figure 5. Children’s global beliefs and challenge seeking, Study 2.
102
Figure 6. Persistence, Study 2. The control condition was significantly different from the treatment conditions.
103
Figure 7. Motivation composite, Study 2. The control condition was significantly different from the treatment conditions.
104
Figure 8. Challenge seeking, Study 3.
105
Figure 9. Challenge seeking by age (older or younger than 4 years, 8 months), Study 3. (Note that all of the younger children in the control condition chose all easy tasks.)
106
Figure 10. Challenge seeking, Study 4, Time 1. The treatment conditions showed a trend toward being higher than the control conditions.
107
Figure 11. Challenge seeking, Study 4, Time 3. The treatment conditions were significantly different from the control conditions.
108
Figure 12. Interaction between booster and age on challenge seeking for children in the treatment condition, Study 4.
109
Figure 13. Self-evaluation composite, Study 5.
110
Figure 14. Challenge seeking, Study 5.
111
Figure 15. Motivation composite (challenge seeking and persistence), Study 5.
112
Figure 16. Mastery-orientation composite (self-evaluations, challenge seeking, and persistence), Study 5.