First-Year Students’ Plans to Volunteer - NSSE Homensse.indiana.edu/pdf/conference_presentations/2006/Cruce_and_Moore... · The purpose of this study is thus to estimate the differences

First-Year Students’ Plans to Volunteer:

An Examination of the Predictors of Community Service Participation

Ty M. Cruce

John V. Moore, III

Indiana University

Scholarly Paper Session

Track 4: Institutional Management and Planning

2006 Annual Forum for the Association for Institutional Research

May 18, 2006

An electronic copy of this paper is located at: http://nsse.iub.edu/conferences/index.cfm

2

Abstract

The impacts of community service participation on college student development are extensive

and well-documented. The characteristics of students that predict volunteerism, however, are not

well understood. The purpose of this study is thus to estimate the differences in first-year

students’ plans to volunteer while in college (i.e., ‘Done,’ ‘Plan to do,’ ‘Do not plan to do,’ or

‘Have not decided’) by their background characteristics and by the characteristics of the

institutions that they attend. Study results suggest changes to several campus policies and

programs that may remove barriers to successful community service participation among first-

year students.

3

Educating people for responsible citizenship has been a part of the missions of colleges

since their inception in the United States. The founding of the first institutions of higher

education in this country was for the purpose of educating a new generation of civic and

religious leaders for the communities of the new world. It was within this spirit that colleges

came of age, and it is a tradition that has stayed with them up to the present, as evident in their

current mission of teaching, research, and service (Rudolph, 1977; Terenzini, 1994).

Modern ideas about the integration of service and learning in higher education were

shaped by the work of John Dewey, who proposed ideas about hands-on learning and practical

education (Curti, 1965). Dewey asserted that better learning occurs when students have the

opportunity to put into practice the ideas that they are learning in the classroom. As further

research has been done on this concept of service-learning, scholars (e.g., Astin & Sax, 1998;

Astin, Sax, & Avalos, 1999; Eyler & Giles, 1999; Eyler, Giles, & Braxton, 1997; Gray,

Ondaatje, Fricker, Geshwind, Goldman, Kaganoff, Robyn, Sundt, Vogelgesang, & Klein, 1998;

Vogelgesang & Astin, 2000) have found that not only does student volunteerism promote the

civic engagement that universities have historically tried to foster in their students, but

community service also offers a host of educational and extracurricular benefits to the students.

The impacts of volunteerism on college student development are extensive and well-

documented. These benefits can be distilled into thee broad categories – educational/scholastic,

career/vocational, and personal/social – and each of these categories represent areas of student

development that are highly valued by universities and educators. Educationally, students who

participate in community service receive better grades (e.g., Tartter, 1996), demonstrate greater

educational gains (e.g., Eyler & Giles, 1999), and increase their critical thinking skills (e.g.,

Eyler, Root, & Giles, 1998). Vocationally, community service by college students is associated

with a stronger likelihood to participate in both future community service (e.g., Astin,

4

Vogelsang, Ikeda & Yee, 2000) and service-oriented professions (e.g., Astin, Sax, & Avalos,

1999), and to have career aspirations that require a more advanced degree (e.g., Astin and Sax,

1998), thus fostering an interest in further education. Socially, service learning is associated with

stronger leadership skills (e.g., Vogelsang & Astin, 2000), a greater interest in and commitment

to diversity (e.g., Boyle-Baise & Kilbane, 2000), a stronger sense of self (e.g., Wang, 2000),

increased self confidence (e.g., Astin & Sax), and a more developed commitment to social issues

(e.g., Giles & Eyler, 1994).

Although the benefits of community service to the student volunteer are well-

documented, information from a recent survey of college students illustrates that 36% of

graduating seniors at baccalaureate degree-granting institutions never participate in community

service during their college years (National Survey of Student Engagement, 2005). This

percentage of seniors who were uninvolved in community service is as low as 25% for seniors

attending Baccalaureate – Liberal Arts institutions and as high as 40% for seniors attending

Doctoral –Intensive institutions. Only 6% of first-year students at these same institutions

indicated that they did not plan to volunteer during college, and another 13% responded that they

were undecided about volunteering during college. Although these data are not longitudinal, if

this pattern holds true over time, many students’ plans to volunteer during college are not being

realized by the time that they graduate.

Given the important benefits of community service to students’ academic and personal

development, their future career development and civic engagement, the impending question is,

which students are not participating in community service during college? If measured during the

early stages of the college experience, information on the characteristics of students who are not

likely to volunteer may assist campus policy-makers in realigning programs and resources

toward removing the barriers to community service participation and maximizing the number of

5

students who volunteer while in college. Using data from over 125,000 first-year students

attending over 620 four-year colleges and universities in the U.S., the purpose of the current

study is to examine the differences in students’ plans to volunteer during college by their

background characteristics and college experiences, and by the characteristics of the institutions

that they attend.

Literature Review

The characteristics of students who volunteer during college are not well understood.

Only a limited number of studies have examined the predictors of volunteerism during college,

and the evidence from this small body of research suggests that students are more likely to

volunteer during college if they are female (Astin & Sax, 1998; Astin, Vogelgesang, Ikeda, &

Yee, 2000; Fitch, 1991; Marks & Jones, 2004; Pierson, 2002; Serow & Dreyden, 1990); have a

higher socioeconomic status (Marks & Jones); are more religious (Astin & Sax; Astin, et al.;

Fitch; Marks & Jones; Serow & Dreyden); are less materialistic (Astin & Sax; Astin, et al.;

Marks & Jones; Serow & Dreyden); have expressed an earlier commitment to community

service (Astin & Sax; Astin, et al.); have higher college grades (Serow & Dreyden) and graduate

level degree aspirations (Marks & Jones); live on campus (Fitch) and participate in college

organizations (e.g., Greek societies, student government, and religious groups) that encourage or

require community service (Marks & Jones; Serow & Dreyden); are not employed (Fitch) or

work fewer hours per week (Marks & Jones); and attend religiously-affiliated as opposed to

public institutions (Serow & Dreyden). Students are also more likely to volunteer during college

if they volunteered prior to entering college (Astin & Sax, 1998; Astin, et al., 2000; Pierson,

2002), although some evidence suggests that the direction of the effect or prior community

service may depend on whether the students were either required (i.e., negative effect) or simply

6

encouraged (i.e., positive effect) to volunteer while in high school (Marks & Jones). Finally,

some evidence suggests that propensity to volunteer does not differ by race or ethnicity (Mark &

Jones; Pierson).

Although there is a considerable degree of consistency in the above findings, there are

several limitations to this small body of research that render this evidence suspect. The first

limitation, evident in two of the studies cited above, is the lack of generalizability of the findings.

These studies are based on convenience samples at either a single institution (i.e., Fitch, 1991) or

a limited set of institutions in a single state (i.e., Serow & Dreyden, 1990), and, accordingly, the

student samples in these studies are not necessarily representative of the college student

population. Although there are studies (Astin & Sax, 1998; Astin, et al., 2000; Marks & Jones,

2004; Pierson, 2002) that are based on more representative samples of college students across

multiple institutions, these studies do not take into consideration (or either report) the unique

differences in the student propensities to volunteer by the characteristics of the institution

attended (e.g., control, size, and location).

The second limitation of the extant literature is the type of analysis employed to estimate

student differences in the propensity to volunteer. Two of the studies cited above (Fitch, 1991;

Serow & Dreyden, 1990) rely on a series of Chi Square tests to examine differences between

students. This approach is not as powerful as regression analysis, which would allow the

researchers to statistically control for other characteristics of the students and to isolate the effect

of a single variable. Although providing an improvement over the Chi Square test, Pierson

(2002) estimates a model of volunteering – represented by a dichotomous dependent variable –

using ordinary least squares (OLS) regression. It has been demonstrated (Dey & Astin, 1993)

that logit, probit, and OLS regression analyses produce similar results when estimating a model

with a dichotomous dependent variable, yet the assumptions of OLS regression are violated

7

when the dependent variable is not continuous, and coefficients produced by OLS regression are

often uninterpretable because they may produce predicted probabilities that extend beyond the

logical boundaries of 0 and 1 (Cabrera, 1992). Although they presumably use logistic regression,

Astin and Sax (1998) and Astin, et al. (2000) do not report the statistical models that produced

their findings. In these two studies, the researchers refer to their statistically significant results in

passing, but they neither report the effect sizes for these variables nor mention the statistical

controls in their models. The absence of this critical information makes it difficult to ascertain

the soundness of their findings.

Although they provide one of the more unique approaches to the prediction of

community service during college, Marks and Jones (2004) misspecify their models by creating

three overlapping binary contrasts for an outcome with four independent response options (i.e.,

did not volunteer in high school or college; began volunteering in college; dropped volunteering

after high school; and sustained volunteering through high school and college). Specifically, the

researchers set as contrasts ‘dropped volunteering after high school’ with the other three

categories, ‘began volunteering in college’ with the other three categories, and ‘sustained

volunteering during high school and college’ with the other three categories. The lack of

independence between these three models, and the inadequate combination of divergent response

options to form reference groups – e.g., both students who did not volunteer in high school or

college and students who had sustained volunteering through high school and college were a part

of the reference group in two of the three models – made the estimates difficult to interpret and

made the implications of their findings less meaningful. A more appropriate approach to their

study would have been either to estimate two binomial models that provide a better set of

contrasts (i.e., did not volunteer in high school or college versus began volunteering in college;

and dropped volunteering after high school versus sustained volunteering through high school

8

and college) or to estimate a multinomial model with the parameter estimates for one of the four

response options (e.g., did not volunteer in high school or college) constrained to be zero.

A final limitation of the extant literature is in the operationalization of volunteerism as a

dichotomous variable indicating whether a student did or did not participate in college. By

operationalizing volunteerism in this manner, we limit our understanding of how the extent to

which students volunteer differs by their background characteristics. Perhaps more importantly,

by dichotomizing volunteerism we fail to illuminate the different intentions of students who have

not yet participated in community service. These intentions could be classified along a

continuum of planning to participate, being undecided about participating, and not planning to

participate in community service. If measured during the early stages of the college experience,

information on the characteristics of students who fall within each classification along this

continuum would assist campus policy-makers in realigning programs and resources toward

maximizing the number of students who volunteer while in college.

The purpose of this study is thus to estimate the differences in first-year students’

intentions to volunteer while in college (i.e., ‘Done,’ ‘Plan to do,’ ‘Do not plan to do,’ or ‘Have

not decided’) by their background characteristics and by the characteristics of the institutions that

they attend. Two research questions guide the study:

1. How do students’ plans to volunteer differ by their background characteristics and

college experiences?

2. Controlling for the background characteristics and college experiences of students,

how do their plans to volunteer differ by the characteristics of the institution that they

attend?

This study adds to the extant literature on the characteristics of students who volunteer by

improving upon a number of the limitations of the current research. Specifically, this study is the

9

largest study to date on the predictors of volunteerism in college, relying on data from over

125,000 first-year students at over 620 colleges and universities in the U.S. The findings from

this study are based on an analytic technique that 1) allows for the estimation of the unique

effects of student and institutional characteristics on students’ plans to volunteer, 2) is

recommended for categorical dependent variables with multiple response options, and 3) is

designed for nested sampling schemes (i.e., students randomly sampled within institutions). This

study also contributes to the previous literature by examining in greater detail the differences in

the intentions of students who have not volunteered during their first year of college. Given the

representativeness of the sample and the analytic technique, the findings from this study are

generalizable to a larger segment of the college student population than the findings of previous

research.

Methods

Data Source and Samples

Data for this study originate from the 2004 and 2005 administrations of the National

Survey of Student Engagement (NSSE). NSSE assesses for a random sample of first-year

students and college seniors at participating institutions the extent to which the students engage

in empirically-vetted good educational practices, perceive that their college or university

emphasizes these practices, and perceive that they have developed educationally and personally

as a result of these experiences. NSSE is administered annually in the spring, and the 2004 and

2005 administrations included a total of 761 participating baccalaureate degree-granting

institutions in the U.S. and Canada.

The institution sample for this study was limited to 623 institutions in the U.S. that had a

sufficient sample of first-year students after removing from the study those students with missing

10

data. An important independent variable in this study, the student’s college entrance exam score,

is provided to NSSE by participating institutions on a voluntary basis. Institutions that elect not

to provide NSSE with this information or that do not require the student to complete either the

ACT or SAT for college admission (e.g., Canadian universities) were excluded from this study.

Fifty-four percent of the institutions in the study are privately controlled; 11% of institutions in

the sample have Carnegie classifications of Doctoral-Extensive, 9% are classified Doctoral-

Intensive, 47% Masters I or II, 17% Baccalaureate-Liberal Arts, and 15% have a classification of

Baccalaureate-General. Twenty percent of the sample institutions are categorized as Non- or

Less Competitive on the Barrons’ index of institutional selectivity, 44% are categorized as

Competitive, 24% as Very Competitive, and 11% as Highly or Most Competitive. Other

descriptive statistics for the institutional sample are located in Table 1.

_______________________

Table 1 About Here

_______________________

The student sample for this study comprised 129,597 students who were randomly

sampled from the participating institutions described above. Given the study’s focus on student

plans to volunteer before graduating from college, only first-year students at these institutions

were selected for the study. Four percent of the students in the sample are nontraditionally aged

(i.e., 20 or older), 67% are female, and 20% are students of color. Sixty-one percent of the

students in the sample have at least one parent with a four-year college degree or more, 24%

have a least one parent with some college education, and 15% have parents with no college

education. Other descriptive statistics for the student sample are located in Table 2.

11

_______________________

Table 2 About Here

_______________________

Variables

The dependent variable for this study is a single item from the NSSE instrument that

measures the student’s plans to engage in community service or volunteer work while in college.

The item appears along with seven other items (e.g., study abroad, independent study or self-

designed major) that in combination measure students’ plans to participate in various enriching

educational activities. The question for this set of items reads “Which of the following have you

done or do you plan to do before you graduate from your institution?” and students may choose

from the options ‘Done,’ ‘Plan to do,’ ‘Do not plan to do,’ and ‘Have not decided.’

At the student level, the independent variables comprise measures of the students’

background (i.e., age, sex, race, citizenship status, and parents’ education) academic achievement

(i.e., entrance exam score) and characteristics associated with the college experience (i.e.,

enrollment status, college major, commuter status, fraternity or sorority membership,

participation in varsity athletics, membership in a learning community, and working for pay on-

and off-campus.) At the institution level, the independent variables comprise measures of the

institutions’ characteristics (i.e., sector, size of the undergraduate student body, and degree of

urbanicity).

Analysis

Given the nesting of students with institutions, a hierarchical multinomial logit model

was estimated to study the relationship between students’ plans to volunteer during college, their

background characteristics and college experiences, and the characteristics of the institution. In

order to provide non-redundant contrasts between each of the response options in the dependent

12

variable, the parameter estimates for the response option ‘Plan to do’ were constrained to equal

zero. Odds ratios thus represent the effect of a unit change in the independent variable (e.g.,

being female as opposed to male) on the odds of one of the other response options (e.g.,

volunteered) versus the odds of planning to volunteer. Because of the negative connotation of the

response options “Do not plan to do” and “Have not decided,” odds ratios that are less than one

within these categories should be interpreted favorably. At the student level, all variables were

entered into the model grand-mean centered. An intercept that is then modeled at the institution

level can be interpreted as the adjusted proportion of students within a particular category of the

dependent variable.

Results

Before conducting our analysis, we estimated three binary logit models to test whether

any two contiguous outcomes were indistinguishable with regard to the independent variables in

the model. If the global null hypothesis of any one of these models is retained, combining the

two outcomes would produce a more parsimonious model (Long, 1997). Using a likelihood ratio

test and a very conservative significance level, the null hypothesis was rejected for all three tests

(See Table 3). These findings suggest that there are sufficient differences between the four

outcomes with regard to the independent variables to estimate an unrestricted model.

_______________________

Table 3 About Here

_______________________

Next, we estimated an unconditional – or intercept only – model, which allows us to

examine the between institution variance in the proportion of students that fall within each

13

category. Entering the intercepts for the unconditional model (See Table 4) into the following

equations:

( )∑

=

+== J

jji

ii

xxy

2exp1

1)|1Pr(β

, (1)

( )

( )∑=

+== J

jji

miii

x

xxmy

2exp1

exp)|Pr(

β

β for all m > 1. (2)

we find that, for a typical institution, an estimated 41% of first-year students volunteered during

their first year of college, whereas another 40% planned to volunteer sometime during college,

14% had not decided about volunteering, and 6% did not plan to volunteer during college. The

estimated percentage of students that fell within any category, however, varied across

institutions. For example, the 95% plausible value interval (Raudenbush & Bryk, 2002) for the

estimated percentage of respondents who volunteered during their first year of college ranged

from 19% to 67% across the institutional sample.

_______________________

Table 4 About Here

_______________________

Finally, we estimated a conditional model that provides the effects of the student and

institutional characteristics on the students’ log odds of volunteering, being undecided about

volunteering, or not planning to volunteer (relative to planning to volunteer). Parameter estimates

and odds ratios are provided in Table 5, and are described in greater detail below.

Student Characteristics

14

All else being equal, nontraditionally-aged students had odds of volunteering during their

first year of college that were 24% greater than the odds of volunteering for traditionally-aged

students. These older students, however, also had odds of being undecided about volunteering or

planning not to volunteer that were greater than the odds for traditionally-aged students by 16%

and 50%, respectively. Females had greater odds than males of volunteering during their first

year of college (by 15%), and they had lower odds than males of being undecided about

volunteering (by 50%) or of not planning to volunteer during college (by 70%).

Although differences in the odds of volunteering between first-year students of color and

white students were statistically non-significant, students of color had consistently lower odds

than white students of being undecided about volunteering or of not planning to volunteer during

college. With regard to being undecided about volunteering, Asian American students had the

lowest odds relative to white students (0.56), followed by African American students (0.64) and

Hispanic students (0.66). With respect to not planning to volunteer in college, the same pattern is

evident, with Asian American students having the lowest relative odds (0.51), followed again by

African American students (0.57) and Hispanic students (0.61). International students had odds

of volunteering during their first year of college that were 13% lower than the odds for their

peers. These students, however, did not differ significantly from their U.S. peers with regard to

their odds of being undecided about volunteering or of not planning to volunteer while in

college.

Models estimates suggest that greater levels of parent education increase the students’

odds of volunteering during the first year of college, all else being equal. Relative to students

with one or more parents with at least a four-year degree, students with neither parent attending

college had the lowest odds of volunteering during their first college year (0.85), followed by

students with one or more parents with some college education (0.94). Differences by parent

15

education in the students’ odds of being undecided about volunteering or of not planning to

volunteer during college were statistically non-significant.

All else being equal, model estimates suggest that students who entered college with

higher levels of academic achievement had greater odds than their peers of volunteering during

the first year of college. Relative to first-year students in the highest interval of ACT composite

scores (i.e., 33 to 36), students in the lowest ACT interval (i.e., 13 to 15) had the lowest odds of

volunteering (0.54), followed by students in ACT interval 16 to 19 (0.63), ACT interval 20 to 23

(0.64), ACT interval 24 to 27 (0.73), and ACT interval 28 to 32 (0.84). Differences in the odds of

being undecided about volunteering or of not planning to volunteer during college by the

students’ entering level of academic achievement were not statistically significant.

_______________________

Table 5 About Here

_______________________

College Experiences

Students who were enrolled part-time during their first-year of college had odds of

volunteering that were 19% lower than the odds for full-time students. Part-time students also

had greater odds than full-time students of being undecided about volunteering (by 35%) and of

planning not to volunteer during college (by 60%). Students who had pledged with or joined a

fraternity or sorority had odds of volunteering during their first year of college that were 123%

greater than the odds of their peers. Their odds of being undecided about volunteering or of

planning not to volunteer were markedly less than that of their peers (by 62% and 57%,

respectively). Although students who participated in varsity athletics had the same odds as non-

athletes of volunteering during their first year of college, student athletes had lower odds than

16

non-athletes of being undecided about volunteering (by 18%) and of planning not to volunteer

(by 15%).

Although students who resided on campus had roughly the same odds of volunteering

during their first college year as students who commuted to campus, model estimates suggest

that living on campus had a positive influence on the students’ plans to volunteer, all else being

equal. Specifically, campus residents had lower odds than commuters of being undecided about

doing community service (by 28%), and of planning not to volunteer while in college (by 33%).

Residing in a learning community had a strong positive influence on students’ odds of

volunteering that was over and above the influence of residing in other living arrangements.

Learning community members had odds of volunteering during their first year of college that

were 152% greater than the odds of nonmembers, and members had odds of being undecided

about volunteering and of not planning to volunteer that were lower than the odds of

nonmembers by 30% and 43%, respectively.

Students who worked on campus from one to fifteen hours per week had 14% greater

odds than students who did not work on campus of volunteering during their first year of college.

Students in the sixteen to thirty hour range of on-campus work per week also had greater odds of

volunteering than students who did not work on campus (by 12%). Relative to students who did

not work on campus, working on campus one to fifteen hours per week or sixteen to thirty hours

per week decreased the students’ odds of being undecided about volunteering by 16% and 28%,

respectively; and working on campus one to fifteen hours per week decreased the students’ odds

of not planning to volunteer by 11%.

A similar pattern was evident for moderate levels of off-campus work per week.

Specifically, relative to students who did not work off campus, working off campus one to

fifteen hours per week or sixteen to thirty hours per week increased the students’ odds of

17

volunteering during the first year by 15% and 7% respectively, and it decreased the students’

odds of being undecided about volunteering by 12% and 8%, respectively. Working off campus

sixteen to thirty hours per week, however, also increased the students’ odds of not planning to

volunteer in college by 18%, and working off-campus more than thirty hours per week increased

the students’ odds of not planning to volunteer in college by 38%.

Relative to students who are enrolled in arts and humanities disciplines, education majors

had the greatest odds of volunteering during their first year of college (by 29%), followed by

social sciences majors (by 13%), and business majors (by 8%). Education majors and social

sciences majors also had lower odds of being undecided about volunteering (by 31% and 30%,

respectively) and of not planning to volunteer during college (by 47% and 33%, respectively).

Students who belonged to academic programs within the biological sciences, engineering,

physical sciences, and professional occupations had odds of volunteering that were not

statistically different than the odds for students within the arts and humanities. Biological

sciences majors and professional occupations majors, however, had lower odds of being

undecided about volunteering (by 44% and 30%, respectively) and of not planning to volunteer

during college (by 61% and 48%, respectively), relative to arts and humanities majors. Finally,

students who were undecided about their major had the lowest relative odds of volunteering

during the first year of college (by 16%). Undecided majors had 12% greater odds of being

undecided about volunteering than their peers majoring in arts and humanities, but they also had

18% lower odds of not planning to volunteer during college.

Institutional Characteristics

After adjusting average institutional plans to volunteer by the student background

characteristics and college experiences represented in the model, students who attended private

institutions still had 40% greater odds of volunteering during the first year of college than their

18

public institution peers. Students who attended private institutions also had lower odds than their

peers of being undecided about volunteering and of planning not to volunteer (by 31% and 37%,

respectively). Evidence from the model also suggests that undergraduate student body size had a

negative impact on students’ odds of volunteering during their first year of college. Relative to

students at institutions with less than 1,000 undergraduates, students at institutions with 20,000

undergraduates or more had the lowest odds of volunteering during the first year of college (by

28%), followed by students at institutions with 10,000 to 19,999 undergraduates (by 26%).

Although, differences in odds of being undecided about volunteering or of planning not to

volunteer by institution size do not suggest as clear of a pattern, students who attended

institutions with 1,000 undergraduates or more typically had lower odds than students at

institutions with less than 1,000 undergraduates of being undecided about volunteering or of

planning not to volunteer.

Relative to students who attend institutions within large cities, students at institutions

located in mid-size cities and students at institution located in large towns had greater odds of

volunteering (by 21% and 53%, respectively). Students at institutions in these locales also had

lower odds of being undecided about volunteering (by 15% and 20%, respectively) and of not

planning to volunteer while in college (by 16% and 26%, respectively) than their peers at

institutions in large cities. Overall, students at institutions in small towns or in rural areas had

plans to volunteer that were not significantly different from students at institutions in large cities.

Discussion and Implications

Student Characteristics

Consistent with the findings of previous research (Astin & Sax, 1998; Astin,

Vogelgesang, Ikeda, & Yee, 2000; Fitch, 1991; Marks & Jones, 2004; Pierson, 2002; Serow &

19

Dreyden, 1990), we find that females are more likely than males to volunteer during their first

year of college. Our findings go beyond the current state of knowledge, however, to suggest that

male students are not only less likely to volunteer during their first year of college, but they are

also more likely than females either to be undecided about volunteering or not plan to volunteer

during college. In other words female students that have not volunteered by the end of their first

year of college often have plans to volunteer at some time during college; male students do not

share these plans. As opposed to rationalizing this finding away as the difference in the ethic of

caring between women and men, more needs to be done on college campuses to increase the rate

of volunteering among male students. Given the intentions of male students as found in this

study, simply informing these students of the available opportunities for community service is

not enough. Instead, campus resources should be directed toward instilling in these students the

individual and societal benefits of community service. Learning more about the different

motivations of these students (e.g., Fitch, 1987; Serow, 1991; Winniford, Carpenter, & Grider,

1995, 1997) may also be necessary to more effectively target programming to encourage male

students to volunteer.

As with the studies by Marks and Jones (2004) and Pierson (2002), we find no

differences by race or ethnicity regarding the students’ propensity to volunteer during the first

year of college. Our findings do suggest, however, differences by race or ethnicity in the

students’ intentions to volunteer during college. Specifically, students of color are more likely

than white students to plan to volunteer during college versus being undecided about

volunteering or not planning to volunteer. Knowing this, outreach to these student populations is

important for the university officials in charge of community service. Students of color may feel

less connected to the organizational structure and, although they intend to participate, they may

not be aware of the appropriate channels for participation (Hurtado, Milem, Clayton-Pedersen,

20

and Allen, 1999). Working with student organizations that primarily serve students of color may

help with their inclusion in the system and assist in turning their community service intentions

into practice.

We also find that students who enter college with greater educational capital (i.e.,

parents’ education and prior academic achievement) are more likely than their peers to volunteer

during their first year of college. It may be the case that these students were also more likely to

volunteer during high school, a behavior of the students that is predictive of volunteerism during

college (Astin & Sax, 1998; Astin, et al., 2000; Pierson, 2002) but that we were unable to

measure in this study. It may also be the case that these students need to devote less time to their

academic work, or that they are better able to manage their academic load, freeing up time for

such activities as volunteering. These students may also be more likely to join first-year honor

societies, or to take part in first-year honors programs, which often emphasize or require

community service.

Programmatic endeavors need not only focus on the honors students, however, colleges

and universities can encourage participation by first year students in cascading mentoring

programs wherein students in need of tutoring receive assistance from older students, but also

provide tutoring for high school students, becoming both student and teacher, recipient and

provider of service. Practices such as these can not only promote community service habits, but

also foster an increased academic self-efficacy in students that need assistance by allowing them

to be experts in an educational setting. The academic benefits of community service could also

be very helpful for students in need of remediation or who are classified as at-risk, providing

them with stronger connections to academic material and encouraging them to structure their

time more effectively.

21

College Experiences

Unlike the results of Fitch (1991), we find that students who live on campus do not

necessarily have a greater likelihood than their peers of volunteering during the first year of

college. The differences in our results may be due to the differences in sampling between the two

studies, and it may also be due to our inclusion of learning communities in our model. Regarding

students intentions, we do find that campus residents are more likely than their peers to plan to

volunteer as opposed to being undecided about or not planning to volunteer, suggesting some

benefit to living on versus off campus. More important than simply living on campus, however,

are the benefits of on-campus living arrangements that are educationally purposeful. Learning

community membership was the single most important factor in predicting students’ propensity

to volunteer during the first year of college. Encouraging on-campus residence (where

appropriate) and increasing the use of learning communities on campus are two areas in which

institutions may be able to influence service participation among their students. The positive

effects of implementing these practices on campus extend beyond community service

participation to include other forms of student engagement and development during college

(Pascarella & Terenzini, 1990, 2005).

Consistent with previous research (Marks & Jones, 2004; Pierson, 2002; Serow &

Dreyden, 1990) students who were members of a fraternity or sorority were more likely than

their peers to volunteer during the first year of college. Greek membership was the second most

important factor in predicting volunteerism, and this finding suggests that membership in student

organizations where volunteerism is encouraged or even required joins the spirit of service with

involvement in a peer community and creates an experience for students that may reap benefits

similar to learning communities and service-learning initiatives.

22

Additionally, institutions can take away some lessons from learning communities and

Greek organizations, even when on-campus housing or large numbers of student organizations

are not available. Both communities have strong structures in place that both encourage students

to participate and provide channels for them to do so. The students in each community receive

consistent messages about the value and the importance of service. While requiring community

service may lead to a negative opinion of the process and decrease participation in the long run

(Marks & Jones, 2004), when students perceive a strong institutional commitment and receive a

consistent message from the university about the importance of and the opportunities for

participation, students are more likely to volunteer (Ward, 1996).

Compared to students in education, the social sciences, and business, students in the arts

and humanities, biological sciences, engineering, physical sciences, and professional occupations

are less likely to volunteer during the first year of college. Although the particular grouping of

majors in this study may be crude, the differences in plans to volunteer by major suggest some

broad areas or disciplines in which service learning could be further implemented. The on-going

challenge for several of these disciplines (e.g., engineering, and physical sciences) is to

incorporate service into the curriculum in a way that provides a meaningful and seamless

learning experience for students.

The fact that undecided majors are less likely than others to volunteer and are more likely

than their peers to be undecided about volunteering suggests that tying volunteerism to such

student support services as career planning and placement. Among the imperatives for the field

of career services provided by Rayman (1993) are accepting the position of career planning and

placement offices as “the most obvious and continuing link between corporate America and the

academy” and developing “alternative means of facilitating the transition from college to work”

(p.105). To the extent that local business are also involved within the community, linking the

23

academy and businesses together through joint participation in community service will reinforce

in college students the importance of service while giving the students an opportunity to explore

career opportunities and relationships with potential employers.

Contrary to previous research (Fitch, 1991; Marks & Jones, 2004), our findings suggest

that working on- or off-campus is not a detriment to the students’ propensity to volunteer. In fact,

at moderate levels, working may even be a positive factor regarding the students’ community

service and plans to volunteer. Only when the number of hours per week exceeds thirty does

working off-campus start to have a negative impact on students’ plans to volunteer during

college. This finding is similar to that of Rago and Moore (2004) who found that moderate

amounts of work (up to 20 hours off campus and up to 30 on campus) were associated with more

hours spent in both academic preparation and participation in extracurricular activities. Rago and

Moore hypothesized that students with moderate workloads had developed better time

management skills and were therefore able to successfully incorporate more activities into their

lives.

Institutional Characteristics

Even after controlling for a number of the characteristics (i.e., nontraditional age,

employment status, commuting status, part-time enrollment status) of the students that these

institutions serve, students at institutions located within urban areas are less likely to volunteer or

even to plan to volunteer than students at institutions in less urban areas. This finding is

disappointing given the numerous opportunities for these institutions to connect with their local

communities through service (Palm & Toma, 1997). It may be the case that urban institutions,

although having many opportunities for partnering with the local community, may face some of

the greatest challenges to effective partnerships, including power differences, culture or race

24

issues, differences in rhetoric, limited resources, differences in leadership, lack of measurement

and documentation of effectiveness, and the limited visibility of partnerships (Holland, 2005).

Another troubling finding of our study is that students who attend larger institutions are

less likely than students at smaller institutions to volunteer. This finding suggests that these

institutions may have more difficulty than smaller institutions in creating and maintaining an

environment that demonstrates a commitment to community service. Although these institutions

often have a larger number of student organizations and activities than smaller institutions, the

fact that these student activities are competing for student involvement may create students who

are more consumer oriented toward selecting those activities and organizations that have the

highest benefit-cost ratio. In this type of student marketplace, student organizations may be more

lenient in their community service requirements to attract and maintain their membership.

Other Implications

On the whole, the results of our study suggest that college students who are unengaged in

community service activities are not a single entity with regard to either their background

characteristics or their intentions to volunteer in the future. Consequently, a one-policy-fits-all

approach to attract these students to community service opportunities may be costly and

ineffective. Creating initiatives that effectively target these students for community service

opportunities requires a better understanding of the interactions between the students’ intentions,

their background characteristics, and the characteristics of the institutions that they attend. Using

coefficients from the current study or replicating this study with data from a single institution,

institutional researchers and other campus stakeholders who are interested in community service

initiatives at the postsecondary level can simulate the community service participation and

intentions of entering students to better realign programs and resources toward maximizing the

number of students who participate in community service while in college.

25

Limitations and Directions for Future Research

One of the major limitations of this study is the absence of independent variables that

have been demonstrated elsewhere to have an impact on college students’ propensity to

volunteer. An important predictor that was not available for this study is a measure of the

students’ pre-college community service. The Beginning College Survey of Student Engagement

(BCSSE), a survey being piloted by NSSE, asks entering college students to provide detailed

information on their high school academic and extracurricular experiences, as well as the value

that they place on their engagement in college. Using data from 70 institutions that participated

in the 2005 pilot administration of BCSSE and the 2006 administration of NSSE, we plan to

address this limitation in the current study by examining the relationship between the extent of

students’ involvement in high school service clubs and organizations, the value that they place

on service learning in college, and their propensity to volunteer during the first year of college.

Although we were able to demonstrate differences in students’ likelihood of volunteering

by the structural characteristics of the institutions that they attend, finding more meaningful

measures of institutional programmatic efforts is essential to understanding institutional impact

on students’ engagement in community service in college. To this end, we hope to combine

NSSE data with data from the Campus Compact survey of its member institutions to examine 1)

the differences in institutional commitment to community service (e.g., types of service

programs offered on campus, campus support for community service, campus obstacles to the

extension of service-learning, etc.) by the characteristics of institutions, and 2) and the effects of

institutional commitment to community service on the community service participation of

students.

26

Conclusions

Although the impacts of volunteerism on college student development are extensive and

well-documented, the characteristics of students who volunteer during college are not well

understood. Only a limited number of studies have examined the predictors of volunteerism

during college, and several limitations to this small body of research render this evidence

suspect. Using data from over 125,000 first-year students attending over 620 four-year colleges

and universities in the U.S., this study contributed to the literature by examining the differences

in students’ plans to volunteer during college by their background characteristics and college

experiences, and by the characteristics of the institutions that they attend. On the whole, the

results of our study suggest that college students who are unengaged in community service

activities are not a single entity with regard to either their background characteristics or their

intentions to volunteer in the future. Consequently, a one-policy-fits-all approach to attract these

students to community service opportunities may be costly and ineffective. Study results suggest

changes to several campus policies and programs that may remove barriers to successful

community service participation among first-year students.

27

References Astin, A. W., & Sax, L. J. (1998). How undergraduates are affected by service participation.

Journal of College Student Development, 39, 251-263. Astin, A. W., Sax, L. J., & Avalos, J. (1999). Long term effects of volunteerism during the

undergraduate years. Review of Higher Education, 22, 187-202. Astin, A., Vogelgesang, L., Ikeda, E., & Yee, J. (2000). How service learning affects students.

Higher Education Research Institute. Boyle-Baise, M., & Kilbane, J. (2000). What really happens? A look inside service-learning for

multicultural teacher education. Michigan Journal of Community Service Learning, 7, 54-64.

Cabrera, A. (1994). Logistic regression analysis in higher education: An applied perspective. In

Smart, J. C. (ed). Higher Education: Handbook of Theory and Research, 10 (pp 225-256). New York: Agathon Press.

Dey, E. & Astin, A. (1993). Statistical alternatives for studying college student retention: A

comparative analysis of logit, probit, and linear regression. Research in Higher Education, 34, 569 – 581.

Eyler, J., & Giles, D., Jr. (1999). Where's the learning in service-learning? San Francisco, CA:

Jossey-Bass. Eyler, J., Giles, D., Jr., & Braxton, J. (1997). The impact of service-learning on college students.

Michigan Journal of Community Service Learning, 4, 5-15. Eyler, J., Root, S., & Giles, D., Jr. (1998). Service-learning and the development of expert

citizens: Service-learning and cognitive science. In R. G. B. & D. K. Duffy (eds.), With service in mind: Concepts and models for service-learning in psychology. Washington DC: American Association for Higher Education.

Fitch, R. (1987). Characteristics and motivations of college students volunteering for community

service. Journal of College Student Personnel, 28, 424-431. Fitch, R. (1991). Differences among community service volunteers, extracurricular volunteers,

and nonvolunteers on the college campus. Journal of College Student Development, 32, 534-540.

Giles, D., Jr., & Eyler, J. (1994). The impact of a college community service laboratory on

students' personal, social and cognitive outcomes. Journal of Adolescence, 17, 327-339. Gray, M., Ondaatje, E., Fricker, R., Geschwind, S., Goldman, C., Kaganoff, T., Robyn, A.,

Sundt, M., Vogelgesang, L., & Klein, S. (1998). Coupling service and learning in higher education: The final report of the evaluation of the Learn and Serve America, Higher Education Program. RAND.

28

Holland, B. (2005). Reflections on community-campus partnerships: What has been learned?

What are the next challenges? In Pasque, P., R. Smerek, B. Dwyer, N. Bowman, & B. Mallory (eds.), Higher education collaboratives for community engagement and improvement. The National Forum on Higher Education for the Public Good.

Hurtado, S., Milem, J., Clayton-Pedersen, A., & Allen, W. (1999). Enacting diverse learning

environments: Improving the climate for racial/ethnic diversity in higher education. ASHE-ERIC Higher Education Report, Volume 26, no. 8. Washington, DC: The George Washington University, Graduate School of Education and Human Development.

Long, J. (1997). Regression models for categorical and limited dependent variables. Advanced

Quantitative Techniques in the Social Sciences Series, 7. Thousand Oaks, CA: Sage. Marks, H., & Jones, S. (2004). Community service in the transition: Shifts and continuities in

participation from high school to college. The Journal of Higher Education, 75, 307-339. National Survey of Student Engagement (2005). NSSE 2005 grand frequencies by class and

Carnegie classification. Indiana University Center for Postsecondary Research. Palm, R., & Toma, J. (1997). Community relationships and partnerships. In Dietz, L., and V.

Triponey (eds). Serving students at metropolitan universities: The unique opportunities and challenges.. New Directions for Student Services, No. 79 (pp 57-65). San Francisco: Jossey-Bass.

Pascarella, E., & Terenzini, P. (1991). How college affects students. San Francisco: Jossey-Bass. Pascarella, E., & Terenzini, P. (2005). How college affects students, vol. 2: A third decade of

research. San Francisco: Jossey-Bass. Pierson, C. (2002). Volunteerism in college: Impacts on cognitive outcomes, learning

orientations, and educational aspirations. Unpublished dissertation, University of Iowa, Iowa City.

Rago, M., & Moore, J. (2004). Disengaged and ignored: Are working students a lost cause.

Paper presented at the annual meeting of the Association for Institutional Research, Boston, MA.

Rayman, J. (1993). Concluding remarks and career services imperatives for the 1990s. In

Rayman, J. (ed). The changing role of career services. New Directions for Student Services, No. 62 (pp 101-108). San Francisco: Jossey-Bass.

Raudenbush, S., & Bryk, A. (2002). Hierarchical linear models: Applications and data analysis

methods (2nd). Advanced Quantitative Techniques in the Social Sciences Series, 1. Thousand Oaks, CA: Sage.

Rudolph, F. (1977). Curriculum: A history of the American undergraduate course of study since

1636. San Francisco: Jossey-Bass.

29

Serow, R. (1991). Students and voluntarism: Looking into the motives of community service

participation. American Educational Research Journal, 28, 543-556. Serow R., & Dreydan, J. (1990). Community service among college and university students:

Individual and institutional relationships. Adolescence, 25, 553-566. Terenzini, P. T. (1994). Educating for citizenship: Freeing the mind and elevating the spirit.

Innovative Higher Education, 19, 7-21. Vogelgesang, L. J., and Astin, A. W. (2000). Comparing the Effects of Service-Learning and

Community Service. Michigan Journal of Community Service Learning, 7, 25-34. Wang, W. (2000). Service Learning: Is it good for you? Paper presented at the Annual Meeting

of the American Educational Research Association Conference Roundtable, New Orleans, LA.

Ward, K. (1996). Service-learning and student volunteerism: Reflections on institutional

commitment. Paper presented at the annual meeting of the American Educational Research Association, New York, NY.

Winniford, J., Carpenter, D., & Grider, C. (1995). An analysis of the traits and motivations of

college students involved in service organizations. Journal of College Student Development, 36, 27-37.

Winniford, J., Carpenter, D., & Grider, C. (1997). Motivations of college student volunteers: A

review. NASPA Journal, 34, 134-146.

30

Table 1. Descriptive Statistics for Institution-Level Predictor Variables

Variable Mean SD Private sector 0.543 0.499 Undergraduate enrollment size less than 1,000 † 0.088 0.284 Undergraduate enrollment size 1,000 to 2,499 0.327 0.470 Undergraduate enrollment size 2,500 to 4,999 0.215 0.411 Undergraduate enrollment size 5,000 to 9,999 0.173 0.379 Undergraduate enrollment size 10,000 to 19,999 0.148 0.355 Undergraduate enrollment size 20,000 or more 0.048 0.214 Located in a large city † 0.350 0.477 Located in a mid-size city 0.372 0.484 Located in a large town 0.051 0.221 Located in a small town 0.177 0.382 Located in a rural area 0.050 0.218 N = 623 † Reference group

31

Table 2. Descriptive Statistics for Student-Level Predictor Variables

Variable Mean SD Nontraditional age 0.044 0.204 Female 0.666 0.472 African American 0.052 0.222 Asian American 0.055 0.228 Hispanic/Latino 0.046 0.209 White/Caucasian † 0.796 0.403 Other race 0.052 0.222 International student 0.032 0.175 At least one parent with a bachelor's degree † 0.613 0.487 At least one parent with some college 0.240 0.427 Neither parent attended college 0.147 0.354 ACT composite score 13 to 15 0.017 0.128 ACT composite score 16 to 19 0.129 0.335 ACT composite score 20 to 23 0.289 0.453 ACT composite score 24 to 27 0.316 0.465 ACT composite score 28 to 32 0.226 0.418 ACT composite score 33 to 36 † 0.024 0.152 Part-time enrollment status 0.017 0.128 Resided in campus housing 0.769 0.421 Learning community member 0.142 0.349 Fraternity/Sorority member 0.107 0.310 Student athlete 0.125 0.330 Did not work on campus † 0.743 0.437 Worked on campus 1 to 15 hours per week 0.220 0.415 Worked on campus 16 to 30 hours per week 0.034 0.181 Worked on campus over 30 hours per week 0.003 0.050 Did not work off campus † 0.720 0.449 Worked off campus 1 to 15 hours per week 0.148 0.355 Worked off campus 16 to 30 hours per week 0.109 0.312 Worked off campus over 30 hours per week 0.023 0.149 Arts and humanities major † 0.147 0.354 Biological sciences major 0.087 0.281 Business major 0.138 0.345 Education major 0.106 0.308 Engineering major 0.054 0.225 Physical sciences major 0.036 0.187 Professional occupations major 0.107 0.309 Social sciences major 0.132 0.339 Other major 0.138 0.345 Undecided major 0.055 0.228 N = 129,597 † Reference group

32

Table 3. Specification Tests for Combining Contiguous Outcomes

Dependent Variable Deviance d.f. Sig. ‘Done' vs. 'Plan to do' 327147.06 46 < 0.0000 ‘Plan to do' vs. 'Have not decided' 202588.44 46 < 0.0000 ‘Have not decided' vs. 'Do not plan to do' 75557.39 46 < 0.0000

33

Table 4. Results of Unconditional Hierarchical Multinomial Logit Model of Volunteering Done Have not decided Do not plan to do

Independent Variable b Std. Error b Std. Error b Std. Error Intercept 0.02 0.02 -1.12 0.02 -2.02 0.03

34

Table 5. Results of Conditional Hierarchical Multinomial Logit Model of Volunteering

Done Have not decided Do not plan to do Independent Variable b Odds Ratio b Odds Ratio b Odds Ratio

Intercept 0.02 -1.24 -2.30 Nontraditional age 0.21 †† 1.24 0.15 †† 1.16 0.41 †† 1.50 Female 0.14 †† 1.15 -0.70 †† 0.50 -1.21 †† 0.30 African American -0.03 -0.45 †† 0.64 -0.55 †† 0.57 Asian American -0.03 -0.59 †† 0.56 -0.68 †† 0.51 Hispanic/Latino 0.05 -0.41 †† 0.66 -0.49 †† 0.61 Other race 0.05 -0.17 †† 0.84 -0.10 International student -0.14 † 0.87 0.09 0.02 Parent with some college -0.06 †† 0.94 0.04 0.02 Parents with no college -0.16 †† 0.85 0.03 -0.04 ACT composite score 13 to 15 -0.61 †† 0.54 -0.02 -0.07 ACT composite score 16 to 19 -0.47 †† 0.63 0.02 -0.18 ACT composite score 20 to 23 -0.44 †† 0.64 -0.00 -0.20 ACT composite score 24 to 27 -0.32 †† 0.73 -0.03 -0.14 ACT composite score 28 to 32 -0.17 † 0.84 -0.06 -0.10 Part-time enrollment status -0.21 † 0.81 0.30 †† 1.35 0.47 †† 1.60 Resided in campus housing -0.03 -0.33 †† 0.72 -0.40 †† 0.67 Learning community member 0.92 †† 2.52 -0.36 †† 0.70 -0.56 †† 0.57 Fraternity/Sorority member 0.80 †† 2.23 -0.97 †† 0.38 -0.85 †† 0.43 Student athlete -0.03 -0.20 †† 0.82 -0.16 †† 0.85 Work on campus 1-15 hours 0.13 †† 1.14 -0.17 †† 0.84 -0.12 † 0.89 Work on campus 16-30 hours 0.11 † 1.12 -0.33 †† 0.72 0.07 Work on campus 31+ hours 0.21 0.07 0.16 Work off campus 1-15 hours 0.14 †† 1.15 -0.13 †† 0.88 0.09 Work off campus 16-30 hours 0.07 † 1.07 -0.08 † 0.92 0.16 † 1.18 Work off campus 31+ hours 0.13 -0.03 0.32 †† 1.38 Biological sciences major 0.07 -0.58 †† 0.56 -0.93 †† 0.39 Business major 0.08 † 1.08 -0.10 -0.09 Education major 0.26 †† 1.29 -0.37 †† 0.69 -0.63 †† 0.53 Engineering major -0.01 -0.04 -0.14 Physical sciences major 0.01 -0.08 -0.17 Professional occupations major 0.06 -0.35 †† 0.70 -0.65 †† 0.52 Social sciences major 0.13 †† 1.13 -0.35 †† 0.70 -0.39 †† 0.67 Other major 0.13 †† 1.14 -0.01 -0.09 Undecided major -0.17 †† 0.84 0.11 † 1.12 -0.19 † 0.82 Private sector 0.34 †† 1.40 -0.37 †† 0.69 -0.46 †† 0.63 Size 1,000 to 2,499 -0.11 -0.11 -0.19 Size 2,500 to 4,999 -0.19 -0.30 † 0.74 -0.34 † 0.71 Size 5,000 to 9,999 -0.23 -0.14 -0.16 Size 10,000 to 19,999 -0.30 * 0.74 -0.29 * 0.75 -0.36 Size 20,000 or more -0.32 * 0.72 -0.38 † 0.68 -0.41 † 0.66 Located in a mid-size city 0.19 †† 1.21 -0.16 † 0.85 -0.17 † 0.84 Located in a large town 0.43 †† 1.53 -0.23 † 0.80 -0.30 † 0.74 Located in a small town 0.12 -0.02 -0.06 Located in a rural area 0.25 -0.21 * 0.81 -0.21 * p < .05, † p < .01, †† p < .001