every dollar matters: examining young alumni …...donor predictor model findings tested 18...

Presented by:Meredith S. Billings

Tufts UniversityOffice of Institutional Research & Evaluation (OIRE)

November 2009

EVERY DOLLAR MATTERS: EXAMINING YOUNG ALUMNI GIVING BEHAVIOR

Office of Institutional Research, November 09

Agenda

Background/IntroductionResearch QuestionsParticipants

Sample demographics

InstrumentMethodology

Donor Predictor Model Donor Comparison Profiles

FindingsConclusion & Future Research Questions


BACKGROUND


Institutional Background

Research University – Very high activity4-year Liberal Arts & Engineering undergraduate schools5,164 undergraduates; 5,088 graduate/ professional students; 10,252 total enrollment


School of Arts & Science

36%

School of Engineering

13%

School of Medicine19%Tisch College of

Citizenship & Public Service

5%

School of Dental Medicine

3%

The Fletcher School8%

Friedman School of Nutrition Science

and Policy4%

Cummings School of Veterinary

Medicine8%

Cross University Initiatives

4%

Background for Research Study

Beyond Boundaries: The Campaign for TuftsUniversity-wide initiative to raise $1.2 billionStarted in Fall 2006

Goals per School


Purpose of Study

Inform Tufts Funds’ young alumni fundraising strategy for remaining years of the CampaignUnderstand priorities and motivators of recent graduatesCreate positive experience for respondentsIdentity future opportunities for research


Significance of Study

Young alumni typically have lower participation rates of giving than older alumni (Untapped resource!)Too large of a population to ignore

Young alumni are 23% of alumni in the Schools of Arts, Sciences, and Engineering at Tufts

Current & future giving is correlated with past giving behavior (Okunade & Justice, 1991; Lindahl & Winship, 1992)

Need effective strategies to target potential young alumni donors


RESEARCH QUESTIONS


Main Research Questions

What are the current motivators and priorities for young alumni?

Are there differences in demographics, attitudes, and experiences between donors and non-donors? Among different type of donors?

To what extent can the advancement office use these characteristics to provide a more targeted strategy for fundraising?


PARTICIPANTS


Participants

Alumni from the last ten graduating classes (Classes 2008 - 1999), N = 1,405

DemographicsSex: 62.2% Female, 37.8% Male

Graduation Years: 59.1% Classes 2008-05; 26.0% Classes 2004-02; 14.8% Classes 1999 -2001

Location: 25.6% Boston, 19.9% NYC, 8.6% California, 7.8% Washington D.C., 1.1% Florida, 37.0% Other

Donor status: 60.0% Donated in last six years


INSTRUMENT


Young Alumni Instrument

Developed by OIRE and University Advancement25 Items

Satisfaction with undergraduate and alumni experiencesGauge level of interest in particular alumni events and activitiesEvaluate motivations to give along with satisfaction with the giving processAssess likelihood of donating to particular areas/funds for the university


Selected questions

Low priority, 36.9%

Mid priority, 41.7%

High priority, 14.2%

Top priority, 2.0% N/A, 5.2%

How would you rank Tufts on your list of philanthropic priorities?


Selected questions

I do not support charities/causes

Religious

Arts & Culture

Political

Environment

Human Service

Healthcare

Education

7.5%

19.8%

30.6%

33.7%

35.0%

36.3%

36.4%

45.0%

Excluding Tufts, what causes/charities do you support? (% Yes)


METHODOLOGY


Methodology

Donor Predictor ModelBinary Logistic Regression

Young Alumni (coded as 1 for “yes” and 0 for “no”)

SPSS 15.0 with MLEData for Predictors

Young Alumni Survey ResponsesAdvancement ‘s Giving Information DatabaseTufts Data WarehouseNational Student Clearinghouse Enrollments


Logistic Regression

Fitting this model (mathematically):

Fitting this model (in words):Odds of an event occurring (i.e. alumnus/a being a donor) for given values of the predictor variables

nnxxx βββα ++++=⎟⎠⎞

⎜⎝⎛ ...

P - 1Pln 2211


Variables for Donor Predictor Model

Predictor Data Source Literature

Sex Advancement Database Eckel & Grossman (1998) WomenDugan et al. (2000)Clotfelter (2003)Monks (2003)

In reunion year Advancement Database Wunnava & Lauze (2001)

Distance from institution Advancement Database Hueston (1992) NearVolkwein & Parmley (1999)Dugan et al. (2000)

Financial aid Office of FinancialServices

Clotfelter (2003) LoansDugan et. al (2000) Grants, LoansMonks (2003) Loans

College experience Survey Harrison et al. (1995)Volkwein & Parmley (1999)Clotfelter (2003)Monks (2003)

GPA Data Warehouse Dugan et al. (2000)

Major Data Warehouse Dugan et al. (2000) Math/Engr/EconMonks (2003) Bus/Engr/Math/Soc Sci

Multiple Degree Holder Advancement Database Okanade et al. (1994)Volkwein & Parmley (1999)Dugan et al. (2000)

Education in progress National Clearinghouse Volkwein & Parmley (1999)

Charitable behavior Survey Dugan et al. (2000)


Methodology

Donor Comparison ProfilesDonors were classified into five categories based on amount, frequency, and how recently alumni donated

Current Year = donated in current FY 2009 excluding Annual donorsLYBUNT = “Last Year But Not This Year”SYBUNT = “Some Years But Not This Year”Annual = given every year since graduatedMajor = given $500 or more in their lifetime


Methodology

CY, LYBUNT, SYBUNT, & Annual Donors comparedMajor Donor to All Donors compared

Conducted ANOVAs, independent sample t-tests, and chi square tests to compare demographic and attitudinal data


FINDINGS


Donor Predictor Model Findings

Tested 18 predictor modelOverall model significantTen n.s. predictors and eight significant predictorsRemoved 10 n.s. predictors and reran model

Tested 8 predictor modelOverall model and predictors were significantModel correctly predicts whether an alumnus is a donor 82.0% of the time and non-donor 56.6% of the time for all overall success rate of 72.1%Model accounts for 26.4% of the variance in the outcome variable, donor statusHosmer & Lemeshow test was n.s.


Donor Predictor Model Findings

Ten Non Significant PredictorsDistance from institution in milesDouble Jumbo statusReceived undergraduate grantsReceived undergraduate loansSelf reported “active” alumnus/aUndergraduate major by disciplinesCurrently enrolled in higher educationConsidering oneself “philanthropic”Attended HomecomingVolunteering for Tufts in some capacity

1.33

1.36

2.00

0.79

1.26

0.91

0.96

1.28

1.56

1.70

0.50

0.43

0.85

2.08

0.61

1.03

0.83

1.36

0.74

1.02

0.88

1.17

0.86

0.80

1.23

1.10

1.12

1.07

Female*

Years Out*

Living in/around Boston*

Living in/around NYC

Living in/around DC

Living in/around FL

Living in/around CA

Overall undergrad exp*

Attended alumni event*

Rank of Tufts in "giving priorities"*

GPA 2.00 ‐ 2.49

GPA 2.50 ‐ 2.99*

GPA 3.00 ‐ 3.49

In reunion year*

Live 100 miles or less from Tufts

Live between 249‐101 miles from Tufts

Double Jumbo

Received grants

Received loans

Active alumnus/a

Social Science

Natural Science

Arts & Humanities

Interdisciplinary

Currently enrolled

Attended homecoming

Philanthropic

Volunteer for Tufts

Figure 1. Odds Ratio for Young Alumni in the Full Donor Predictor ModelMore Likely to DonateNo EffectLess Likely to Donate


Eight Significant Predictors

Demographics (2)Sex, Residency

Undergraduate Experience (3)Overall rating of undergraduate experience, number years out from graduation, GPA

Current Alumni Engagement (3)In reunion year, attended at least one alumni event (excluding Homecoming), rank of Tufts in philanthropic priorities


Final Donor Predictor Model Findings


Final Donor Predictor Model Findings

Significantly Increases OddsFemaleLiving in/around BostonIn reunion yearAttended at least one alumni eventSatisfied with undergraduate experienceRanked Tufts high in philanthropic prioritiesEach year out from bachelor’s degree

Significantly Decreases OddsEarning undergraduate GPA between 2.50 –2.99


Donor Comparison Findings

Different types of Donors were compared on:SexAgeDistance from institution (in miles)College AffiliationGPATotal amount of grants receivedTotal amount of loans receivedTotal financial aid receivedAverage donation in last five yearsAverage lifetime donation



Significantly higher percentages of Annual Donors willing to donate to:

Unrestricted funds (58.8%)Faculty salaries (38.2%)Student activities (61.8%)Endowment Funds (64.7%)

Significantly lower percentages of SYBUNTs willing to donate to:

Unrestricted funds (22.0%)Endowment funds (29.5%)


Donor ComparisonFindings

Significantly higher percentage of Major donors willing to donate to:

Unrestricted funds (59.3%)Endowment funds (59.3%)


Limitations of Study

Not collecting income information from young alumni

Monks(2003) and Melchiori (1988) found personal and household income to have significant effect on donating behaviorHowever, Melchiori noted that only 84% of alumni answered personal-income questions and 65% answered household-income questions

Compared to almost perfect response rates for demographic questions on same survey


CONCLUSION&

FUTURE RESEARCH


Conclusions & Future Research

Set of eight characteristics that Tufts Fund staff can use to identify and target potential donorsRemove alumni unlikely to donate from solicitation lists to save scarce office resourcesNot many significant differences among donors

More research is needed to understand why some young alumni are consistent donors and some are only occasional donors


QUESTIONS ???


Contact Information

Meredith S. Billings, Research Analyst, Office of Institutional Research & Evaluation, Tufts University, [email protected]

Slides will be available at:http://institutionalresearch.tufts.edu/?pid=6&c=4

every dollar matters: examining young alumni …...donor predictor model findings tested 18...

Documents