Loftman, Guy R., Study Habits and Their Effectiveness in Legal Education, 27
J. Legal Educ.
418 (1975) and Stevens, Robert, Law Schools and Law Students, 59
Va. L. Rev.
551 (1973) are the two major reported studies that offer some information on the question. Loftman conducted a survey at Indiana University School of Law in 1973 on the relationships between study habits and grades among first- and second-year law students. He concluded, among other things, that study habits, including the amount of time spent, had only small relationship to law school grades. Id. at 445. He gathered his data, however, by having 329 second- and third-year law students (roughly a 60 percent response) recall their study pattern in two courses each had taken in the previous year. He acknowledged that this method might reflect the students' perceptions of their own study habits more than what they had actually done. Id. at 422. Loftman, at 419–21, also reviews the scanty literature on the effect of study time and study techniques on performance in undergraduate settings. A major study on law student activity patterns, conducted by the American Bar Foundation under the direction of Ronald Pipkin, is in the analysis phase. Two partial reports appear in Legal Education: The Consumers' Perspectives, 1976
1161, and Law School Instruction in Professional Responsibility: A Curricular Paradox, 1979
A.B.F. Res. J.
247, but neither addresses issues dealt with in this study. Some additional information from Pipkin's study appears in Law Schools and Professional Education: Report and Recommendations of the Special Committee for a Study of Legal Education of the American Bar Association 38–41 (Chicago: American Bar Association, 1980).
Kimball, Edward L. & Farmer, Larry C., Comparative Results of Teaching Evidence Three Ways, 30
J. Legal Educ.
196, 202–6 (1979).
Most of the students who did not take part came from the lower part of the class when measured by both admission criteria and performance. When classified by first-year performance, 1 student from the top third did not participate, 5 from the middle third, and 12 from the bottom third. This left 46 participants from the top third, 43 from the second, and 35 from the bottom third.
See 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.
793, 808 n. 46. The general characteristics of the BYU student body are described by Hedegard at 812–19. He gives reasons for thinking that in most respects they are typical of law students everywhere. Their time-use patterns could be affected by the fact that most were married and their average age was a year or so higher than at other schools, but there is no reason to think that the relationships among ability, effort, and performance would be different. Stevens reports that 38 percent of married students at Yale reported feeling time pressure, compared with 11 percent of single students, but that the majority of all students felt that marriage aided adjustment to law school by providing emotional stability. Stevens, supra note 1, at 646. The BYU participants' average LSAT was 652 and their average undergraduate GPA was 3.58.
At BYU, law school grades range from 50 to 90 and correspond with letter grades as follows:
Each law school course was taught in a large section of about 115 and a small section of about 30, with different teachers for each section. In some courses, both fall and winter semesters, the average grade given in one section of the course differed significantly from the average grade given in the other section. The differences in averages could not be attributed to chance or to a difference in the average ability level of the students in the two sections. As a result, an adjusted grade was given to all students in all courses, which took into account the average grade difference between given large and small sections as well as the ability levels of the students in these sections. The adjustment was made by regressing first-year grades on ability (ability as used here is LSAT+ 200 GPA, the law school entrance criterion), while using the class section as a covariate. The remainder become the adjusted grade. Performance was measured by this adjusted grade.
Students seemed to have no great difficulty filling out a form on Monday requiring an hour-by-hour description of study time for a period as much as three days earlier.
As the year progressed, lateness in returning forms increased, especially during moot-court exercises and the final examination period. Students' carelessness in filling out the forms also caused some problems. So that all the forms could be read by machine, considerable time and effort had to be spent darkening boxes that students had left either too light or only partially filled in. On occasion, a student would indicate he attended a class on a Saturday or Sunday when, of course, no classes were held. Probably the most frequent error occurred when a student accidentally filled in the wrong time-line boxes by filling in the boxes immediately above or below the correct line. We wrote computer programs to first identify and then correct detectable student errors of this nature. For example, if a student indicated he attended a class but did not fill in the corresponding time-line boxes for that hour we credited him with having filled out those boxes. At the end of the project we discovered that perhaps ten students each semester were missing 1 or 2 of their 42 forms or had failed to include their identification number. To avoid problems in analysis with missing data we constructed replacement forms based on the average for the same weekday on the students' other forms Overall, the care and diligence of students impressed us. In spite of the problems listed, we felt this data-gathering method was much more accurate than the methods used in previous studies.
Because we gathered data for only 2 days each week for the first 14 weeks of each semester, reported effort for those weeks had to be muliplied by 3.5 to obtain an estimate of actual effort.
These data, and the data in the following tables, are based on 133 student participants in the fall and 124 in the winter semesters. The 9 students who chose not to participate during the winter semester averaged 889 hours of effort during fall semester, compared with the class average of 904 hours. If these students had participated during the winter semester and had again averaged 15 hours less than the rest of the class, the overall class average during the winter semester would have been about 1 hour less effort than it was without them.
Students thus studied approximately 3.4 times as many hours outside the classroom as time spent attending class. This does not mean they spent 3.4 hours in preparation for each class, since some of the study time was to catch up for classes missed and other time was devoted to Legal Writing and moot-court preparation. Maximum classroom time would have been about 415 hours. Hedegard, James M., The Course Perceptions Questionnaire: Development and Some Pilot Research Findings, 1981
A.B.F. Res. J.
463, 476 table 2, shows that members of the 1974–75 first-year class at BYU estimated their mean out-of-class study at 10.39 hours per course per week. For five courses, excluding Legal Writing, that would be 51.95 hours per week. If 13 hours of class (90 percent attendance at 15 class meetings) is added, the total estimated effort by those students would be 65 hours plus the time spent on Legal Writing. The students' estimates of effort obtained this way are substantially higher than what we found based on daily records.
Loftman, supra note 1, at 423–24. Loftman asked students to recall the total time they spent in each of several activities related to a six-credit first-year course in Civil Procedure and a four-credit first-year course in Torts. He then calculated that the first-year students spent on the average 183.9 hours per three-credit course in class attendance and study. Adding 3 hours for examinations and dividing by 16 weeks of classes and examinations suggests 11.7 hours per week time expenditure on each three-credit course. For 15 credits that would be 58.5 hours per week; for 16 credits, as at BYU, it would be 62 hours per week.
In the study reported by Stevens, supra note 4, at 650, the median for hours of study per week during law students' first semester was 30 (to which one should add up to 15 hours of class attendance to compare with our data). In a sample of 527 students, 25 percent studied more than 40 hours, 24 percent studied 30 to 40 hours, and 51 percent studied less than 30 hours. (These figures are derived from Stevens's table 46 at 650.) Preliminary information from Pipkin's study, reported in Law Schools and Professional Education, supra note 1, at 38–41, shows that in his sample, first-year students studied 36 hours per week in the middle of their first semester and 33 hours per week near the end of their second semester (to which one should add up to 15 hours of class attendance for comparison). The results of both these studies are lower than hours given in table 1.
Loftman, supra note 1, at 424.
Stevens's study, supra note 4, at 652–57, and Pipkin's study, supra note 1, at 40–41, document the decline in students' study time. In a study of seven schools, Stevens found that in three the major decline was between the first and second semester, with more gradual decline thereafter; in four schools the major decline came between the second and third semesters. In the first semester almost no one studied less than ten hours per week, but by the fifth semester about 25 percent studied less than ten hours per week. Some students maintained essentially level study patterns, many declined, and almost no one increased the amount of study. Only low grades in the first semester seemed to have retarded decline in study. Stevens in table F.7, at 657, reported this relationship between class rank after the first semester with the percentage of students reporting that they studied as much in the second as in the first semester:
Pipkin reported average total involvement in law school per term as follows:
Law Schools and Professional Education, supra note 1, at 40. This shows decline in involvement but not the great disengagement suggested by Stevens's data.
Distortions in effort/performance relationships would occur if effort were highly correlated with ability and if ability rather than effort were the major determinant of law school performance. Greater effort by the more capable students might then cause one to believe incorrectly that effort was an important factor in performance. Thus, to the extent effort was found to be highly correlated with ability, its use as a predictor of law school performance would be confounded by the corresponding relationship between ability and performance. However, upon examination we found little correlation between effort and our measures of student ability, as shown below:
The findings are consistent with Loftman's results, supra note 1, at 433. Loftman found a .10 correlation (p <.05) between study time and grades. Loftman's data, however, suggest a weaker relationship between effort and performance than was found in this study. The BYU results may more accurately estimate the true relationship between first-year effort and performance than the Loftman findings, since he did not use current time records but relied on students' estimates made at the end of the first year. Other studies outside the law school environment show similar results. Allen, Lerner, and Hinrichsen, in a study of 122 undergraduate students in a psychology course at the University of Illinois, found the relationship between number of days studied and grades to be .23, with an associated probability of less than .05. Allen, George J., Lerner, Wayne M., & Hinrichsen, James J., Study Behaviors and Their Relationships to Test Anxiety and Academic Performance, 30
407, 408 (1972). This value is close to the .27 correlation produced by this study. Williamson found the following correlations between final grades and study hours in a one-week experimental period among freshmen students at four major universities: Syracuse = .32, Yale = .00, Minnesota = - .06, and Iowa= - .28. Williamson, E. G., The Relationship of Number of Hours of Study to Scholarship, 26
J. Educ. Psych.
682, 685 (1935). The experimental period was short and the research conditions dissimilar; hence, the results are strikingly varied. At the University of Utah, questionnaires distributed to 190 lower-division science and engineering majors revealed that the estimated number of hours students spent studying per week was significantly correlated with grades. The correlation was .31. Rollie Wagstaff & Hoda Mahmoudi, Relation of Study Behaviors and Employment to Academic Performance, 38
380, 382 (1976). None of the previous studies have taken the procedural precautions incorporated into this study, nor are they directly comparable. They do, however, also suggest a moderate positive relationship between grades and study time. Hedegard, supra note 12, at 502, found “the absence of a significant relationship between amount of studying and performance, even after LSAT-measured abilities have been statistically eliminated.” We did find a significant relationship.
The .27 correlation between effort and performance means that effort accounted for 7.5 percent (that is, .272) of the variation in performance over the year. When out-of-class study was separated out of effort the results were very similar to those in table 4. The correlations between out-of-class study and performance were. 18 for all (p= .02), .30 for winter (p <.01), and .24 for the entire academic year (p < .01). Out-of-class study explained 5.8 percent of the variation in performance over the year. When attendance was separated out of effort, the correlations between attendance and performance were .14 for fall (p= .01), .33 for winter (p < .01), and .35 for the entire academic year (p < .01). Attendance explained 12.5 percent of the variation in performance over the year.
Ability alone accounted for 40.6 percent of the variation in performance, compared with the 7.5 percent that effort alone explained. Together, ability and effort accounted for 46.3 percent of performance variability over the whole year. The individual effects are not quite additive because of a small correlation between effort and ability. Loftman's study also showed that LSAT and undergraduate GPA (which together constitute ability in this study) were much more strongly associated with law school grades than were any study habits, including total study time.
Predicting performance from ability was done by calculating a regression model on the basis of all 142 first-year students. After constructing the model, each student's predicted performance was subtracted from his actual performance to compute his disparity score.
The top effort group averaged 1,735 hours of effort per year, the middle group 1,407 hours, and the bottom group 1,142 hours.
Though effort and ability were nearly uncorrelated when first-year participants were considered as a whole, dividing the class into three ability groups enabled us to see whether there might be greater correlation at some ability level. As used here, ability=LSAT+ 183 GPA. When performance was regressed on LSAT and GPA, this combination gave the highest correlation with performance for the entire experimental group.
Only those students who participated in the project were considered in making the divisions: hence, equal numbers of students comprised each third. In establishing disparity, we established a predicted performance independently for each of the three groups. The top ability group had a law school average performance of 75.7 at the end of the year; the middle third had a 72.8 average; the lowest entering abilty group had a 71.5 average.
A similar analysis of the correlations between effort and disparity from predicted grades produced similar results:
Another analysis of disparity from predicted performance showed that the top third in out-of-class study performed on the average 1.16 grade points higher than predicted from their ability, while the bottom third scored 1.10 grade points lower than expected. These data are shown below:
In addition to recording attendance data from the time-use forms, we asked each student in a year-end questionnaire to indicate the number of classes he or she had missed in each course. See appendix 2 for a description of this effort. We believe that the sample obtained from the time-use forms is the more reliable, and therefore we used this measure of attendance in calculating the effects of attendance on other variables.
Compare these correlations with the .27 correlation between effort and performance reported supra in the text at note 19.
The BYU average law school grade in major courses at the end of the first year was best predicted by the following equation, based on 400 possible classes to attend:. average grade = 27.43 + .02087 (LSAT) + 3.81 (GPA) + .04555 (classes attended) + .00206 (hours of out-of-class study). When attendance and out-of-class study were combined into effort, the equation became average grade = 20.32 + .02257 (LSAT) + 4.02 (GPA) + .00301 (effort). This suggests that if two students had the same LSAT and undergraduate GPA we might expect a difference of one point in law school GPA at the end of the first year (where the difference between A and B is seven points) if one student studied 332 hours more than the other (i.e., 1/.00301).
Reversing table 11, average grade point disparity (after adjusting for out-of-class study) by attendance is shown below:
Note: The standard deviations of these numbers ranged from 0.31 to 0.49. a
Disparity figures in parentheses are for comparison. They are not adjusted for out-of-class study.
We divided the total effort expended by each student during the last four and one-half weeks of classes in the semester and the two weeks of the final examination period by the student's effort for the whole semester. We called this ratio cramming. Since this ratio might also partly reflect a student's total effort, another variable was computed which adjusted this ratio for effort by using a simple linear regression of performance on effort. We called the residual adjusted cramming. The inverse of these variables is used in table 13.
We divided the student's total effort during the semester by the number of study interruptions of at least one-half hour. This became a measure of how long a student would study at one sitting, which we called concentration. Since this ratio might also correlate very highly with effort, another variable was computed which adjusted this ratio for effort by using a simple linear regression of performance on effort. We called the residual adjusted concentration. It is used in table 13.
Only 6 of the 36 students who had been employed had law-related jobs. Lunneborg, Clifford E. &Lunneborg, Patricia W., Relations of Background Characteristics to Success in the First Year of Law School, 18
J. Legal Educ.
425, 435 (1966) reported that for 395 students entering the University of Washington School of Law in 1956–59, first-year grades and outside employment had a correlation of — .29. This should be compared with John E. Hay & Carl A. Lindsay, The Working Student: How Does He Achieve? 10 J. College Student Personnel 109 (1969), which found in undergraduate work no difference in grades unless students worked 16 or more hours per week. Research conducted at the University of Utah also showed that undergraduate students holding part-time jobs spent about the same amount of time studying as those who did not hold jobs. Wagstaff & Mahmoudi, supra note 18.
Participation in study groups declined from fall to winter semesters both in numbers of students involved (55 percent/29 percent) and in the average time spent by those who did participate (2.3 hours per week to 1.7 hours per week).
The square root of this percentage is the correlation, e.g., 40.6 percent equals a .64 correlation between ability and performance.
The four most important factors, combined in one predictive model, accounted for 50.6 percent of the variation in law school performance. This corresponds to a multiple correlation of .71 between performance and the combination of these four factors. In a stepwise regression, concentration as a fifth factor failed to enter the regression model at a .10 level of significance. Lunneborg & Lunneborg, supra note 32, at 435 table 5, found that LSAT explained 20 percent of the variation in first-year grades and undergraduate GPA explained 19 percent; outside employment explained 8 percent, and the student's age 5 percent. The latter two had a negative relationship with first-year grades.
Stevens, supra note 4, at 652–53.
Loftman, supra note 1, at 425 table 2.
Anikeeff, Alexis M., The Relationship Between Class Absences and College Grades, 45
J. Educ. Psych.