Who Gets theFellowship?
GRE General Test scores can play a useful rolein the awarding of on-campus fellowships when used in conjunction with other relevantdata, such as grade point average, faculty recommendations, and related course work.There are a number of strategies for using GREscores in the fellowship selection process.Two strategies are described below.
The Fellowship SelectionProcess at Institution X
Institution X offers graduate degrees in a widerange of disciplines and routinely processes asubstantial number of fellowship applicationseach year. In the selection process, Institution X places a heavy emphasis on GRE GeneralTest scores and compares an applicant’s verbal, quantitative, and analytical scores tothe mean for the total group applying to theinstitution. For students applying to Institution X, the mean verbal score is 470, the meanquantitative score is 570, and the meananalytical score is 480.The figure that follows shows the percentageof students in the English literature, psychol-ogy, and engineering departments who meetor exceed the institutional mean scores for thetotal group for each of the three General Testmeasures. Sixty percent of the engineeringstudents fall into this group, as do 50 percentof the psychology students and 30 percent of the English literature students. On this basisalone, Institution X might decide to awardproportionally more fellowships to engineeringstudents than to those in the psychology orEnglish literature departments. In doing so,however, Institution X is favoring studentsfrom disciplines that, as a rule, have higherGeneral Test scores; thus, the institutionmay be introducing a bias that they are notovertly aware of.
An Alternate Approach
However, as an alternative, Institution X might want to consider how their students in eachof the three departments perform relative tonational mean scores
. As thefigure that follows indicates, 50 percent of theirEnglish literature students submit General Testscores that meet or exceed the national meanfor students in that particular field. Forty-fivepercent of their psychology students meet orexceed national means for students in theirfield, as do 40 percent of their engineeringstudents. Clearly, then, comparing students toothers within a particular field of study paintsa very different picture than when students arecompared to the total group of institutionalapplicants. By making the within-field com-parison, Institution X would be giving fellow-ship awards to students who have relatively high test scores within their field of study.
Use Scores with Other Data
The GRE Board encourages graduate schools,departments, and fellowship selection commit-tees to consider GRE scores as a meaningfulsource of information about an applicant’s chancefor success in graduate school. GRE scores arebest used, however, in conjunction with othersources of information about the applicant. As the alternative strategy suggests, institu-tions are also encouraged to look at GREscores within the context of the applicant’sfield of study. In this way, students arecompared to what may possibly be moreappropriate groups, and the fellowships are