The importance of lecture in general chemistry...
0 downloads
106 Views
2MB Size
Svmposiurn: lecture and learning: Rre Thev Compatible?
The Importance of Lecture in General Chemistry Course Performance James P. Birk and John Foster Arizona State University, Tempe, AZ 85287-1 604
Arguments could be advanced that, to be effective,a lecture should have several characteristics, including good organization, appropriate structure to cover prerequisites before a topic is introduced, frequent use of demonstrations, use of appropriate visual aids, efforts to show relevance to the real world in an attempt to excite students and encourage them to study chemistry, and a degree of interaction between the instructor and the students. Perhaps it is no accident that many modem textbooks have many of these characteristics. One of the authors had occasion to question the efficacyof lecture as an aid to learning while teaching one of three sections of general chemistry. The three lecturers had quite different styles in terms of the features mentioned above. However, when a common final exam was administered to the three sections, the averages were 62%,61%, and 60%.If any real learning was occurring as a direct result of the lectures, it would have seemed reasonable for us to experience a greater discrirnination between sections. We decided to test the hypothesis that learning occurs as a result of lecture. We wish to report on three studies that bear on this hypothesis.
Experiment 1 The final exam results cited earlier do not support this prediction. In addition, we carried out a more extensive study. Students from the first-semester course (CHM-113) were tracked into the second-semester course (CHM-115 or CHM-116) over a period of years. If the hypothesis is correct, the second semester performance should vary with identity of the first-semester instructor. The instructors in this studv ranged from newlv facultv " aaduated tem~orarv to seasoned veterans, and from individuals whose primary interest was teaching to research-oriented faculty To obtain significant numbers, results from each instructor
-
First Semester: CHM-113 Second Semester: CHM-115 or CHM-116 CHM-115/116 Professor
Professor
If learning occurs during lecture, the extent of learning should vary from one instructor to another. Table 1. Summaries of Standard Final Exam Scores for CHM-115 lnstructor X
-
Table 2. Performance of Students in the SecondSemester General Chemistry Course
CHM-113
Study 1 Prediction I
-
Wa 116
X 115
X 116
Y
zb
115
115
-
ND
'd
ND
B
ND -
ND
'
ND
gf
C
ND
ND
ND
-
D
-
ND
*
ND
-
E
ND
ND
*
ND
IT eT gt
F
ND
ND
'
ND
fl
-
First-Semester lnstructor
Mean
Std Dev
Cases
For Entire Populationa
-0.0254
0.9820
144
CHM-113 Instructor B
4.0700
0.9216
60
CHM-113 Instructor E
0.0385
0.9465
26
CHM-I13 Instructor F
0.01 08
1.0099
26
CHM-I13 Instructor G
0.0044
1.1331
32
-
'Th s s a s ~ n f i c a n ts a s e l of tne em re pop. ason. M the mean H a s not exanly Zen) A Sma n x m e r of s l A e n l nth s m r s e *no 0 0 no1 la6e CrlMI13 a! ASU w in one of me slea nsthctom were not n c l m o in i n s samp e.
180
Journal of Chemical Education
-
A
0.001
0.003 0.005 0.002
el gJ
G
ND
ND
*
ND
ft et gt
0.018 0.007
o m
'ND = no significant difference. 'f = final exam, e = exam average, g = murse grade; the arrow indicates
whether the scores were higher or lower than average. 'P = probability that the two samples could have been selected randomly. he symbol ' means that differences were found only for specific pairs of instructors and these differences did not fall into any coherent pattern.
were collected over several years, when data were availqble. In order to make such comparisons valid, scores were converted to standardz-scores, which have a mean of 0 and a standard deviation of 1( I ) . Results 1 There was some variation in second-semester ~ e r f o r mane, but most differences were not statistically significant. Statistical studtes were curried out with SPSS 2, on an IBM PC clone. Typical mean final exam scores are shown in Table 1for one course. Variations in mean final exam performance between groups associated with different CHM-113 instructors are quite small. T-tests were carried out for each pair of means for all the second-semester courses. For the data listed in Table 1, t-tests gave twotailed probabilities in the range of 0.60.9,indicating that no two groups were significantly different. Results for seven CHM-113 instructors whose students went on to five sections of CHM-115 and CHM-116 with four different instructors are shown in Table 2. In two cases, a s noted, t-tests indicated significant differences, though the differences for final exams, hour exams, and overall course performance were not all significant in most cases. Data were examined in more detail to see if there were any other differences among the groups. It was noted that the performance in CHM-1151116 varied by instructor only in those cases where significant differences were Table 3. Analysis of Variance in CHM-115 (Instructor Z in Table 2) Final Exam by CHM-113 Instructor and CHM-113 Final Exam
Source of Sum of Degrees of Mean Variation Squares Freedom Square Main Effects
Significance of F
34.348
11
3.123
7.144
0.000
4.265
3
1.422
3.252
0.029
24.189
8
3.024
6.918
0.000
CHM-113
Instructor CHM-113
Final Exam 14.637 17 0.861 1.970 0.032 2-way Interactions 28 1.749 4.003 0.000 Explained 48.985 Residual 22.727 52 0.437 Total 71.71 2 80 0.896 81 Cases were processed. 'F is the ratio of the mean square variance anributed to that source relative to the unexplained or residual mean square variance. The greater this number, the Oreaterthe amount of variance that can be attributed to thissource. Table 4. Mean Exam Performance in CHM-113 by Number of Absences
Absences no data 0 1 2 3 4
5 7-€ &20
For Entire Population
Mean
Std Dev
Cases
found in CHM-113 performance. An analysis of variance (ANOVA)study of these cases indicated that the only significant contribution to the variance came from previous performance and not from previous instructor, as shown in Table 3. For some reason, which we have not identified, some CHM-113 instructors were sending a different distribution of students on to the second-semester course than others. Only in these cases did we observe statistically different performanee in the second semester. Study 2
Prediction 2 If lecture has an effect on learning, students who attend lecture should perform better on exams than students who are absent. Table 5. Course Performances for Students Who Attend or Miss Lecture.
CHM-113 Fall 1991 Attended Final Exam (Std) Exam Ave. (Std)
Total Hoursat ASU ASU GPA HS GPA HS Rank
Transfer GPA ACT total ACT math ACT science ACT reading SAT total SAT math SAT verbal Reasoning Skills Test
0.33 0.28 45.5 2.93 3.30 21.9 3.00 23.4 23.9 23.6 23.7 1022 557 465 2.09
Absent -0.32 -0.28 42.6 2.34 3.04 29.4 2.87 23.6 23.5 23.5 23.7 1015 548 469 2.06
Significant by t-test? Yes Yes NO
Yes Yes Yes NO NO NO
NO NO
No NO
No NO
CHM-I01 Fall 1991
Attended Final Exam (Std)
Exam Ave. (Std) Total Hours at ASU ASU GPA HS GPA HS Rank Transfer GPA
ACT total ACT math ACT science ACT reading
SAT total SAT math SAT verbal Reasoning Skills Test
0.15 0.18 48.6 2.71 3.07 29.5 2.90 21.3 21.2 20.7 20.9 927 496 430 1.65
Absent -0.23 -0.27 51.1 2.29 2.88 39.5 2.69 22.3 21.9 21.9 23.7 947 512 434 1.62
Significant by t-test? Yes Yes
No Yes Yes Yes Yes NO
No NO
Yes
No NO
No NO
Volume 70 Number 2 February 1993
181
Experiment 2A
Students in one section of CHM-113 were asked on the second exam to list the number of lectures they had missed; 70% responded to the questionnaire. Average performance on the first two exams was calculated for groups falling into different numbers of missed lectures (0-20). If learning occurs during lecture, performance should decrease as absences increase. Results 2A
Mean performances on the first two exams are shown in Table 4. T-tests indicated that there were no significant variations between any pair of groups, while ANOVAindicated that absence was not a significant factor (P= 0.47) in influencing the variance in exam performance. Experiment 28
Student attendance in four CHM-113 (General Chemistry, 750 students) sections and in three CHM-101 (Introductory Chemistry, 580 students) sections were sampled on a random Wednesday that was not near any special day (such as an exam or a holiday). Attendance in these sections varied between 30% and 70%. Mean performance would be exoected to be hieher for the attenders if lecture and learning are connected. Course scores were converted to standard z-scores (1) to allow com~arisonsof multiole sections.
-
Results 28
Table 5 compares course performance for the two groups in the two courses. As can be seen, course performance is enhanced in the group that attends lecture, by about 0.40.6 standard deviations, which corresponds to only about 5 points (per 100)improvement in scores on a normal curve. This conclusion is supported by t-test and ANOVAstudies. The higher scores by students who attended lecture, though the differences are small, are statistically significant and may, in fact, be a result of attendance a t lecture. However, because these results seemed at variance with the other two experiments, we decided to see if any other factors were different for the two groups. We were able to do such a study because we had accumulated other data for
182
Journal of Chemical Education
a study of reasoning skills (3).The boldfaced lines in Table 5 indicate the other factors that were significantly different for the attenders and absentees. accordine to t-tests. The two groups of students appear td have theiame capability to perform well, as measured by SAT and ACT scores and by scores on reasoning skills tests we have been investigating (3). However, the attenders have a higher GPAin high school and at Arizona State University (ASU). Students in Introductorv Chemistw also have a hieher GPAin work transferred to-ASU, from community colleges. Interestingly, students in Introductory Chemistry who do not attend have a higher ACT reading score. Perhaps these students feel more comfortable using the textbook to study the material. In general, students who attend chemistry lecture are better students in all their coursework at ASU, but reasons for this are not revealed by this study in its present state. Factors that we intend to investigate in a further study include time spent studying chemistry, time spent working for a living, home location and environment (includine on- or off-camous housing.). and attitude towards e h e r n h y In any evek, this study orovides little evidence to sueeedt that substantial learning occurs as a result of att&;ance a t lecture. While we have not vet investieated the eficacv of other wavs of using leetire time, we believe these studies provide"a license for change. It is unlikely that other uses of classroom time would be less effective. Some possible alternatives to lecturing have been discussed by Brooks (41, Strauss and Levine (51, and Fasching and Erickson (6). We h o ~ the e studies reported here mieht encouraee others to siklarly question the effectivenessof the manner in which they are teachine their general chemistrv courses and to accumulate dataon the ;elative effectiveness of alternate approaches.
-..
Literature Cited 1. Glas8.G. ';Hopkins, KD. StotistlmlMDUlodri n E d u e o t i o n ~ d P ~ h o l o g y . 2 n d e d . ; Rentke-Hall: Engl-md Cliffs, NJ, 1984: pp 5%73. 2. SPSSIPC* 4O;SPSS he.: Chicago, 1990. 3. Birk, J. R; Foster, J.; Ku*, M. J.; Woodward, 5 . Te~easoningSknls and Mi-ncep tiom in General Chemistry- ARationsle for Change",presented atthe 12th Biennial Canference m Chemical Education. Davis, CA, August 1992. 4. BmLs,D. W.J. Ckm. Edue., 1984 61, 8-9. 5. Strausr,M.J.;Le%ne,S. H. J. Chrm Educ, 188S.62.317. 8. Fasching, J.L.;E"ckson,B. L. J. Chern. Edue., lS85,8,ffl2-6.
