Chapter Four

An Investigation of Learning Styles in General Chemistry Students


RESULTS

Overview

The research study was designed to first determine the strength of the model being used, and secondly to begin to determine whether the model, so confirmed, might help to improve the quality of education for the student. As such, we begin our analyses by examining the goodness of fit of the model to the students it attempts to categorize. Subsequently we examine the subgroups by category for differences in achievement in general chemistry, and finally whether the differences between treatment (i.e. instruction in study techniques specific to cognitive quadrants) and control (i.e. no such instruction) has any effect on grade in the course, overall or for any identifiable subgroup.

Demographics

Subjects for the study were all students of Clemson University, during the Fall semester, 1994, for the pilot study with Psychology 101; Spring semester, 1995, for the CH 102 initial segment and Fall semester, 1995, for the CH 101 replication. All students were regularly enrolled in the courses and sections, and as such represent a highly self selected population. Following the pilot of the cognitive profile inventory instrument with psychology students, the initial study was conducted with the general chemistry 102 students, and replicated with the general chemistry 101 students.

First phase, CH 102, Spring 1995

The students in the first phase, the general chemistry 102 group, consisted of 957 students who responded to the initial questionnaire on the first day of class. The questionnaire consisted of the cognitive profile instrument and multiple response items on gender, major, college year and ethnicity. Of those 957 students, 765 finished the course and received a final grade.

Gender

Analysis of the data showed the group consisted of 608 men and 339 women, plus 10 students who omitted the response to the gender question. 485 men finished the course for an attrition rate of 20.2%. 280 women finished the course, for an attrition rate of 17.4 %. Overall attrition rate of all students was 19.2%.

Table 1. Student Responses to Gender Question

Gender
Initial No.

Final No.

Attrition
Male
608

485

20.2 %
Female
339

280

17.4 %
No Answer
10






Major

The students were requested to choose one of five categories into which their major best fit. The categories were a. Sciences, Engineering, Mathematics, other technical; b. Business, Liberal Arts, pre-law; c. Elem. education, nursing, other helping professions; d. Creative Arts; e. Undecided. It was judged important to have the student's own evaluation of category of major in order to more appropriately evaluate their intended career path, since freshmen students might be expected to change specific majors.


Table 2. Students' Identification of Themselves by Major


 

Major


Initial No.

Final No.

Attrition
A
825

674

18.3 %
B
42

29

31.0 %
C
66

49

25.8 %
D
4

4

0.0 %
E
10

9

10.0 %
N/A
10






Year

Students responded to a multiple choice question of college year. The choices were limited to a. freshman, b. Sophomore, c. Junior, d. Senior. Some students had already completed an undergraduate degree and were completing requirements for a second undergraduate degree or pre-requisites for admission to a graduate program. Those students who asked were instructed to mark "d" Senior, since they had the educational experience of a senior.


Table 3. Student Responses to Academic Year

Year
Initial No.

Final No.

Attrition
1
751

610

18.8 %
2
134

107

20.1 %
3
42

31

26.2 %
4
20

17

15.0 %




Ethnicity

Students responded to a multiple choice question worded "ethnic background." The choices were limited to a. Caucasian, b. African, c. Asian, Indian Subcontinent, d. Asian, not Indian, e. Hispanic. These responses were selected based on prevailing politically correct descriptions as much as could be determined, in efforts not to offend any student. The line was added: "Please leave question #4 blank if A-E are not adequate descriptors."

Table 4. Student Responses to Ethnic Background

Ethnicity
Initial No.

Final No.

Attrition
Caucasian
814

655

19.5 %
African
84

69

17.8 %
Asian, Ind.
6

5

16.7 %
Asian, Oth.
10

10

0.0 %
Hispanic
8

7

12.5 %
No Answer
35






Group Assignment

The student population in phase one was divided into treatment and control groups by the last digit of the social security number. The students with an even digit ending their social security number were assigned to the control group. The students with an odd digit ending their social security number were assigned to the treatment group. Frequency analyses with Chi Square tests were run on the result for every demographic and for lecture section. All analyses resulted in P values in the .87 or higher range, indicating the division by last digit of social security number to be random. Some foreign students do not have a social security number, and are assigned a similar number by the University as a student ID number. In those cases the student ID number was used. For students who have a Social Security number, that number serves as their student ID number.


Second Phase, CH 101, Fall 1995

The students in the second phase, the general chemistry 101 group, consisted of 1208 students who responded to the initial questionnaire on the first day of class. The questionnaire consisted of the same cognitive profile instrument and multiple response items on gender, major, college year and ethnicity as was used in the first phase. Of those 1208 students, 914 finished the course and received a final grade. This is not to imply that all 914 passed the course, only that they remained registered for the course through the end of the semester and received a grade.

Gender

Computer analysis of the data showed the group consisted of 719 men and 464 women, plus 25 students who omitted the response to the gender question. 562 men finished the course for an attrition rate of 22% for all reasons for male students. 338 women finished the course, for an attrition rate of 27% for all reasons for female students. Overall attrition rate of all students was 24%

Table 5. Student Responses to Gender

Gender
Initial No.

Final No.

Attrition
Male
719

562

22 %
Female
464

338

27 %
No Answer
25

14

44 %




Major

The question of major was presented as a multiple choice of 5 categories. The students were requested to choose one category into which their major best fit. The categories were a. Sciences, Engineering, Mathematics, other technical; b. Business, Liberal Arts, pre-law; c. Elem. education, nursing, other helping professions; d. Creative Arts; e. Undecided. It was judged important to have the student's own evaluation of category of major in order to more appropriately evaluate their intended career path. Since freshmen students might be expected to change specific majors, the categories were considered more appropriate.

Table 6. Student Self-Identification Of Major

MajorInitial No.Final No.Attrition
Sciences, Engineering, Mathematics104881222 %
Business, Liberal Arts, Pre-Law201430 %
Elementary Education, Nursing, other helping professions.966136 %
Creative Arts5420 %
Undecided171135 %
N/A221245 %




Year

Students responded to a multiple choice question of college year. The choices were limited to a. freshman, b. Sophomore, c. Junior, d. Senior. Some students had already completed an undergraduate degree and were completing requirements for a second undergraduate degree or pre-requisites for admission to a graduate program. Those students who asked were instructed to mark "d" Senior, as they had the educational experience of a senior.

Table 7. Student Responses to Academic Year

Year


Initial No.


Final No.

Attrition
1
1051

810

23 %
2
99

70

29 %
3
27

16

41 %
4
10

8

20 %
N/A
21

12

43 %



Ethnicity

Students responded to a multiple choice question worded "ethnic background." The choices were limited to a. Caucasian, b. African, c. Asian, Indian Subcontinent, d. Asian, not Indian, e. Hispanic. These responses were selected based on prevailing politically correct descriptions as much as could be determined, in efforts not to offend any student. The line was added: "Please leave question #4 blank if A-E are not adequate descriptors."


Table 8. Student Responses to Ethnic Background

Ethnicity
Initial No.

Final No.

Attrition
Caucasian
1017

775

24 %
African
106

86

19 %
Asian, Ind.
4

3

25 %
Asian, Oth.
18

13

28 %
Hispanic
9

5

44 %
No Answer
54

29

46 %


Group Assignment

The student population in the phase two also was divided into treatment and control groups by the last digit of the social security number. The students with an even digit ending their social security number were assigned to the control group. The students with an odd digit ending their social security number were assigned to the treatment group. Frequency analyses with Chi Square tests were run on the result for every demographic and for lecture section. All analyses resulted in P values in the .87 or higher range, indicating the division by last digit of social security number to be random. Some foreign students do not have a social security number, and are assigned a similar number by the University as a student ID number. In those cases the student ID number was used. For students who have a Social Security number, that number serves as their student ID number. In each phase of the study, additional students took the course, but were not present on the first day of class and did not complete the questionnaire. These additional students were not included in any analyses.

Cognitive Profile Results

The cognitive profile of each student is reported as a set of four numbers, denoting relative preference in each of the four quadrants. The four quadrants, taken together, make up the individual profile. The numbers represent the sum of the ranked and weighted choices of the 20 sets of words in the cognitive profile instrument administered to the students on the first day of class. The four quadrant sums are then summed, and each individual's quadrants divided by the grand sum to result in a percentage of preference by quadrant. This differs from the pilot study pencil and paper instrument where student profiles were expressed as proportions of 180. It was deemed easier for the students to understand if each quadrant expressed as a percentage. This emphasized the concept that every individual has some preference in each area, and most have a dominance in one or two quadrants. The raw percentages were further divided into interval data for some analyses.

In phase one of the study, with the CH 102 students during Spring semester 1995, the results were tabulated by frequency of occurrence of gender and by students' reported classification of major using the interval data. Intervals are reported as 1, very low preference to 4, very strong preference in each quadrant. The data were further evaluated to record which single quadrant showed the highest preference, where a strong singular dominance existed.

Of the 608 men, 101, or 16.6%, did not show a strong dominance in any one quadrant. Of the 339 women, 55, or 16.2%, did not show a strong dominance in any one quadrant.

Numbers in the following tables represent the percentage of subjects, by demographic descriptors, who scored at each level by profile quadrant. Columns for each quadrant will sum to 100% except for rounding anomalies.

Gender

For these and the following tables:

SF = Sensor Feeler ST = Sensor Thinker NF = Intuitive Feeler NT = Intuitive Thinker

1 = very low preference2 = low preference3 = mid to high preference4 = very high preference

Table 9. Cognitive Profile Dominances by Gender, % Men, Phase 1



SF

ST

NF

NT
1 very low preference
4.1

2.0

1.0

0.8
2 low preference
33.7

21.2

24.3

19.1
3 mid to high preference
43.4

41.4

52.0

54.6
4 very strong preference
18.7

35.4

22.7

25.5




Table 10. Cognitive Profile Dominances by Gender, % Women, Phase 1



SF

ST

NF

NT
1very low preference
1.5

3.8

0.9

1.8
2 low preference
15.3

28.9

28.9

35.1
3 mid to high preference
40.4

36.9

49.9

53.7
4 very strong preference
42.8

30.4

20.4

9.4



Table 11. Cognitive Profile Dominances by Gender, % Men, Phase 2



SF

ST

NF

NT
1
5.1

2.9

0.0

0.7
2
30.5

19.9

23.9

22.0
3
42.8

45.9

49.0

56.6
4
21.6

31.3

27.1

20.7




Table 12. Cognitive Profile Dominances by Gender, % Women, Phase 2



SF


ST

NF

NT
1
1.7

4.1

0.4

1.1
2
14.4

26.9

25.4

44.0
3
36.0

46.8

49.8

45.0
4
47.8

22.2

24.4

9.9

In phase one, 42.8% and in phase two, 47.8% of all women scored 4, very high, in the SF quadrant. In Chi square analysis of frequency data for cognitive dominances by gender in phase two subjects resulted in a chi square value of 70.412, with P <0.000. This means there is virtually zero probability of these results occurring by chance.

Major

Table 13. Cognitive Profile Dominances by Major, % Sciences, Phase 1 N=825



SF

ST

NF

NT
1
3.4

2.5

0.8

1.1
2
29.9

22.4

25.3

24.2
3
41.7

40.2

52.0

53.2
4
25.0

34.8

21.8

21.5



Table 14. Cognitive Profile Dominances by Major, % Business, Phase 1 N=42



SF

ST

NF

NT
1
2.4

4.8

2.4

0.0
2
11.9

26.2

28.6

26.2
3
59.5

40.5

47.6

54.8
4
26.2

28.6

21.4

19.0





Table 15. Cognitive Profile Dominances by Major, %: Helping Professions. Phase 1 N=66



SF

ST

NF

NT
1
1.5

3.0

1.5

3.0
2
1.5

37.9

36.4

31.8
3
37.9

31.8

45.5

63.6
4
59.1

27.3

16.7

1.5


Table 16. Cognitive Profile Dominances by Major, %: Creative, Phase 1 N=4



SF

ST

NF

NT
1
0.0

0.0

0.0

0.0
2
0.0

75.0

0.0

25.0
3
100.0

25.0

25.0

75.0
4
0.0

0.0

75.0

0.0



Table 17. Cognitive Profile Dominances by Major, %: Undecided, Phase 1 N=10



SF

ST

NF

NT
1
0.0

0.0

0.0

0.0
2
40.0

30.0

10.0

20.
3
30.0

60.0

50.0

70.0
4
30.0

10.0

40.0

10.0



Phase 2

Table 18. Cognitive Profile Dominances by Major, %: Sciences, Phase 2 N=1048



SF

ST

NF

NT
1
4.2

3.1

0.2

0.9
2
25.7

22.0

24.1

28.8
3
40.6

46.7

49.3

52.6
4
29.6

28.1

26.3

17.7


Table 19.Cognitive Profile Dominances by Major, %: Business, Phase 2 N=20



SF

ST

NF

NT
1
5.0

10.0

0.0

0.0
2
25.0

20.0

30.0

30.0
3
25.0

55.0

45.0

50.0
4
45.0

15.0

25.0

20.0


Table 20. Cognitive Profile Dominances by Major, %: Helping Professions, Phase 2 N=96



SF

ST

NF

NT
1
0.0

2.1

0.0

1.0
2
7.3

28.1

29.2

46.9
3
38.5

41.7

56.2

45.8
4
54.2

28.1

14.6

6.3


Table 21. Cognitive Profile Dominances by Major, %: Creative, Phase 2 N=5



SF

ST

NF

NT
1
0.0

20.0

0.0

0.0
2
0.0

20.0

0.0

60.0
3
60.0

60.0

20.0

40.0
4
40.0

0.0

80.0

0.0



Table 21. Cognitive Profile Dominances by Major, %: Undecided, Phase 2 N=17



SF

ST

NF

NT
1
0.0

0.0

0.0

0.0
2
40.0

30.0

10.0

20.0
3
30.0

60.0

50.0

70.0
4
30.0

10.0

40.0

10.0


Responses are omitted for those 10 students in phase 1 and those 25 students in phase 2 who did not respond to the demographic questions.

Chi square analysis of cognitive dominances by major for phase two subjects resulted in a _ 2 value of 49.972 and P>0.000. This demonstrates correlation between cognitive profile and choice of major. Chi square analyses of cognitive dominances by ethnicity and year resulted in no significant differences. This demonstrates there is no significant correlation between ethnicity or college year and cognitive profile.

The Research Questions

Univariate analyses were performed on all groups and demographic variables and preceded each analysis relevant to the research questions. All populations and samples were normally distributed and no assumptions were violated. Where certain samples or subgroups were too small to be statistically valid such is noted in the tables or text as appropriate.

RQ1Does knowledge of learning profile and appropriate study techniques improve achievement in General Chemistry?

The results in phase one showed no significant difference in the overall mean final grades between the treatment and control groups in CH102, but there was a significant difference in the distribution of those scores. The control group showed a significantly flatter distribution of grades (kurtosis = 1.01) than the treatment group (kurtosis = 3.79) in phase one, CH102.

Table 22. Overall Mean Grades: Control vs. Treatment, Phase 1



Control Group

Treatment Group
Overall Mean
74.73

74.48
Std Deviation
12.68

12.11
Kurtosis
1.01

3.79
Mean SAT V
49.68

50.08
Mean SAT M
59.68

59.37
F Test Results:
F Value = 57.01

Pr> 0.0001



Table 23. Overall Mean Grades: Control vs. Treatment, Phase 2



Control Group

Treatment Group
Overall Mean
72.77

72.82
Std Deviation
13.80

13.56
Kurtosis
0.57

0.23
Mean SAT V
48.90

48.78
Mean SAT M
58.48

60.06
F Test Results
F = 19.10

Pr > 0.0001


RQ2. Which, if any, of SAT scores, math and/or verbal, normalized high school rank in class, or predicted GPA, may predict achievement in general chemistry?

    Using student data as retrieved from Institutional Research in phase one, SAT math score was shown to be the best predictor of all factors tested of final grade in CH102 by regression analysis, with P<0.0001. This confirms, for this population, the prediction factor of SAT math. No other variables were significant predictors of final grade.

    RQ3(a) When a predictor variable such as identified in question 2 is used as a covariate, eliminating performance as a factor, does knowledge of learning profile and appropriate study techniques improve achievement in general chemistry?

    Using SAT math score as a covariate to eliminate this difference still shows no significant difference between the control and treatment group over mean final grades in CH 102, however, there was a significant difference in SAT math score distribution between the two groups, with the control group having the higher SAT math mean score. This difference itself is unexplained, but emphasizes the importance of using SATM as a covariate in these analyses.

    Using the same analysis, there was also no significant difference between the treatment and control groups overall grades in the second phase, CH101.

    RQ3 (b) Does knowledge of learning profile make more of a difference for in achievement in general chemistry for some types of learners than for others?

    Regression of grade versus learning profile by group (without SAT math entered as a variable) showed that the strength in the ST quadrant is the strongest predictor of grade for the control group, having an F value of 7.27, P > 0.0073 for the one variable model, F = 6.64, P > 0.0103 for the two variable model and F = 5.99, P > 0.0148. However, for the treatment group, strength in the NT quadrant became the strongest predictor, with F = 4.56, P > 0.0333 in the one variable model, and a nearly significant ( =0.05 for all tests) F = 3.17, P > 0.0756 in the two variable model, and F = 3.24, P > 0.0725 in the three variable model. The caveat to be considered is that the R2 shows less than two percent of the grade is predicted by learning profile alone.

    Table 24. Differences by Major, Mean Scores

    Major
    Grade

    SAT V

    SAT M
    A (N=522)

     

    75.4A


    50.0

    59.8
    B (N=24)
    76.3 A

    52.0

    61.2
    C (N=37)

     

    67.5 A B


    47.4

    54.3
    D* (N=3)
    76.2 A

    49.5

    59.8
    E* (N=8)
    63.6 B

    48.6

    59.7

    * may not be considered representative of a population due to small n. A B Means with the same letter are not significantly different.

    RQ4. Do some types of learners do better in general chemistry than other types of learners?

    In phase one, CH102, regression analysis showed a significant negative correlation between NF and grade in both treatment and control groups. Using the standard correlation coefficient notation: for the control group ( ß = -0.1227, P < 0.0371), for the treatment group (ß = 0.09918, P < 0.0440 ). ST was significantly positively correlated with grade in the control group (ß = 0.13138, P < 0.0073), and still positively correlated but not significantly so in the treatment group (ß = 0.07494, P < 0.1284). NT was significantly positively correlated in the treatment group, and not correlated at all (P < 0.9907) in the control group. SF was negatively correlated with grade, but not significantly so, in both treatment and control groups.

    In phase two, CH101, general linear models tests of means were performed with or without using SAT math as a covariate, the SF learners did the worst in the course, and the NT's scored the highest. There is a significant difference between the SF mean of 72.050 and the NT mean of 76.048. With correction for SAT math score in the equation, the least squares mean for SF's is 75.40, and for NT's is 79.13, still significantly different.

    Table 25. Final Mean Grades by Style, Phase 1 and Phase 2., With and Without SATM Covariate

    Phase One, CH102
    Phase Two, CH101


    no SAT

    SATM

    no SAT

    SATM
    SF
    73.95

    74.47

    72.05 a

    75.40 a
    ST
    75.47

    75.88

    73.66 a b

    77.02 a b
    NF
    73.79

    73.33

    73.46 a b

    76.84 a b
    NT
    75.74

    74.40

    76.05 b

    79.13b

    A B Means in the same column with the same letter are not significantly different.

    RQ5.Does knowledge of learning profile and study techniques make more of a difference in achievement in general chemistry for some learning profiles than for others?

    Final grade in CH101 or 102 was evaluated based on relative strength in each quadrant of the Jungian Profile. The null hypothesis of no difference in final grade by cognitive profile, overall and by group, was tested.

    Analyses of grades for each demographic and style dominant group by treatment and control showed differences in order of grade average by cognitive dominance in phase two, CH101, and when averages by style grouping are taken, the averages in the treatment group are higher than in the control group. In phase one, CH102, there was no significant difference in the grades between the treatment and control groups.

    When not corrected for SATM, in phase one, CH102, there were no significant differences in grade average within the control group when analyzed by cognitive dominance. In the treatment group, the NT's and the SF's scored significantly better than the NF's. In phase 2, CH101, within the control group, the NT and NF's scored significantly better than the SF's. In the treatment group, the NT's scored significantly better than the NF's. All other differences were not significantly different. These data include only those students for whom we were able to obtain SAT data from Institutional Research. SAT data were not available for non-traditional students.

    Table 26. Grade by Cognitive Dominance, Control vs. Treatment, Both Phases, Corrected for SATM.

    Style
    102; C

    102; T

    101; C

    101; T
    SF
    76.10A

    69.02 A,

    70.13 A

    75.45AB
    ST
    79.17 A

    68.44A

    71.78A

    76.99AB
    NF
    76.65 A

    66.11 A

    73.25 A

    74.00 B
    NT
    76.33 A

    69.22 A,

    71.98 A

    79.28 A




    Table 27. Grade by Cognitive Dominance, Control vs. Treatment, Both Phases, not Corrected for SATM.

    Style
    102; C

    102; T

    101; C

    101; T
    SF
    73.27A

    74.75 A,

    70.78 B

    73.14AB
    ST
    76.85 A

    74.14AB

    72.99A B

    74.15AB
    NF
    75.06 A

    72.34 B

    75.54 A

    71.33 B
    NT
    74.78 A

    76.92 A,

    75.85 A

    76.26A


    For both tables 26 and 27: C=control group, T=treatment group.

    A,B Means in the same column with the same letter are not significantly different.

    RQ6. How well did students in the CH 102 portion of the study feel the profile described themselves?

    Students in the experimental group were asked on their final exam to respond to additional questions in survey form, using a Likert scale (e.g. A = all the time, B = most of the time, C = frequently, D = occasionally, E = never) relating to their use of the study techniques taught to them, or perception of fit of the profile descriptors (Appendix E ). Responses were tabulated for frequency by cognitive dominance.

    SQ 1. I used the Learning Style Profile techniques in CH 102 : A = all the time, B = most of the time, C = frequently, D = occasionally, E = never.

    The most frequently occurring response was B, most of the time, to C, frequently, with no difference in frequency between students of differing cognitive dominances.

      SQ 2. I used the Learning Style Profile study techniques in my other classes: A = all the time, B = most of the time, C = frequently, D = occasionally, E = never.

    The most frequently occurring response was B, most of the time, with a broad and normally distributed distribution over the remaining answers, with no difference in frequency between students of differing cognitive dominances.

      SQ 3. I felt my individual profile A. fit me perfectly, B. fit me very well, C. fit me pretty well, D. was a bit like me, E. was not like me at all.

    The mode of the pattern of responses was between C and D, with the SF and NF students ranging more towards B and A, and the ST's more towards E. The Chi Square test showed a significant difference in frequency of occurrence between categories of answers. Prob. < 0.049.

      SQ 4. I feel that the study techniques improved my grade in chemistry A. a lot, B. quite a bit, C. maybe a little, D. not at all, E. made my grade worse.

    Responses were evenly divided between B and C, with the exception of the ST's, who responded more frequently with B. quite a bit.

      SQ 5. I feel the study techniques improved my grades in my other classes A. a lot, B. quite a bit, C. maybe a little, D. not at all, E. made my grade worse.

    Responses centered around B and C for all learners.

      SQ 6. The learning styles techniques made my study time A much more productive, B. quite a bit more productive, C. a little more productive, D. no more productive, E. less productive.

    Responses were centered about C, with B second.

      SQ 7. Overall, how do you feel about the learning style profile and its impact on your personal approach to your studies? A. very positive, B. somewhat positive, C. no feelings on way or the other, D. somewhat negative, E. very negative.

    Responses centered about C, no feelings one way or the other, to D. somewhat negative.

      SQ 8. If we continue to use the learning styles profile for future students of general chemistry, would you recommend we A. ask all students to do a profile and attend a workshop? B. ask all students to do a profile, with a workshop optional? C. make the profiles and feedback available but don't hold workshops? D. make the profiles and workshops optional? or E. don't do it at all, it's a waste of time?

    Responses were evenly divided between B, D and E.

      RQ 7. Do students use study techniques appropriate to their learning profile, as measured by the evaluation instrument used? Do those students in the experimental group, who were taught appropriate study techniques, use them to a greater extent than those in the control group?

    Students in both the experimental and control portions of the CH 101 portion of the study were asked about the study techniques they use, in additional survey questions added to their final exam (Appendix F).

      SQ 1. I study with someone else: A = all the time, B = most of the time, C = frequently, D = occasionally, E = never.

    All learners in the control group (no significant differences) responded in a flat distribution centered about answer C, with 38.29% of all students responding frequently. The distribution was more kurtic for the treatment group, also centered about C (42.56%) with fewer E, never or A, all the time responses. The NT's in the treatment group were least likely to study with someone else of all students in the study. Treatment group chi square = 27.152, P<0.007.

      SQ 2. I talk out loud about what I am studying. A = all the time, B = most of the time, C = frequently, D = occasionally, E = never.

    All learners in the control group (no significant differences) responded in a flat distribution centered about C, with the exception of the NF's who were somewhat more likely to chose B (35.56 %).

    In the treatment group ( chi square= 21.042, P<0.050 ) SF's (B = 40.48%, C = 35.71%) and NF's (B = 26.42%, C = 43.40%) responded they talked out loud (B) most of the time or (C) frequently, while ST's and NT's distribution of answers was flatter.

      SQ 3. I memorize lists, equations, or facts. A = all the time, B = most of the time, C = frequently, D = occasionally, E = never.

    All learners in the control group reported memorizing, with especially strong practice in the ST's and NT's, but not significantly different from the SF's and NF's. In the treatment group, also, most students reported memorization, with the NT's lower, but not significantly different in distribution. Approximately 90% of all students in both groups reported memorizing all the time, most of the time, or frequently.

      SQ 4. I practice by doing problems or writing things out over and over. A = all the time, B = most of the time, C = frequently, D = occasionally, E = never.

    Responses centered on B for all learners in the control group. Within the treatment group, SF's and ST's were most likely to respond A or B, and NF's and NT's were most likely to respond B or C. Treatment group chi square= 28.511, P<0.005.

      SQ 5. I read the introductory and summary material for the general overview first, before I read the chapter. A = all the time, B = most of the time, C = frequently, D = occasionally, E = never.

    In the control group, most responses for all learners fell within the C to E range, with the lowest numbers for the ST's, showing few students read the introductory or summary material first. A far greater proportion of NT's, and a somewhat greater proportion of NF's in the treatment group report reading introductory or summary material first. (Treatment group chi square = 22.983, P<0.028.) ST's and SF's in the treatment group responded much the same as those in the control group.

      SQ 6. I look for patterns or logical explanations when I study. A = all the time, B = most of the time, C = frequently, D = occasionally, E = never.

    Virtually all students (>95%) in both groups report looking for patterns or logical explanations at least frequently, with better than 80% at most of the time.

      SQ 7. I use metaphors, or figure out how what I am trying to study is like something I already understand. A = all the time, B = most of the time, C = frequently, D = occasionally, E = never.

    Control group responses centered about B, with SF's and NF's very slightly stronger. In the treatment group, SF and NF learners were more likely to report using metaphors. (P<0.143, not significant.)

      SQ 8. I use color, or pictures, or other creative tricks when I study. A = all the time, B = most of the time, C = frequently, D = occasionally, E = never.

    Even in the control group, NF learners were significantly more likely to use creativity in studying. Control group chi square = 24.514, P<0.017. Other learners responses fell into C-E. The pattern was quite similar in the treatment group, but not statistically significant.