Rachelle Sharpe June 15, 2012 SFARN

Rachelle SharpeJune 15, 2012

SFARN

How Washington low-income community college students finance their education during a recession: dependent &

independent students’ choices to borrow, work &/or attend part-time

Washington StudentAchievement Council

Introduction

• WA State needs to educate more students at higher levels (HECB, 2008)

– There are individual & societal benefits to postsecondary education

• The “great recession” has led to higher tuition at a time when more students enroll– Many students attend community colleges (SBCTC, 2010)

– More low-income students are enrolling

Problem Statement

Indebtedness, working too much, and attending part-time are risk factors associated with low-income students.• Low-income students are more likely to be

independent, work full-time & attend part-time (Berkner & Choy, 2008)

• Increased work hours negatively impact academic progress (Horn & Nevill, 2006)

• Research has not addressed directly why students choose to work, borrow &/or attend part-time in combination

Purpose of Study

• Examine choices made by independent & dependent low-income community college students

• Describe the students’ reasons reported for choice of work, loan &/or part-time enrollment

• Expand knowledge for campuses and state policymakers– Tailor programs and policies to financial status

Scope of StudySurvey of WA community college students enrolled Fall 2009 &

Winter 2010 about their financing choices.• Assumptions

– At least 5% would respond– Students would have access to a computer– The online survey would produce a complete data set– Students would report their decisions and be truthful

• Limitations – Low-income students– Community college students– Patterns during a recession may differ

• Delimitations– Removal of records: younger than 18, no email on FAFSA, attending a CC that

doesn’t offer federal Stafford loans, or those included in the pilot study

Significance of Study

• Policy makers will be able to affect social change if they better understand how & why certain financing options are chosen

• Examining the financing strategies of low-income students during a recession improves understanding

• Statewide strategies to address any patterns of over-use of finance options are suggested

Student Change & Student Demand Theories

• Traditional student change theories review growth and development with limited consideration of diversity

• Student demand researchers examine whether economic policies affect student choices

• Price-response studies found (McPherson & Schapiro, 1991; St. John, 1990)

– first-time attendance decisions were more responsive to aid than tuition

– persistence affected by tuition & aid

Financial Decisions• Students face many decisions affecting cost (where,

enrollment rate) & methods of financing• Working more than part-time & attending college part-time

negatively affect persistence for low-income students (King, 2002)

• Financial aid plays a role in the college choice process (Dongbin, 2004; Kim, 2004; Lillas & Tian, 2008)

• Students who receive aid are more likely to persist (Heller, 2003; Fike & Fike, 2008; Wei & Horn, 2009)

• Current recession has affected how college students choose a college & their ability to fund their education (College Board, 2009; HERI, 2010)

Risk Factors

• Low-income students are less likely to study full-time (King, 2002)

• Independent students withdraw at higher rates, tend to be older & work more than dependent students (Berkner & Choy, 2008)

• Part-time attendance & working full-time are among several risk factors for low-income students progress to degree (Lee & Horn, 2009)

Loans, Work & Part-time

• Full-time students working 25 or more hours per week were more likely to borrow (King & Bannon, 2002)

• Low-income students who both worked part-time & borrowed performed better than those who only worked part-time (King, 2002)

• Low-income students who did not work were less likely to attend full-time than the part-time workers (King)

Research Design• Non-experimental quantitative study using three phases to

examine the patterns among the financing choices of low-income community college students– independent variables not controlled– students not randomly selected– no treatment in the study

• Comparison of groups used rather than predicting relationships as there is no logical order among the variables (Johnson & Christensen, 2004)

• Many confounding variables exist such as academic preparation or family obligations

Approach

• Phase 1: Test for differences by financial status (independent or dependent) related to three different financing options (working, borrowing &/or part-time enrollment)

• Phase 2: Focus on students who use working or borrowing as a finance option– Determine differences in level (low, medium, high)

of each option

Research Questions

• Are there any patterns in the differences by financial status for low-income community college students in the state of Washington related to the strategies that they use to finance their educations?– Are there any patterns in the differences among dependent and

independent students who borrow based on their level of borrowing?

– Are there any patterns in the differences among dependent and independent students who work based on their level of working?

• What factors do students report as affecting their decisions to choose a financing option (i.e. whether to work, borrow and/or enroll part-time)?

Population & Sample

• The 2009-10 State Need Grant (SNG) program serving 71,000 will be used to identify low-income students– filtered by age >18, email on FAFSA, SNG Fall &

Winter, CC that offers loans & removal of rejected email addresses (n = 19,718)

– results can be generalized to larger CC population• 5% return expected – 10% achieved

Survey Instrument

• Survey designed specifically for this study• Web-based survey to reduce cost• Separate survey for D vs. I students– Differentiate levels among groups– Ensure dependency status known

• Validity of survey items supported through generating items from literature review

• “Skip-logic” used to present appropriate questions & reduce respondent fatigue

Survey Instrument

Data Collection & Procedure

• Letter of cooperation from the Higher Education Coordinating Board re. access to data

• “Access code” given to students to provide “consent”– personally identifiable information was removed

• Data was screened for missing or extreme values and underrepresented cells– representation in all but 1 of 16 combinations (part-time

Dependent borrowers)

Statistical Approach

• Log-linear analysis used to examine the cross-tabulations– categorical variables– all variables can be analyzed at once reducing

Type 1 error • The log-linear approach extends the basic chi-

square approach to the analysis of multidimensional contingency tables

Statistical Approach

Variable Code Output

IV = Dependency Status

Independent or DependentDS I or D

DV1 = Borrow BW Y or N

DV2 = Work WK Y or N

DV3 = Part-time PT Y or N

DV4 = Borrow Level BWL L, M H

DV5 = Work Level WKL L, M H

• A 2x2x2x2 contingency table for Phase 1 from the cross-tabulations of the IV (dependency status) and DVs (working, borrowing & attendance)

• Phase 1 Explanatory Model: DS = WK + BW + PT • Phase 2 Explanatory Models: DS = WKL and DS = BWL

Variable Coding

Phase 1 Findings

• Expected frequencies found to be acceptable• Explanatory Null Model

Null = STATUS + PT * WK * BW• Null Model found statistically significant (L2 =

107.924, df = 7, p < .001)• Zero Order Model (3 dependent variables) not

significant

Phase 1 FindingsModel/Source Component L2 df P

Null 107.924 7 <.001

Due to Part-Time 8.376 1 .004

Due to Working 8.536 1 .003

Due to Borrowing 82.693 1 <.001

Due to All Interactions 7.218 4 .125

• Strongest difference in the borrowing component• Null hypothesis rejected as significant differences

found

Phase 1 Findings• Post hoc analysis of Standardized Residuals

considered all 3 DVs• Bonferroni technique used for Type I error a

priori alpha (.05)/16 (# of cells) gives threshold of p = .003 for significance

• Only 3 of 16 combinations were significant– Dependent working (SR = 5.980, p < .001)– Dependent borrowing (SR = -5.171, p < .001)– Dependent working, borrowing (SR = -2.899, p = .002)

Phase 2 Findings

• None of the differences was found to be significant for either working or borrowing levels ~ null hypotheses supported– working level (L2 = .303, df = 2, p = .859)– Borrowing level (L2 = 6.753, df = 2, p = .034)• post hoc analysis did not produce any significant

findings

• Larger sample size may have led to significant results

Phase 3 Findings• Primary reasons to work– earn funds to pay expenses– desire to gain skills & experience

• Dependent comments: buy health benefits, pay bills, secure funds between quarters, & support family

• Independent comments:health benefits, pay bills, family, have a good job & keep a routine

Phase 3 Findings

• Primary reasons to borrow for Dependent– cover expenses– borrow the maximum– get immediate funding

• Primary reasons to borrow for Independent– borrow the maximum– just what was needed– get immediate funding

Phase 3 Findings

• Primary Dependent student reasons to attend part-time– work– avoid scheduling conflicts

• Primary Independent student reasons to attend part-time borrow the maximum– work– meet family obligations

Interpretation of Findings• The importance of borrowing may represent an emerging

pattern during a recession of low-income community college students receiving significant grant aid who still need to borrow

• Although Independent students are using each option, the differences between themselves and Dependent students were not significant– Perhaps this finding reflects that during a recession the

differences between the two groups are less apparent• It appears students pursue a variety of financing methods to

balance their limited fiscal and time resources

Interpretation of Findings• The reasons students reported they chose to select each of the finance

options reflects a tension of balancing time and finances to support their success in school

• It appears that during a recession students may be more likely to desire the maximum funding available, even in the form of a loan to cover expenses

• The reasons align with idea that financial resources are the major factor that determines attendance status (King, 2002)

• Students appear to consider the impact of their financing choices on their academic success

• The self-reported reasons students select finance options reflect the importance of factors such as academic preparation, family situation, work obligations, and affordability

Some Descriptive Patterns• Patterns do not mirror literature review– Independent students worked at lower rate (50% vs. 69%)

& borrowed at higher rate (45% vs. 18%)– Perhaps students are unable to find work in the recession

& need to borrow to fill gaps• Half or more of students in both groups worked more than 30

hours• About ¾ of Dep students & 84% of Ind borrowed at least

“medium” levels• Most in part-time students in both groups attended ¾ time

Implications for Social Change• Financial aid is at the heart of social change, providing

access & opportunities to low-income students • The amount & type of the financial aid may make a

difference – for low-income students receiving grant aid during a

recession, there was still a need to attend part-time, borrow & work

• Examination of students’ financing strategies better positions policymakers & campus administrators to make program, policy & procedural adjustments

Recommendations for Action• Adjust policies to address unique challenges of

Independent students– i.e. early incentive grant to transfer students

• Examine part-time levels on campuses• Information campaigns re. scholarships, budgeting,

minimizing debt & repercussions of too many work hours• Target work-study policies to improve retention of needy

students• Evaluate costs of both groups students to determine

proper budgets

Suggestions for Further Study

• Replicate study & improve sample size• Compare to non-recessionary periods• Survey students from all sectors of institutions• Explore reasons more deeply through focus

groups• Examine behavior of students without full

grant packages• Control for confounding variables

Conclusion

• There were differences in the financing strategies – Part-time enrollment and working were statistically

significant & borrowing was especially statistically significant

– Dependent students were over-represented in working & under-represented in borrowing

• Reasons the financing strategies chosen related to the need for funds & time to support their expenses & academic success

Hypotheses

• H1˳: There will be no differences between dependent or independent low-income community college based on their chosen financing options (work, borrow, and/or part-time).

• H1A: There will be significant differences between dependent or independent low-income community college students based on their financing options chosen (work, borrow, and/or part-time).

• H2A˳: There will be no differences between working dependent or independent low-income community college students based on their level of work.

• H2AA: There will be significant differences between working dependent or independent low-income community college students based on their level of work.

• H2B˳: There will be no differences between dependent or independent low-income community college student borrowers based on their level of loans.

• H2BA: There will be significant differences between dependent or independent low-income community college student borrowers based on their level of loans.

Decision Making Theories

• Two social cognitive theories evaluated behavioral, cognitive & environmental factors that influence decisions– Skinner (1971) suggested that behavior is

controlled primarily by environment– Bandura (1986) focused on cognitive reflection

and social environment