1 Hedegard, James M., The Impact of Legal Education: An In-Depth Examination of Career-relevant Interests, Attitudes, and Personality Traits Among First-Year Law Students, 1979 A.B.F. Res. J. 791 [hereinafter cited as Hedegard I].Google Scholar

2 Id., The Course Perceptions Questionnaire: Development and Some Pilot Research Findings, 1981 A.B.F. Res. J. 463 [hereinafter cited as Hedegard II].Google Scholar

3 Hedegard I, supra note 1.Google Scholar

4 Hedegard II, supra note 2.Google Scholar

5 Examples of each kind of change during the first year among the BYU sample are as follows:. 1) Some students appear to have changed their conceptions of administrative law. 2) Some students apparently came to see solo practice as centrally involving work on various “family law” problems. As a result, students who entered law school with a strong interest in solo practice but a strong dislike of family law could have ended the year with a weakened interest in solo practice. Some students with interests in poverty law at entrance apparently also felt that federal agencies were especially effective settings within which to work on legal problems of the poor. By the end of the year, some of these students came to feel that state and local governments were more effective in this area. For this reason, their interests in federal agency careers may have waned. 3) Students interested in constitutional law at entrance may have had such interests because they had more general interests in abstract, conceptual problems. If, during the first year, they came to see themselves as less interested in abstraction than they thought at entrance, and if they continued to perceive constitutional law in the same way, their interests in constitutional law might have diminished. 4) A faculty member much admired by many first-year students stressed the equivalent moral “value” of two aspects of legal practice (criminal defense and criminal prosecution) that at least several students had tended to evaluate somewhat differently. Our data suggest that students' admiration of the instructor was tied to changes of opinion regarding the “value” and hence of relative interest in the two forms of practice.Google Scholar

6 See table 2 at pp. 802–3 infra.Google Scholar

7 All students may have had a particular professor who kindled or dampened the interests of all in a particular content area, many students may have attended a particular lecture by an outside lecturer or a symposium on a special topic, the pervasive ethos of the school may have tended to draw student interests in a particular direction, and so forth.Google Scholar

8 E.g., the interest focusings mentioned earlier.Google Scholar

9 The variance of a set of scores is the mean squared deviation of individual scores around the mean score. The square root of the variance is called the standard deviation.Google Scholar

10 Hedegard I, supra note 1, at 820 table 3, 825 table 4, with discussion of dispersion data at 823.Google Scholar

11 See table 3 and accompanying text at p. 804 infra.Google Scholar

12 Hereinafter I will refer to Brigham Young University as BYU and to the student group we are studying as the BYU students or the students.Google Scholar

13 We used the random student numbers as code numbers to provide anonymity and confidentiality. All student data were received and processed with only the code numbers for identification. Students affixed their own code numbers to response sheets. The only links between the code numbers and student names were (1) a list kept in the dean's office at BYU and (2) students' personal records of their own code numbers. The American Bar Foundation has no copy of the list. Thus we could link student responses to code numbers, and the BYU dean's office personnel could link code numbers to student names, but neither could link student names to their responses. The name-code number list was maintained so that students completing questionnaires could obtain their code numbers if they had forgotten them and dean's office personnel could send information such as grades identified to us only by code numbers.Google Scholar

14 The design, content, and scoring of the CPQ is fully explained in Hedegard II, supra note 2.Google Scholar

15 Personality characteristics are considered by psychologists to be relatively enduring behavioral patterns or cognitive and emotional qualities, more general than specific behaviors. Viewed as less easily changed than specific behaviors, personality characteristics are viewed as changeable by intense experiences (such as religious conversion experiences or military combat) and by intense environments (such as university environments). For a listing of the OPI traits, see Hedegard 1, supra note 1, at 811–12 table 1.Google Scholar

16 In what follows I will use interest to refer both to interest in areas and to acceptability of work setting.Google Scholar

17 This article focuses on interest scale changes. For information on response changes on the individual interest items, see Hedegard I, supra note 1, at 820.Google Scholar

18 These are not actual items but are similar in form and content to items on the OPI.Google Scholar

19 Hedegard I, supra note 1, at 811–12 table 1.Google Scholar

20 For descriptions of all the OPI traits making up these factors see id.Google Scholar

21 The assignment of faculty members to course sections is depicted in the table below. The 11 first-year instructors are numbered 1–11 in the table. The five major courses are labeled A-E.Google Scholar

22 Although it was for the sake of objectivity that these measures of a student's experiences of instructional techniques were based on course section mean scores, these scores also permitted measuring the experience of students who did not complete the CPQ in all course sections, since course section mean scores could be calculated on the basis of the other students' CPQ responses. Nearly all students completed the CPQ, however, in all course sections.Google Scholar

23 A regression analysis yields correlations between “predictor” variables and a “criterion” variable. Such correlations do not mean that the qualities measured by the predictor variables cause the qualities indicated by values of the criterion variable. Causality could run in the reverse direction, or both qualities could be the effect of yet another, perhaps unknown, cause. Nonetheless, a regression analysis suggests possible causal relations, or the points at which some criterion-affecting causal agent appears to be operating. Later in this article I will offer speculations about causes of interests change which may serve as hypotheses for other researchers. When a series of studies firmly establishes a causal relationship, it should be possible to craft law school environments, or student selection procedures, which can more strongly or more reliably produce desirable changes in students (assuming agreement on what is desirable).Google Scholar

24 For example, a positive correlation between students' interests in, say, constitutional law and the degree to which the students were exposed to the Socratic method (as measured by the CPQ Socratic method index) would indicate that the more exposure students had in their first-year classes to the Socratic method, the more positive (or less negative) their changes in interest in constitutional law. For such a correlation to be possible, some students would have to have had more positive (or less negative) changes in interest than did other students, and some students would have to have had more extensive exposure to the Socratic method than did other students.Google Scholar

25 Those who would like to find out “who done it” (i.e., the relative impacts of the various kinds of predictor variables on end-of-year interests) can look at table 11 at p. 837 infra. The rest of this article deals with the “how” and the “why.”.Google Scholar

26 The at entrance and end-of-year mean scores were linearly transformed into scores that would permit easy scale-to-scale comparisons. On each of the resulting scales, a score of +4.00 indicates that all BYU students found all areas of law included in the scale to have “strong interest” and/or work settings included in the scale to be “highly acceptable.” A mean score of 1.00 on a scale indicates, on the contrary, that all BYU students indicated “no interest” in the constituent law areas and/or found the work settings “not acceptable.” Mean scores for individual interest items (including mean change scores) are shown in Hedegard I, supra note 1, at 820 table 3, 825 table 4.Google Scholar

27 Id. at 838.Google Scholar

28 To understand the correlations in table 3, one can imagine a group of ten students (A-J) and consider their scores on a particular interest scale at the beginning and at the end of the first law school year, as in the illustrations below.Google Scholar

For each student in illustration I the scale score at entrance was identical with the score at the end of the year. Thus the correlation between scores at entrance and at the end of the year was + 1.00.Google Scholar

Although scale scores at entrance and at the end of the year were identical for some students in illustration II, they changed for other students. Some students (C, G, and J) reported drops in interest, others (D and H) reported rises in interest, while the rest (A, B, E, F, and I) reported no change. This set of scores yielded a correlation between scores at entrance and at the end of the year of + 0.68. This correlation coefficient indicates that students who at entrance had strong interests relative to other students tended to have relatively strong interests at the end of the year; students who began the year with moderate interests relative to other students tended to end the year with relatively moderate interests; students who began the year with relatively weak interests tended to end the year with relatively weak interests.Google Scholar

The word tended has been italicized to emphasize the fact that although some students maintained their initial level of interests relative to other students, other students did not. That some students changed their relative positions within the group of ten is reflected in a correlation coefficient that was not + 1.00 but + 0.68.Google Scholar

Although all students in illustration III reported loss of interest, each student ended the year with the same relative level of interest throughout the group: each student maintained his initial position. Thus the correlation between scores at entrance and at the end of the year was + 1.00.Google Scholar

As in illustration III, the mean interest scale score of students in illustration IV dropped from 3.00 to 2.00. Also, the variability among student scores dropped over the year: the standard deviation among student scores dropped from 0.82 at entrance to 0.41 at the end of the year. Despite these two kinds of changes over the year, each student maintained the same relative level of interest: the difference in interest scores between student B and student E at entrance was the same size as the difference in interest scores between student E and student J. At the end of the year, the score difference between students B and E was also the same as between E and J. Because the relative scores of students were identical, in this sense, the correlation between scores at entrance and at the end of the year was + 1.00, as in illustration III.Google Scholar

Although the final mean interest score was identical to the mean score at entrance (3.00) in illustration V, scores at the end of the year for individual students had no consistent relationship with scores at entrance. One of the three students who had scores of 4 at entrance maintained this interest score through the year, but the second reported a drop in interest with a score of 3, and the third reported a more marked drop in interest, with an end-of-year score of only 2. Similar diverse changes in score occurred among students with “average” interest scores at entrance (students D, E, F, and G) and among students with relatively “low” interest scores at entrance (students H, I, and J). Because of such diverse interest score changes, end-of-year scores bore no consistent relationship to scores at entrance, reflected in a correlation between scores at entrance and at the end of the year of 0.00.Google Scholar

29 In this article the reliability coefficient describes the correlations between two measures of the same quality at two points in time. Reliability expresses the extent to which the measuring device, an interest scale in this instance, yields the same score for the same individual at both times (more precisely, the extent to which the measuring device duplicates the scores of individuals relative to other individuals. See, especially, illustration IV in note 28 supra).Google Scholar

30 This proportion of accountable end-of-year score variability is the square of the reliability coefficient in the left-hand data column for the same scale. Thus, for the corporate law scale, the proportion of accountable end-of-year score variability is 0.48 (which is 0.69²). Slight discrepancies between the two columns in table 3 are due to rounding of numbers.Google Scholar

31 Thus, for example, the proportion of end-of-year score variability on the corporate law scale not predictable from, or consistent with, score variability at entrance is 0.52 (1.00-0.48 = 0.52). Description of end-of-year scores as “consistent with” or “predictable from” scores at entrance is a statistical notion, illustrated in note 28 supra. In the formal statistical sense, the “perfect” correlation of + 1.00 in illustrations I, III, and IV implies that through linear regression techniques end-of-year scores were completely consistent with, and perfectly predictable from, scores at entrance. Again, “complete consistency” means that the interest scores of a pair of students relative to each other at the end of the year were the same as their relative scores at entrance. In fact, one could generate a linear equation to generate (or “predict”) a student's end-of-year score from a student's score at entrance. Because of the consistency between relative scores at entrance and relative scores at the end of the year in illustrations I, III, and IV, one could construct linear equations that would generate, or “predict,” the actual end-of-year scores in each of these illustrations. However, the linear equation that would most accurately predict end-of-year scores from scores at entrance in illustrations II and V would not generate the actual end-of-year scores for all students. In illustration II, for example, the linear equation that would generate end-of-year scores as similar as possible to the actual end-of-year scores would yield the scores shown in the following table:Google Scholar

The correlation between entrance and end-of-year scores in illustration II was + 0.68. Squaring this correlation coefficient, my reliability coefficient, yields 0.48, which is the proportion of end-of-year score variability predictable from the scores at entrance. Thus, 0.52 is the proportion of end-of-year score variability that cannot be accounted for by, or predicted from, the scores at entrance. This latter proportion is reflected in the discrepancies in the table above between actual end-of-year scores and end-of-year scores predicted from my linear equation. If one squares each of the difference scores shown in the right-hand column of the above table, adds these squared differences together, and divides the total by 9 (the number of students minus 1), one obtains a measure of the variability of actual scores around the predicted scores. The variability index turns out to be 4.733/9 = 0.52, which is about 52 percent of the total variability among actual end-of-year scores (0.99).Google Scholar

32 Random or chance events that might temporarily change a student's responses to an interest item (and thus change the student's score on an interest scale) could include interest-specific experiences having only a temporary effect on the student's response to the item. Suppose, for example, that at entrance into law school a student felt “moderate” interest in tort law which persisted generally through the first year and then just before taking the LSI near the end of the second semester, read and discussed an especially interesting tort law case. Such a student's interest in tort law might be raised temporarily and declared a “strong” interest on the LSI. Although this student's score on the tort law scale would be raised relative to some other students' (who had not been so struck by the case or who already had “strong” interests in tort law), this change in interest could be temporary, later reverting back to the more permanent “moderate” level of interest. For a discussion of such phenomena, see L. L. Thurstone & E. J. Chave, The Measurement of Attitude: A Psychophysical Method and Some Experiments with a Scale for Measuring Attitude Toward the Church (Chicago: University of Chicago Press, 1929). Students can also make errors in recording their responses to questionnaire items by checking the wrong boxes, and occasionally coding and keypunching errors occur despite elaborate checking procedures.Google Scholar

33 If focusing of interests is a process in which lesser interests fall away while stronger initial interests are reinforced during the first year, then the initially stronger interests determine their own continuation and the decline of other, competing interests.Google Scholar

34 These characteristics can also be viewed as noncentral determinants of behavior, for psychologists are in considerable disagreement over the centrality and utility of personality concepts, some arguing that situations are primary determinants of behavior, with personality characteristics relatively unimportant. For a discussion of the interaction of personality characteristics and social aspects of situations as determinants of behavior, see David Marlowe & Kenneth J. Gergen, Personality and Social Interaction, in 3 Gardner Lindzey & Elliot Aronson, eds., The Handbook of Social Psychology 590 (2d ed. Reading, Mass.: Addison-Wesley Publishing Co., 1968).Google Scholar

35 Table 4 indicates the extent to which the personality characteristics students brought to law school are related to the changes in interest scores of some students relative to the scores of other students. For example, the positive correlation between scores on the antitrust scale and on the intellectuality factor means that students who entered law school professing strong interests in intellectual abstraction and scientific problem solving were more likely than students with weaker intellectual and problem-solving interests to maintain or increase their interests in antitrust law during the first law school year. Thus, differences in intellectual interests among entering students could account for some of the differential changes in interests in antitrust law during the first law school year. The negative correlations between scores on the tort law scale and on the expressive autonomy factor indicates that entering students with low scores on the expressive autonomy scale (i.e., low in impulsiveness, independence of judgment, and tolerance for ambiguity and relatively high in adherence to traditional Christian beliefs and rituals) were more likely than higher scorers to maintain or increase their interests in tort law during the first law school year.Google Scholar

36 Only 11 of the 14 OPI scales are discussed here and presented in table 5. The scales omitted are Aestheticism (a less important scale for the purposes of this study), Masculinity-Feminity (omitted because it measures adherence to what may be outdated sexual role stereotypes and would engage the study in fruitless controversy), and Response Bias (omitted because scores on this scale correlate substantially with scores on intellectuality and psychological well-being scales).Google Scholar

37 The differences can be illustrated by the following. Table 5 shows interests in constitutional law to be positively correlated, both at entrance and at the end of the year, with scores on the three OPI scales comprising the intellectuality factor. Despite this consistent relationship, table 4 shows no relationship between the intellectuality factor score at entrance and constitutional law interest scale scores at the end of the year after variability in the interest scores predictable from interest score variability at entrance has been removed. This means that differential interest score changes during the year are not predictable from intellectuality scores at entrance. Given this situation, we might expect that intellectuality scores at entrance would predict constitutional law interest scale scores at the end of the year before end-of-year interest score variability predictable from interest score variability at entrance is removed. In fact they do.Google Scholar

38 Examples are the strong and significant correlations in table 5 between interests in constitutional law and scores on the three intellectuality scales of the OPI, which indicate that both at entrance and at the end of the year students with relatively high general intellectual interests were more interested in constitutional law than were students with lower general intellectual interests. The absence of similar significant correlations in table 4 indicates that the students did not change their conceptions of constitutional law over the first year in any marked uniform fashion that was related to personality characteristics, perhaps because their contact with the subject was slight and derivative. If there had been such a shift in students' conceptions of constitutional law, and if, in reaction to this shift, students entering law school with relatively strong intellectual interests had experienced interest changes different from those of students with weaker intellectual interests, then one could expect a significant correlation between the intellectuality factor and constitutional law interest scale scores in table 4.Google Scholar

39 The percentages of BYU students finding the various work settings highly acceptable at the end of the year were as follows:Google Scholar

40 Hedegard I, supra note 1, at 835.Google Scholar

41 I also have indicated that such mean interest changes might not represent common aspects of interest changes but rather could be the accidental result of diverse interest changes, each characterizing one or a subgroup of students.Google Scholar

42 See table 2 at pp. 802–3 supra.Google Scholar

43 If students' interests in each of the 19 content areas and work settings are assumed to be independent of one another (my analyses indicate these interests were not statistically independent of one another), 14 of the 19 mean scale score changes would have to have been in the direction of faculty mean scores to indicate the possibility of a real faculty pull on student interests (at the .05 level, one tailed).Google Scholar

44 Making a similar assumption of interest scale independence, 9 of the 12 statistically significant changes would have to have been toward the mean faculty scores to have been significant at the .05 level, one tailed.Google Scholar

45 I lowered the significance level for two, perhaps debatable, reasons. First, the end-of-year interest score variability shown in relationship to small section variables in table 6 has had variability predictable from interest score and personality variations at entrance removed. Thus the error variance remaining in the interest score variance represents a high proportion of the total variance remaining and might mask some effects of small section variables. Second, effects on interests of exposure to one instructor (out of from seven to ten different instructors) might be expected to be relatively small. This being a pilot study, it seemed useful to lower the significance level to get a better sense of the patterning of the data, given relatively weak expected effects.Google Scholar

46 See pp. 810, 814 supra.Google Scholar

47 See Hedegard II, supra note 2, at 497, for an earlier discussion of this phenomenon.Google Scholar

48 The correlations among the ten small section variables are as follows:Google Scholar

Why these intercorrelations? Imagine students, for example, who were assigned to the small section of contract law 1. All those students would also have been assigned to the large sections of the other first-semester courses—hence the negative correlations of the contract law 1 small section variable with the small section variables for the other first-semester courses. In the second semester, none of these students would be assigned to the small section of contract law 2—hence another negative correlation. Those students would instead be spread among the small sections of the other four second-semester courses—hence the slight positive correlations with the small section variables for these courses.Google Scholar

Now suppose the small section variable for contract law 1 was positively correlated with interests in, say, labor law (there was in fact such a correlation). Because of the correlations among the small section variables, correlations between interests in labor law and several other small section variables (those for Cp 1, Cr 1, Pr 1, To 1, and Co 2) would tend to be negative. In fact, none of these were significantly negative: one (the correlation with Cp 1) was significantly positive. Other correlations (with the small section variables for Cp 2, Cr 2, Pr 2, and To 2) would also tend to be positive: one (the correlation with Cp 2) was in fact positive (and significant at the 10 percent level).Google Scholar

49 This conclusion can be checked further by comparing the patterns of correlations in table 6 against the pattern of correlations among the small section variables.Google Scholar

50 On all 16 of the instruction index scales, none of the correlations with small section variables of courses in which both sections were taught by the same instructor was larger than the largest correlation with a small section variable in a course in which the two sections were taught by two different instructors. On 15 of the 16 instruction index scales, one or more correlations with a small section variable in one of the latter courses were larger than any of the correlations for the same section variables in courses in which both sections were taught by the same instructor.Google Scholar

51 For descriptions of these scales see Hedegard II, supra note 2, at 473–75.Google Scholar

52 One qualification should be added: the student's comments must not be too far off the path the instructor wishes to follow. The speculation index, which measures an instructor's tolerance of quite off-beat and speculative comments, correlates negligibly with interests (in fact, correlations with interests tend to be negative).Google Scholar

53 Grades in law courses at BYU (as they probably are generally) were determined almost entirely by grades on the final examinations in the courses.Google Scholar

54 Industrial psychologists have looked in some detail at relations among skills, motivation, and work performance. See, e.g., Victor H. Vroom, Work and Motivation 141 (New York: John Wiley & Sons, 1964).Google Scholar

55 The argument here would be that personality characteristics are psychologically more central patterns of interest and motive than are content area or work setting interests. Thus, if interest differences cause grade differences, it is because interest differences are linked to these more central interest and motive differences that affect how one learns, what aspects of course material one focuses on, and the like. Thus personality differences should be linked to interest differences on the interest scales on which score differences are linked to grade differences.Google Scholar

56 See tables 4 and 5 at pp. 810, 811–13 supra.Google Scholar

57 It is, of course, possible that differences in interest level on scales such as the antitrust and criminal law scales are linked to interests and motives more central to learning which are not tapped by the particular personality measures included in the OPI.Google Scholar

58 I say “maintain or increase their interests” in this context because average scores on the various interest scales tended to drop over the year.Google Scholar

59 See table 2 at pp. 802–3 supra.Google Scholar

60 This need not have been the case. The mean scores of the three GPA-defined groups on the patent law scale actually converged during the year, and the correlation between GPA and adjusted end-of-year patent law scale scores, while not statistically significant, approached significance.Google Scholar

61 This hypothesis assumes that the information the students received in the first year about content areas and work settings on these five scales was consistent with, and perhaps elaboration and detailing of, the concepts with which they entered law school. The hypothesis is also consistent with the interest-focusing hypothesis I have discussed earlier.Google Scholar

62 It may be that as students gained information about these three interests during the first year, differences in levels of interest were linked to differences in learning and exam-taking behaviors reflected in differences in examination performance and hence in grades.Google Scholar

63 Zemans and Rosenblum, in their study of career paths within the profession, have indicated that public criminal prosecution provides entry-level career positions for many lawyers who eventually move into other areas of the law. Frances Kahn Zemans & Victor G. Rosenblum, The Making of a Public Profession (Chicago: American Bar Foundation, 1981).Google Scholar

64 As shown in table 5, interests in district attorney positions were related to traditional religious beliefs in that the greater the interest in district attorney careers, the greater the adherence to traditional religious beliefs. The instructor of three of the four sections of the criminal law and procedure courses had been a public prosecutor and was an active church member. In the classroom it appeared to me that his religious beliefs were closely linked to his perspectives on criminal law and prosecutor values and strategies. Students also may have perceived these links, and some may have come to view public criminal prosecution as being, in some sense, an activity having a religious component.Google Scholar

65 Since many of the BYU law students had been BYU undergraduates, they may have been more accurate in their preconceptions of law school academic performance than would students at other law schools, partly because of familiarity with their competition. It would be interesting to see whether beginning and end-of-year correlations of interests with GPA are parallel at schools such as Harvard, Yale, Chicago, and Stanford.Google Scholar

66 Of course, one can question how much performance on examinations in law school courses does measure the acquisition of practice-relevant knowledge and skills, but this article is not the setting in which to argue that issue. I will, however, in the conclusions section indicate a possible discrepancy between the qualities rewarded on examinations and the qualities rewarded within the classroom.Google Scholar

67 Although the attractions and repulsions of public agency careers might be linked to their bureaucratic nature, suggesting less autonomy and greater constraints, our data do not support this possibility: our measures of the importance students attach to independence and personal expressiveness do not correlate with interests in such work settings, even at entrance before any possible interest-tempering effects of first-year grades.Google Scholar

68 The correlations between personality measures and end-of-year interest scale scores (see table 11) indicate changes in interest during the year, which, as I have indicated, suggest that changes in interest result from changes in conceptions of a content area or work setting. They are unlikely to reflect attempts to talk up or talk down areas or kinds of practice settings: these latter attempts, if successful, are likely to be reflected in correlations between interest changes and personality.Google Scholar

69 The correlation between the reliability coefficients in table 11 and the total multiple correlation coefficients with all other predictor variables was -0.36. The correlation between the R values (in table 12) for student characteristics at entrance and the R values for experiences within the first law school year was -0.40.Google Scholar

70 As the tables show, the teacher scale has both the highest reliability coefficient and the highest R with other predictor variables, and the district attorney scale has the lowest values of both.Google Scholar

71 Data on changes in correlations among interest scales over the year can shed light on changes of meaning in several ways. If scores on an interest scale had few or no correlations with other scales at entrance but developed significant correlations over the year, this could indicate that the content areas or work settings comprising the scale had little meaning to students at entrance. If correlations between an interest scale and other scales were substantial at both times but changed during the year, this could indicate that perceptions, though “clear” to students at entrance, were inaccurate and became more accurate over the year.Google Scholar

72 Interest score correlations between the two items were about + 0.60 at both times.Google Scholar

73 See table 6 at p. 820 supra.Google Scholar

74 See table 2 at pp. 802–3 supra.Google Scholar

75 Some students found tort law compatible with more central interests and motives but unattractive in other respects (not identified in this study), and other students had the opposite set of contradictory reactions.Google Scholar

76 See table 13 at p. 841 supra.Google Scholar

77 These links contribute to evidence that some students were interested in antitrust law for its antibusiness aspect.Google Scholar

78 Anyone objecting to considering military law a rather “abstract, policy-oriented” area must keep in mind that this scale contained the maritime law item (which, in fact, had a strong, positive correlation with the labor law item) and also that some students indicated that military law implied more than courts martial.Google Scholar

79 See table 2 at pp. 802–3 supra.Google Scholar

80 See comments on criminal law with table 18 at p. 846 supra.Google Scholar

81 Some BYU students indicated great moral reservations about defending criminals whose guilt was clear. Such reservations, brought into the open in the criminal law class discussions, may have added to mean score drops.Google Scholar

82 Most students, including most of the less academically successful students, ended the year with stronger interests in small private practice settings than in government careers, indicating that the bulk of students with relatively strong interests in family, estates, and tort law areas would end up in private rather than public work settings.Google Scholar

83 If all scores on a variable are identical, that variable can have no correlation with other variables. Since virtually all students indicated, both at entrance and at the end of the second year, strong or moderate interests in small partnership and small law firm careers, correlations with other variables were constrained.Google Scholar

84 See table 6 at p. 820 supra.Google Scholar

85 I did entertain the possibility that the correlations between small section variables and interest scale scores occurred only by chance. Because of the nonindependence of the small section variables and among the interest scales, I could not determine how many significant correlations one might expect by chance. However, I could compare this correlation matrix with another matrix of correlations in which one would expect only chance correlations: the correlations between personality scale scores at entrance and small section variables (there should have been only chance correlations in this latter matrix since student assignments to small sections were random). Indeed, the general magnitude of correlations and the number of significant correlations were far smaller in the latter matrix.Google Scholar

86 Because we were focusing on the processes rather than on the content of instruction we did not record the specific concepts and issues being presented or discussed in class, instructors' expressions of interests and attitudes, and so forth.Google Scholar

87 Issue resolution versus ambiguity in students' performance on examinations has been taken up earlier in this study: see Hedegard II, supra note 2, at 500. Students who thought instructors stressed acknowledging alternative reasonable concept applications and issue resolutions tended to perform better on exams than did students who perceived instructors to be more “dogmatic.” Instructors claimed that they sought exam answers that explored alternative resolutions, but then they offered a single resolution.Google Scholar

88 Sherry Diamond (personal communication) has suggested a contrary hypothesis. Since the BYU students were somewhat conformist, religiously traditional, and raised in a strong, pervasive religious culture, their basic interests and attitudes might have been well formed prior to beginning law school and, therefore, not likely to be affected by outside pressures, even from such credible sources as BYU law faculty. These possibilities await adjudication by replications of this kind of study at other schools.Google Scholar

89 For some students, international law may have remained a relatively exotic and exciting alternative to better known content areas.Google Scholar