methodological and technical appendix

12
Methodological and Technical Appendix Research on the Impact of Student Loan Debt on Household Financial Health APRIL 2019 Authored by David Ansong, Julia Barnard, Jess Dorrance, and Kate Sablosky Elengold

Upload: others

Post on 30-Jan-2022

7 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Methodological and Technical Appendix

Methodological and Technical Appendix

Research on the Impact of Student Loan Debt on Household Financial Health

APRIL 2019

Authored by David Ansong, Julia Barnard,

Jess Dorrance, and Kate Sablosky Elengold

Page 2: Methodological and Technical Appendix

Methodological and Technical Appendix

2 UNC CENTER FOR COMMUNITY CAPITAL

ACKNOWLEDGEMENTS

The Center for Community Capital gratefully acknowledges MetLife Foundation, whose generous support made this work possible. The views and conclusions of this report are those of the authors alone and do not necessarily reflect the opinions of our funders. We would like to thank Nathalie Santos and Eileen Harvey for their assistance with data analysis, Allison Freeman for editorial assistance, and Audrie Lathrop for graphic design. Additionally, the authors wish to express their deep gratitude to all of the individuals who shared their time and their stories with us. We thank you.

ABOUT THE CENTER FOR COMMUNITY CAPITAL

The Center for Community Capital (CCC) is a non-partisan, multi-disciplinary research center housed within the University of North Carolina at Chapel Hill, and is a leading center for research and policy analysis on the power of financial capital to transform households and communities in the United States. It is part of the University of North Carolina at Chapel Hill’s College of Arts and Sciences.

The center’s in-depth analyses help policymakers, advocates, and the private sector find sustainable ways to expand economic opportunity to more people, more effectively.

ABOUT METLIFE FOUNDATION

MetLife Foundation was created in 1976 to continue MetLife’s long tradition of corporate contributions and community involvement. Since its founding through the end of 2017, MetLife Foundation has provided more than $783 million in grants and $70 million in program-related investments to organizations addressing issues that have a positive impact in their communities. In 2013, the Foundation committed $200 million to financial inclusion, and our work to date has reached more than 6 million low-income individuals in 42 countries. To learn more about MetLife Foundation, visit

metlife.org.

Page 3: Methodological and Technical Appendix

Methodological and Technical Appendix

3UNC CENTER FOR COMMUNITY CAPITAL

MetLife Foundation funds research and thought leadership in the spirit of generating dialogue on important societal topics. The opinions expressed in this report are those of the UNC Center for Community Capital and do not necessarily represent those of MetLife Foundation.

Page 4: Methodological and Technical Appendix

Methodological and Technical Appendix

4 UNC CENTER FOR COMMUNITY CAPITAL

With funding from MetLife Foundation, the Center for Community Capital (CCC) at the University of North

Carolina – Chapel Hill conducted exploratory research to better understand individuals’ higher education

financing decisions, their experiences in student loan repayment, and the impact of student loans on overall

financial health. The purpose of this Appendix is to describe (a) the research design and methodologies that

the project employed, (b) the study sample, and (c) the differences and similarities between the CCC study

sample and other nationwide samples, including the National Post-Secondary Aid Study (NPSAS) and the U.S.

Census. This document serves as a technical appendix to a series of research and policy briefs published using

the CCC data described herein. The Appendix is organized into three parts. Part 1 provides an overview of the

study’s research methods, including the research design, sampling and recruitment strategy, data collection

techniques, and a description of the instruments developed for the study. Part 2 describes the CCC sample,

with particular emphasis on the study participants’ demographic profiles and socioeconomic characteristics.

Part 3 offers a summary of the comparison between the CCC sample and the NPSAS sample.

METHODOLOGY

This study used a mixed-methodology design to collect and analyze qualitative data (in-depth interviews) and quantitative data (paper surveys) alongside with analysis of publicly available data from the National Post-Secondary Aid Study (NPSAS)1 to better understand the implications of student loans on household financial health. Inclusion of NPSAS data places CCC’s original qualitative and quantitative data in context and allows researchers to assess the study sample relative to the NPSAS’s nationally representative sample of undergraduates and graduates. The mixed-methods approach was necessary to provide insight into individual experiences without losing the context for that insight. Through this project, we were able to corroborate existing research, locate areas of mismatch between data and lived experience, and fill gaps in existing knowledge by allowing the nuanced and complex stories of borrowers themselves to deepen and challenge the ways we think about student loan debt and the value of higher education.

RECRUITMENT & SAMPLE

The CCC sample was limited to individuals who (a) used loans to finance portions of their higher education, (b) were not currently enrolled at a higher education institution

Part 1: Research Methods

at the time of the interview,2 and (c) were available to meet in person for an interview at selected locations. Based on these criteria, we used a combination of three nonprobability sampling strategies—convenience, maximum variation, and snowball sampling—to recruit 65 participants living in 10 cities in 6 states depicted in Figure 1. To a lesser extent, we used consecutive sampling within some geographic locations. Interview participants were primarily recruited via flyers and Facebook announcements, and all of them received a $35 Target gift card as an incentive and

to thank them for their time.3

INSTRUMENTATION

Each participant completed a self-administered paper survey and participated in a semi-structured interview. The paper

1 The purpose of NPSAS is to compile a comprehensive research dataset, based on student-level records, on financial aid provided by the federal government, the states, postsecondary institutions, employers, and private agencies, along with stu-dent demographic and enrollment data. NPSAS is the primary source of informa-tion used by the federal government (and others, such as researchers and higher education associations) to analyze student financial aid and to inform public policy on such programs as the Pell grants and Direct/Stafford loans. https://nces.ed.gov/surveys/npsas/about.asp2 Two individuals in the sample were still pursuing a degree at the time of the interview.3 Consistent with standard social science research protocols involving human par-ticipants, a monetary gift was offered as an incentive to encourage participation.

Page 5: Methodological and Technical Appendix

Methodological and Technical Appendix

5UNC CENTER FOR COMMUNITY CAPITAL

survey included questions about the participant’s educational institutions, degree type, completion level, tuition amount(s), loan type(s), loan amount(s), other financing sources, and general financial management, including savings habits and bill payment, the perceived effects of student loan debt on a range of future financial decisions, and demographic questions.

An interview guide was used to provide general structure to the interview and ensure the consistent asking of key questions, however, interviewers were able to probe responses for additional clarity and ask relevant follow up questions. Interview questions focused on the participant’s motivation for higher education and areas of study, the participant’s sources of information about how to pay for school, and the particpant’s experiences with employment, housing, family obligations, and extracurricular activities while in school. Respondents were also asked to discuss their experiences repaying student loans, the perceived effects of student loan debt on future financial decisions, and their advice to others who may be considering higher education.

DATA COLLECTION

Primary data collection activities spanned 10 months, from August 2017 to May 2018. Depending on the logistical arrangements, participants completed the survey immediately before or after their interview. The in-depth interviews ranged in length from 30 minutes to 75 minutes and were audio recorded with the participants’ consent. In addition to the individual

interviews, we conducted six interviews with key informant experts to help inform the research direction and the analysis of the data.

DATA MANAGEMENT & ANALYSIS

Qualitative data

Once the original data was collected and transcribed,

we used a deductive approach to begin the analysis.

We identified key themes and topics to monitor across

the interview transcripts. Two research team members

independently reviewed the transcripts for substantive

thematic findings. Although we did not formally code in

teams or track inter-coder reliability, the coding approach

was driven by and analyzed using team-identified themes.

We used Atlas.ti, a qualitative analysis software, to manage

the data and facilitate identification of themes and specific

critical passages.

Quantitative data

We generated summary statistics from both participant

questionnaire data and existing NPSAS data. We relied

on measures of frequency to examine the distribution

of sociodemographic characteristics of the CCC sample.

In select cases, researchers also performed comparative

analyses to highlight how trends in the CCC sample

compare and contrast to the nationally-representative

NPSAS data. Finally, bivariate graphs were used to illustrate

how particular characteristics vary by data source.

Los Angeles

CA

TX

KSIL

NC

NY

LawrenceKansas City

Chicago

New York

Durham/

San Antonio

Houston

Bryan/College Station

Figure 1. A map of cities and states reflecting primary data collection zones

Austin

Chapel Hill

Page 6: Methodological and Technical Appendix

Methodological and Technical Appendix

6 UNC CENTER FOR COMMUNITY CAPITAL

40%

56%

18%

15%

11%

15%

5%

11%

8%

1%

0%2% 3%

16%

CCC Sample NPSAS

Whi

te

Bla

ck/A

fric

an A

mer

ican

His

pani

c or

Lat

ino

Asi

an Am

eric

an In

dian

/Ala

ska

Nat

ive

Oth

er

Mor

e th

an o

ne ra

ce

Demographic details and selected descriptive statistics of the CCC sample are provided below. For context, the sample is compared to national data from National Postsecondary Student Aid Study and Census data.

DEMOGRAPHICS

RaceOver a third of the CCC sample was White (40%). However, the sample had a higher representation of Hispanics (21%), Blacks (16%), Native Americans (10%), and Asians (8%) compared to the 2017 estimates of the racial distribution of the US population.4 Similarly, racial minority groups were over-represented in the CCC sample, compared to the combined NPSAS graduate and undergraduate samples (see Figure 2).

4 https://www.census.gov/quickfacts/fact/table/US/PST045217

AgeThe CCC study participants ranged in age from 19 to 58 years, with an average age of 31 (SD = 8), which was similar to the average age (29) for the NPSAS samples (Undergraduate: 25.73, SD = 9.06; Graduate: M = 32.28, SD = 9.66). However, the oldest persons in the NPSAS graduate (85 years) and undergraduate samples (89 years) were much older than the oldest CCC study participant (58 years). In other words, the NPSAS sample was more heterogeneous in terms of age, compared to the CCC sample.

GenderThe CCC sample of 65 participants was split between females (51%, n=33) and males (49%, n=32). That is a slight difference from the gender distribution of the NPSAS samples. While the gender distribution was nearly balanced (51% female) in the CCC sample, the female respondents in the NPSAS sample (58%) outnumber their male counterparts by 16% points as shown in Figure 3.

Figure 2. Racial composition of the CCC and the NPSAS samples

Part 2: Sample Description

Page 7: Methodological and Technical Appendix

Methodological and Technical Appendix

7UNC CENTER FOR COMMUNITY CAPITAL

40%

56%

18%

15%

11%

15%

5%

11%

8%

1%

0%2% 3%

16%

CCC Sample NPSAS

Whi

te

Bla

ck/A

fric

an A

mer

ican

His

pani

c or

Lat

ino

Asi

an Am

eric

an In

dian

/Ala

ska

Nat

ive

Oth

er

Mor

e th

an o

ne ra

ce

38%

55%

6%

Bachelor’s degree

Advanced academic degree

Associate’s degree

Figure 4 shows that in both the CCC and the NPSAS samples, respondents overwhelmingly identified as single. In the CCC sample, nearly two out of every three respondents (63%) reported that they were either single, divorced, or widowed. In the NPSAS samples, the marital status was even more skewed towards single respondents, with three-fourths of the sample (74%) reporting that they were single.

EDUCATIONAL BACKGROUND AND EXPERIENCES

Highest Education LevelAlthough the entire CCC sample had some form of postsecondary education, the education levels varied. As shown in Figure 5, more than half of the sample (55%) pursued a bachelor’s degree, as their highest degree, and 38% pursued some form of an advanced academic degree (graduate level education). Only 6% of respondents in the CCC sample reported that an associate’s degree was their highest level of education. In the NPSAS undergraduate sample, approximately 65% of the undergraduate sample pursued 2-year degree.

Institutions Attended and Completion StatusOverall, a majority of the CCC sample attended multiple schools in the course of their postsecondary education. Thirty-eight percent of the CCC sample attended only one school, 40% attended two schools, and 22% attended three schools. In the NPSAS samples, it is not clear how many respondents attended multiple educational institutions during their postsecondary school years.

Female

Male

CCC Sample NPSAS Sample

58%

42%

51%49%

Figure 3. Gender distribution in the CCC and NPSAS samples

Figure 5. The highest education level of the CCC sample

Single, divorced, widowed

Married or living with a partner

CCC Sample NPSAS Sample

74%

26%

63%

37%

Figure 4. Marital status for the CCC and NPSAS samples

Page 8: Methodological and Technical Appendix

Methodological and Technical Appendix

8 UNC CENTER FOR COMMUNITY CAPITAL

Public/PrivatePrivate Public

First-Generation Students

Not First-Generation Students

10% 67% 23%

14% 40% 46%

52%

12%

36%

Private

Public

Public & Private

Figure 6 shows that, overall, the CCC study participants who attended multiple institutions were more likely to drop out of school as compared to those who attended only one institution. Among the one-third of the CCC study participants (38%) who reported attending only one school, the overwhelming majority (96%) completed. For the 62% of respondents who reported attending multiple schools, however, nearly a fifth (17%) did not complete a degree at one or more of those institutions. It is unclear if they plan to continue pursuing a degree or not.

Type of Institution Attended Regardless of whether the participants attended multiple institutions, at least half of the CCC sample (52%) attended only public institutions (Figure 7). Just over a fifth of the CCC sample (12%) attended only private institutions, and a third of the sample attended both public and private higher education institutions. It is worth noting that a handful of the 65 CCC study participants reported that they undertook part of all of their higher education either online (n=5), at an HBCU (n=2), or at a tribal college (n=3).

First Generation College StudentsNearly half (46%) of the participants in the CCC sample were the first in their families to attend college. A majority of the participants in the CCC sample (54%) were the first in their families to attend college. Two-thirds of these first-generation students were female (67%). The reverse is true for those participants who were not the first in their families to attend college (63%). Most of the first-generation college students attended public schools. Figure 8 shows that more than two-thirds (67%) of first generation college students attended public institutions of higher learning; fewer than half of those who were not first-generation students (40%) attended public institutions.

Figure 7. Type of higher educational institutions attended by the CCC sample

Figure 8. First-generation college students by school type

Did not complete

Other

Completed

Currently pursuing

Multiple Institutions

One Institution

96%

4%

68%

17%

15%

Figure 6. School completion by the number of schools attended

Page 9: Methodological and Technical Appendix

Methodological and Technical Appendix

9UNC CENTER FOR COMMUNITY CAPITAL

> $55,000

$45,000-$54,999

$35,000-$44,999

$25,000-$34,999

$10,000-$24,999

7%

7%

15%

43%

18%

11%

< $10,000

Figure 11. Household Income of CCC sample

Homeless

Living with partner’s family

Living alone/with non-relatives11%

73%

2%

14%

Living alone/with family

Figure 10. Living situation of CCC sample

Not working

Full-time

Part-time

76%

14%

10%

ECONOMIC WELL-BEING

Employment Employment status and living situation are proxies for economic status, and they offer insights into economic well-being. In the CCC sample, at least 4 in every 5 respondents reported that they had a job at the time of the interview: 74% were full-time employees and 10% were employed on a part-time basis (Figure 9). Only 14% of respondents reported that they were not employed. Of those not employed, one person is a retiree and two are students.

Living ConditionsAt the time of the interview, the majority of the CCC sample (73%) reported either living alone or with an immediate family member (which included living with families of origin and current immediate family), 14% reported living with their spouse/partner’s family, and 11% reported living with non-relatives (Figure 10). One respondent indicated that they were homeless.

IncomeThe CCC study classified household income into six income groups. As shown in Figure 11, nearly half the sample (43%) indicated that they earn $55,000 or more annually; the remainder (57%) earn less than $55,000. About 18% of the participants earn less than $25,000, which is equivalent to 2017 federal poverty level for a 3-person household.

Figure 9. Employment status of CCC sample

Page 10: Methodological and Technical Appendix

Methodological and Technical Appendix

10 UNC CENTER FOR COMMUNITY CAPITAL

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

90%

< $1

0,0

00

$10

,00

0-$

24,0

00

$25,

00

0-$

34,0

00

$35,

00

0-$

44,9

99

$45,

00

0-$

54,9

99

> $5

5,0

00

Associate’s Degree

Bachelor’s Degree

Advanced Degree

When we juxtapose participants’ income level with their academic achievement (figure 12), we find that majority of the top income earners (54%) have some form of advanced academic degree. No participant who attained only an associate’s degree identified themselves as a top income earner in the CCC study.

STUDENT DEBT LEVELS, GRANTS, SCHOLARSHIPS, AND OTHER SOURCES OF EDUCATION FUNDING

Overall, the CCC sample of 65 individuals owed a total of $2,525,236 in student loans for their undergraduate and graduate education. The average amount of debt in the CCC sample was $45,913 compared to the average of $23,366 (for those with graduate degree) and 10,833 (for those with an undergraduate degree). Besides having received more student loans than the NPSAS graduate and undergraduate samples, the CCC study participants’ debt levels varied more widely (Min = 650, Max = 211,850) compared to that of the NPSAS graduate (Min = 108, Max = 125,640) and NPSAS undergraduate samples (Min = 1, Max = 65,095). In other

Figure 12. Academic achievement by income group in the CCC sample

words, the CCC study participants were more heterogeneous than the NPSAS participants regarding student debt levels.

While the overwhelming majority of the CCC sample (82%) received some form of grants, scholarships, and fellowships, fewer respondents in the NPSAS samples reported receiving grants and scholarship (Graduate: 40%, Undergraduate: 63%). On average, CCC study participants received more grants and scholarship (M = $46,645) compared to the NPSAS study participants (Graduate: M = $9,519.07; Undergraduate: M =7,425.22).

Sources of Education FundingFigure 14 shows the primary sources of higher education funding. Participants in the CCC study tap education funds from a wide range of sources. Personal income, savings, and employment represent just more than half (51%) of the funding sources. While participants relied on their parents for funding support, nearly all the parental support came from income and savings (19% of all funding sources). Only a fraction of the parental support came from parents’ retirement and home equity (1%).

Page 11: Methodological and Technical Appendix

Methodological and Technical Appendix

11UNC CENTER FOR COMMUNITY CAPITAL

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

90%

< $1

0,0

00

$10

,00

0-$

24,0

00

$25,

00

0-$

34,0

00

$35,

00

0-$

44,9

99

$45,

00

0-$

54,9

99

> $5

5,0

00

Associate’s Degree

Bachelor’s Degree

Advanced Degree

COMPARISON OF CCC DATA TO NATIONAL DATA:

SummaryAs is common in exploratory qualitative research, the CCC sample was not intended to be representative of the general population. Thus, the goal of the CCC sampling and recruitment strategy was to seek a reasonably diverse sample, but not necessarily to achieve equal representation of all minority groups or for equivalence between the CCC sample and the nationally-representative NPSAS sample. Nonetheless, a few parallels between the two samples are worth highlighting.

1. Gender: Overall, females are adequately represented in both the CCC sample and the NPSAS sample (undergraduate and graduate combined).

2. Age: The CCC sample includes participants of varying ages, though the upper age limit of the NPSAS sample

is significantly higher.

3. Oversampled Groups: Both samples have adequate representation of minority groups. However, the CCC sample includes a slight over-representation of several minorities and often under-represented groups such as first-generation college students, people of color (particularly, Hispanic/LatinX and American Indian/Alaska Native individuals), military veterans, and those with masters or professional degrees.

Part 3: Summary

Own income/savings

Employment

Parent’s income/savings

Work study programs

Veterans programs

Other relatives

Parents’ retirement/home equity

Other

29%

22%19%

15%

4%

6%

3%

1%

Most CommonEducation Funding

Sources

Figure 13. Other sources of education funding in the CCC sample (non-loan, grants, & scholarships)

Page 12: Methodological and Technical Appendix