Generalization of The Results of The Undergraduate Experience Survey, Complementary Form, Spring 2002 to The Spring 2002 Undergraduate Student Body
One issue is the extent to which the Ucues-CF survey results can be generalized to spring 2002 undergraduates. Although the Ucues-CF and the Ucues-Santa Barbara were given to all undergraduates when taken together, each was not given to a representative sample of undergraduates. The Ucues-CF was given to all spring 2002 undergraduates except those given the Ucues-SB. The Ucues-SB was given to the following groups:
- new freshmen in Fall 2001
- new transfers in Fall 2001
- new freshmen in Fall 1998
- new transfers in Fall 2000
Accordingly, the population surveyed by Ucues-CF consists of all spring 2002 registered undergraduates except for those who were new freshmen in the falls of 1998 or 2001 or who were new transfers in the falls of fall 2000 or fall 2001. Although strictly speaking using standard experimental design sampling methodology, the results from the Ucues-CF can only be generalized to the group given the survey, a case can be made that the results obtained from Ucues-CF survey may be viewed as representative of how the spring 2002 undergraduates would reply.
One the reason for saying this is that the Ucues-CF sample was large and overlapped considerably with the spring 2002 undergraduates. Of the 22,964 spring undergraduates, 13,711 were included in the Ucues-CF survey. That is, there is almost a 60% overlap between the Ucues-CF sample and all spring 2002 undergraduates. The response rate of the survey was 34%, i.e., 4,6790 replied.
A second reason for saying that the Ucues-CF is representative of the spring 2002 undergraduate student body is a series of comparisons showing similar compositions of both groups for different key variables as illustrated in the following tables.
|Ucues-CF||Spring 2002 UG|
The ethnicity variable was recoded to a small number of groups. (1) The Underrepresented group contains African American, American Indian, Chicano, and Latino. (2) The Asian group includes Filipino. (3) The White+ group includes Other and Decline to State. (4) The International group contains International students.
|Ucues-CF||Spring 2002 UG|
|2nd Deg Ltd||1.5||0.4|
|Letters and Science||79.8||80.7|
Distributions within the variables listed above show considerable similarity. There are more females than males in both the survey and the population. Categories within class level have similar ranks in that seniors are the most populous and freshmen the least. Categories within ethnicity are close to having similar ranks. There are far more freshman entrants than transfer entrants in both the survey and the population. The College of Letters and Science has about 80 percent of the persons in both the survey and the population.
A third reason for saying that the Ucues-CF results generalize to the spring 2002 undergraduate population is that a cohort-by-cohort analysis shows considerable similarity of response across cohorts. The largest difference between the Ucues-CF and the spring 2002 undergraduate population is that the Ucues-CF has very few persons who entered in either fall 2001 or in fall 1998. Accordingly, an analysis of the key question, i.e., Question 48 will be done by cohort. Whether a person enters as a freshman or as a transfer can have a bearing on their survey replies so this will also be used in the analysis.
The following table shows the proportion of respondents to the question "When you graduate, do you think you will join an alumni association?" by entering cohorts of new freshmen. Cohorts with fewer than 50 members were eliminated from the analysis.
There are several interesting results in the above table. One is that the percentage of both yes and no replies increases for the more recent cohorts. That is the longer persons are at Berkeley the more certain they become of whether or not they will join an alumni association. For example, .46 of those entering in fall 1997 are not sure whereas that number jumps to .63 for those entering in the spring of 2002. The longer students are at Berkeley the more certain they are that they will join an alumni association as the proportion of yes replies goes from .27 for those entering in spring 2002 up to .39 for those entering in fall 1997. The proportion indicating that they will not join an alumni association also increases, for example, from 0.09 in spring 2002 to .15 in fall 1997. As these new freshmen progress through Berkeley, their degree of uncertainty decreases.
Because of the procedure used in constructing the Ucues-CF sample, the usual rational for linking the Ucues-CF sample to the population of spring 2002 undergraduates needs to be supported by additional analyses. The additional analyses showed that (a) there is a considerable, i.e., 60%, overlap between the Ucues-CF sample and the spring 2002 undergraduate population; (b) within both the Ucues-CF replies and the undergraduate population there are considerable similarities of distribution within key demographic and academic variables; and, (c) a cohort-by-cohort analysis for new freshmen shows a uniform effect across cohorts. For practical purposes, it appears as if the Ucues-CF results are representative of the spring 2002 undergraduate population.