S OME seventeen years ago a study was conducted in order to investigate . . attributes [considered to be] desirable in a mate, preferred age at marriage, and number of children hoped for.' Questions which pertained to these points were asked of 600 students at a midwestern state university. Since the year 1939, when these data were gathered, some rather sweeping changes have occurred in age at first marriage, in fertility patterns, and in several other characteristics of the American family.2 Given these changes, it seems reasonable to expect that corresponding modifications might have taken place in the values of young adults regarding marriage and the family. Repetition of this study would provide certain rudimentary information about the direction and magnitude of these changes. Naturally, this information would be applicable only to the investigated population, but at the same time it might yield suggestions for further research and for the construction of theory. Even if such a study should yield no important differences from the original, it would still serve a valuable purpose, that of replication.3 Despite the fact that it suffered several methodological shortcomings, the original study has been widely cited, particularly in texts on preparation for marriage. Among these shortcomings were the facts that the investigated population was neither defined clearly nor sampled randomly. Data were collected in a classroom through the medium of a paper-and-pencil device. Certainly a question may be raised as to the reliability of this approach.4 For these reasons, independent replication of the original study should have some value virtually as an end in itself. With these considerations in mind a repeat study was designed and carried out at the same university as in the original. It must be noted at the outset that, aside from the difference in time, it was impossible to duplicate every condition of the 1939 research. In the original, respondents were students who participated in a noncredit course in preparation for marriage. 600 cases represented approximately five percent of the 1939-40 enrollment at the university. researcher implied that, with these cases, he sought a representative sample of the student body, but that he did not achieve it in every respect.5 It would be impossible to employ a similar group today because the course is no longer offered in its 1939 form. It was decided, as a roughly similar alternate procedure, to define the population explicity as all students enrolled in the university in May 1956, and to draw a one percent sample from this population.6 This procedural difference between the two studies clearly cannot be overlooked in interpreting any differences in outcome between them. * This study was supported in part by the Research Committee of the Graduate School, University of Wisconsin, from special funds voted by the State Legislature. research was conducted jointly by the writer and several of his students in an advanced course. Appreciation is extended to these co-workers as are regrets that, because of space limitations, their names cannot all appear on the credit line. Paper read before the Midwest Sociological Society, Des Moines, Iowa, April 5, 1957. 1 Reuben Hill, Campus Values in Mate Selection, Journal of Home Economics, 37 (November 1945), pp. 554-58. Quote from p. 554. 2 For documentation, see William F. Ogburn and Meyer F. Nimkoff, Technology and the Changing Family (New York: Houghton Mifflin Co., 1955), Part II. 3 For a discussion of replication, see F. Stuart Chapin, Experimental Designs in Sociological Research (rev. ed.; New York: Harper and Brothers, 1955), pp. 187-89 and Raymond W. Mack, The Need for Replication Research in Sociology, American Sociological Review, 16 (1951), pp. 93-94. 4 equally important question of validity cannot even be raised here. With the data at hand, there is no way of knowing the extent to which check-marks on paper correspond to actual values. It must be assumed that there is a positive relationship of at least moderately high magnitude. 'Reuben Hill, op. cit., p. 554. 6A systematic sample of 150 names was drawn from the university student directory which contained names of all regularly enrolled students together with the names of some who had dropped out of school. Questionnaires were mailed and 120 returned. Subsequent examination showed that 14 of the nonrespondents had dropped out of school. This means that 12 percent of the eligible respondents failed to return completed questionnaires. There was no evidence that this was a nonrandom subgroup.