A study conducted last spring and recently released to the Record indicates a strong link between Williams’ athletic success and the amount of alumni giving. The authors, Rob Houle ’01 and Tim Sullivan ’01, both of whom start for Williams football and are economics majors, conducted the study as part of their “Economics of Higher Education” class with Gordon Winston, a professor of economics.
Houle and Sullivan, using various econometric techniques, were able to compare the levels of several different types of alumni giving to win-loss records for Williams sports over the past 50 years. The different categories of alumni giving included annual donations to the Alumni Fund, total annual Gifts and Grants, and Average Gift.
The records, collected from athletic department files, encompassed 10 men’s (hockey, football, basketball, baseball, lacrosse, soccer, squash, tennis, swimming and wrestling) and nine women’s sports (basketball, field hockey, lacrosse, soccer, softball, squash, swimming, tennis and volleyball). Though records were available for men’s sports for the entire time period, records for women’s sports ranged from six to 27 years worth of data.
In addition to comparing the various individual and aggregate categories to alumni giving, Houle and Sullivan accounted for several other variables. In order to gauge the health of the economy in various years, and thus its separate effect on giving, they utilized yearly values of the Dow Jones Industrial Average. The Dow, it turned out, was the largest contributor to fluctuations in alumni giving.
Also, Houle and Sullivan found that fraternities had a negative effect on alumni giving, meaning that donations rose after they were abolished, and that the move to coeducation sparked an increase in giving as well.
“General alumni feelings towards frats were negative, I think,” said Houle. “There was a large push in the ’70s to get rid of frats, and when that happened, it made a lot of alumni happy, and may account for some of the increase in giving.”
In the study, the authors cautioned strongly that “correlation does not prove causation.” In this context, that statement is meant as a warning that increased alumni giving may cause athletics success, and not vice versa, as the study implied.
To defeat this caveat, Houle and Sullivan employed a time lag, running multiple regressions (correlations) for every category with lags of zero to four years. Their strongest relationships came from the two-year lags.
“[For the most successful tests] we were running regressions with a two-year lag,” said Sullivan. “In other words, we would be comparing athletic results in 1996 with the alumni giving in 1998.”
“That was done by simply shifting the data columns, which is used all the time in economics,” added Houle.
For those unfamiliar with economics, the pair explained that data comparisons using the statistics program “E-Views” produce various results, among them an R-squared term. This term reveals how well the variables in the equation account for the results in alumni giving. For example, an R-squared term of .65 means that a model explains 65 percent of the fluctuation in the data being examined.
For the two-year lags, the duo found values of 0.91, 0.91 and 0.83 for the 50-year span of men’s sports and Alumni Fund, Average Gift and Total Gifts and Grants, respectively. In other words, according to this study, an increase in athletic success in 1996 would account for 91 percent of the change in alumni giving between 1997 and 1998.
For women’s sports over the period from 1978 to 1999, they found all three values to be 0.95. Men’s sports over the same span produced values of 0.98, 0.98 and 0.91. The rest of their regressions ran in a similar vein, with the results falling into a range of 0.85 to 0.95. In general, Houle and Sullivan explained, economists accept studies with R-squared values of 0.7 or 0.8 as reasonably acceptable. Their values, they admitted, were somewhat high, implying some inexplicable deficiencies in their model.
“In any econometric model, you can’t factor in all of the error terms [which show the level of fluctuation that cannot be explained],” said Houle. “We were only able to include four explanatory variables, which is a very small number. Williams having high alumni giving probably does have a relationship to athletic success. However, it is probably not determined as completely by athletic success as our numbers indicate.”
Dick Quinn, sports information director, who deals directly with alumni seeking sports results, believes that athletic success contributes to athletes enjoying their time at Williams, and thus to their level of alumni giving. “Athletes have a very positive experience at Williams, and they want to make sure that experience continues for those who follow them.”
Sullivan emphasized that, while athletics are an important benchmark because the data is so readily available, there are other things that contribute to high levels of alumni giving. “I would say that we used athletics because you can gather data easily, and that worked for us. You might be able to do the same study measuring theater ticket sales to alumni, or another, similar dataset, and measure the positive effects of other pursuits of the student body on people’s experience at Williams and their willingness to contribute after they leave.”