Why do parents decide to become involved in their children’s education? An empirical test of the Hoover-Dempsey and Sandler model Christa L. Green Vanderbilt.

Why do parents decide to become involved in their children’s

education? An empirical test of the Hoover-Dempsey and Sandler model

Christa L. GreenVanderbilt University

Joan M. T. WalkerLong Island University

Howard M. Sandler & Kathleen V. Hoover-DempseyVanderbilt University

Acknowledgements

Special thanks to the Peabody Family-School Partnership Lab, the parents and schools that made this study possible and OERI/IES

(grant #R305TO10673-03).

http://www.vanderbilt.edu/Peabody/family-school/

General Overview Part of a large scale, 3-year study based in two

major areas of theory and research. parents’ involvement in students’ education is

associated with improved achievement. social-cognitive theory and research

The full 3-year study was grounded in a theoretical model of the parental involvement process (Hoover-Dempsey & Sandler, 1995, 1997).

For more information, please see our final report (available for free from our website) and visit a session immediately following this one: 10:35am to 12:05pm, Building: Moscone Center West, Room: 2nd Floor, Room 2002: Self-Regulation and Homework Behavior: The Role of Self-Efficacy Beliefs, Homework Quality, and Parental Behavior

Specific Purpose Examine the ability of model constructs to

predict parents’ involvement choices and behaviors. Ability of constructs to predict involvement

after controlling for SES. Examine age-related differences in

predicting parental involvement from model constructs.

Background

Model Constructs

Parent’s motivational beliefs Role construction (“Do I think I’m

supposed to be involved?”) (Drummond & Stipek, 2004; Gonzalez & Chrispeels, 2004; Grolnick et al., 1997, Hoover-Dempsey & Sandler, 1995, 1997; Hoover-Dempsey et al., 2005; Scribner et al., 1999)

Efficacy (“If I’m involved, will it make a difference?”) (Bandura, 1997; Bandura et al., 1996; Eccles & Harold, 1996; Hoover-Dempsey et al., 1992, 2005; Grolnick et al., 1997; Sheldon et al., 2002; Seefeldt et al., 1998)

Model Constructs cont. Parent’s perception of invitations from others

General school invitations (“Is the school inviting? Does the school ‘tell’ me that they want my involvement?”) (Adams & Christenson, 1998; Comer, 1985; Griffith, 1998, 2001)

Specific teacher invitations (“Does the teacher ask me to be involved, make specific requests and suggestions?”) (Balli, Demo, & Wedman, 1998; Corno, 2000; Epstein & Salinas,

1993; Epstein & Van Voorhis, 2001; Hoover-Dempsey et al., 1995; Shumow, 1998). Specific child invitations (“Does my child want or

need my involvement?”)(Balli et al., 1998; Baumrind, 1991; Hoover-Dempsey et al., 1995; Xu & Corno, 1998).

Model Constructs cont.

Parent’s perceived life context Knowledge & Skills (Dauber & Epstein, 1993; Hoover-Dempsey et al., 1995;

Hoover-Dempsey et al., 2005; Kay, Fitzgerald, Paradee, & Mellencamp, 1994; Lareau, 1989). Time & Energy (Hoover-Dempsey et al., 1995; Hoover-Dempsey et al., 2005;

Lareau, 1989; Weiss, Mayer, Kreider, Baughan, Dearing, Hencke, & Pinto, 2003).

Participants

853 parents of 1st through 6th grade public school children.

Parents recruited at 2 different points Questionnaire packets sent home with and

returned by children from participating schools

DemographicsStudy 1 Study 2

Date of data collection Fall 2002 Fall 2003

Participating public schools 3 elementary2 middle

5 elementary4 middle

Parent participants Grades 1-6 Grades 4-6

Number 495 358

Mean parent education High school or equivalent Some college

Mean family income $20,000-30,000 $30,000-40,000

Race (% of total)African AmericanAsian AmericanHispanic AmericanWhiteOtherMissing Value

16.5% 5.6%24.5%30.5% 7.7%15.2%

27.4% 3.9% 6.4%57.3% 4.2%0.8%

Measures

Predictor constructs used a 6-point agree-disagree response scale: Motivational beliefs (role construction and efficacy) Perceptions of invitations to involvement from others

(general invitations from the school, specific invitations from teachers and from the child)

Perceived life context (perceived skills and knowledge, time and energy for involvement)

Outcomes used a 6-point never-to-daily response scale Commonly observed home-based and school-based

parental involvement behaviors

Scale ReliabilitiesConstruct Time/Number of Items Alpha Reliability Coefficient

Role Activity Beliefs Time 1: 7 itemsTime 2: 10 items

.67

.83

Parental Self-Efficacy for Involvement

Time 1: 7 itemsTime 2: 5 items

.78

.80

General School Invitations Time 1: 6 itemsTime 2: 6 items

.88

.79

Specific Teacher Invitations Time 1: 6 itemsTime 2: 5 items

.81

.67

Specific Child Invitations Time 1: 6 itemsTime 2: 5 items

.70

.64

Self-perceived Skills and Knowledge for Involvement

Time 1: 9 itemsTime 2: 6 items

.83

.82

Self-perceived Time and Energy for Involvement

Time 1: 8 itemsTime 2: 5 items

.84

.81

Involvement: Home-based

Involvement: School-based

Time 1: 4 itemsTime 2: 5 itemsTime 1: 6 itemsTime 2: 5 items

.70

.79

.82

.71

Analyses: Primary

Multiple hierarchical regressions were conducted using factors in the following blocks to predict parental involvement:

Block 1: Motivational beliefs (role activity, self-efficacy for helping the child succeed in school)

Block 2: Perceptions of invitations for involvement (general invitation from the school, specific teacher invitations, specific child invitations)

Block 3: Self-perceived life context (skills and knowledge, time and energy for involvement)

Predictors of Home-based Involvement

Significant portion (F = 137.07, p < .000; Adj. R2 = .39) of the variance was accounted for by: Motivational Beliefs

• Role activity ( = .05)• Self-efficacy for helping the child succeed ( = .22)

Perceptions of Invitations • Specific invitations from the child ( = .51)

Life Context• Time and energy for involvement ( = .15)

Excluded variables: General invitations for involvement from the school Specific invitations from teachers Self-perceived skills and knowledge

Predictors of School-based Involvement Significant portion of the variance was accounted for (F =

163.65, p < .000, Adj. R2 = .49) by:Motivational Beliefs

• Role activity ( = .06)• Self-efficacy for helping the child succeed in school ( = -.06)

Perceptions of Invitations• Specific invitations from teachers ( = .37) • Specific invitations from the child ( = .31)

Life Context • Time and energy for involvement ( = .19)

Excluded: General invitations for involvement from the school Skills and knowledge for involvement

Predictors of Total Involvement Regression results were significant (F = 239.39, p < .000,

Adj. R2 = .58):Motivational Beliefs

• Role activity ( = .08)

Perceptions of Invitations • Specific invitations from the child ( = .46) • Specific invitations from teachers ( = .22)

Life Context • Skills and knowledge ( = .12)• Time and energy for involvement ( = .17)

Excluded: Self-efficacy for helping the child succeed in school General invitations from the school

Considering SES

Some have found SES and parental involvement to be positively related (Lareau, 1996).

Others note that SES variables do not explain fully why parents decide to become involved effects of SES have been found to be minimal when

underlying processes are examined (e.g., Grolnick et al 1997)

Added in the first block of the multiple regressions, SES variables were not significant in predicting variance in any of the equations.

Involvement across grades

1.50

2.00

2.50

3.00

3.50

4.00

4.50

5.00

5.50

6.00

1 2 3 4 5 6

Grade level

Means

Home-based Involvement

School-basedInvolvement

Total Involvement

Comparisons between elementary and middle school

School type Number Home-based Involvement Mean (SD)

School-based Involvement

Total Involvement

Elementary (1st - 4th grade)

401 5.22 (.87) 2.50 (1.07) 3.66 (.77)

Middle (5th - 6th grade)

443 4.69 (1.02) 1.86 (.66) 3.16 (.66)

T- test (unequal variances) t (1, 787.65) = 10.07, p < .000

t (1, 656.46) = 10.30, p < .000

t (1, 839.27) =8.10, p < .000

Effect size (Cohen’s d) .70 .73 .55

Exploring age-related differences in the model

Total involvement: Elementary: Adj. R2=.51 Middle: Adj. R2=.56

Home-based involvement: Elementary: Adj. R2=.27 Middle: Adj. R2=.48

School-based involvement: Elementary: Adj. R2=.47 Middle: Adj. R2=.36

Discussion of Results: Primary

Parents’ involvement behavior can be predicted by constructs included in the model.

All involvement types were predicted by Specific child invitations Parents’ perceptions that they have the time and energy to

become involved Parents’ beliefs that they should play an active role in the

child’s education. Home-based involvement:

Also included: Parents’ beliefs that they can help their child succeed in school

School involvement: Also included: Specific teacher invitations

Discussion: Secondary interests

Model continued to predict involvement when controlling for SES

As expected, involvement decreased as the children aged.

The model accounted for different amounts of variance for different age groups.

Applications

Helpful for interventions to know which constructs to target to increase specific types of parent involvement Specific child invitations to increase home-

based involvement in middle school Specific teacher and child invitations to increase

school-based involvement for all grades Role and self-efficacy to increase home-based

involvement in elementary school

Next steps

Explore model’s ability to predict involvement across different cultural groups

Use multiple methods to determine effectiveness of model constructs

Continue to work with the upper levels of model

Thank you!

Special thanks to Kathy Hoover-Dempsey, Joan Walker, Katie Shepard, Kelly Sheehan, and Kathleen Miller for their assistance with this presentation.

http://www.vanderbilt.edu/Peabody/family-school/