Results

 < Day Day Up > 



Results for the experiment were based upon the number of questions attempted as well as the number of correct questions completed. Descriptive statistics for each group's exams are presented in Tables 2 and 3.

Table 2: Descriptive Statistics Number of Questions out of 52 Answered Correctly

Group

First Exam

Second Exam

Group A (Slow Slow)

Mean= 17.49

Std. Dev =8.57

Mean=24.17

Std. Dev =8.87

Group B (Slow Fast)

Mean= 15.41

Std. Dev =8.26

Mean=22.76

Std. Dev =10.33

Group C (Fast Slow)

Mean= 15.97

Std. Dev =8.44

Mean=21.17

Std. Dev =9.08

Group D (Fast Fast)

Mean= 17.36

Std. Dev =9.21

Mean=23.61

Std. Dev =8.33

Table 3: Descriptive Statistics Number of Questions out of 52 Attempted

Group

First Exam

Second Exam

Group A (Slow Slow)

Mean=19.95

Std. Dev =8.59

Mean=27.05

Std. Dev =8.43

Group B (Slow Fast)

Mean=18.03

Std. Dev =8.25

Mean=25.35

Std. Dev =10.15

Group C (Fast Slow)

Mean= 18.77

Std. Dev =8.60

Mean=24.23

Std. Dev =8.90

Group D (Fast Fast)

Mean= 19.95

Std. Dev =9.28

Mean=27.05

Std. Dev =8.48

Descriptive statistics indicating mean and standard deviation for the number of correct answers for each group on each examination are shown in Table 2. A Goodness of Fit test confirmed that results of the first test for four groups were approximately equal. The Chi Squared statistic was 0.165 with three degrees of freedom while the critical value of Chi Squared for an alpha of .01 was 0.20.

Table 3 shows the descriptive statistics for the number of questions attempted on each test for each group. A Goodness of Fit test confirmed that these groups were from the same population. The calculated Chi Squared value was 0.0005 while the critical value with an alpha of 0.01 and three degrees of freedom remained at 0.20.

Table 4 displays the percentage improvement from test one to test two for each experiment group. The results indicate that there was an improvement in the number of correct answers in all four groups. This result is not unexpected. Familiarity with the test, or short-term learning, as indicated by the average improvement by the control groups, accounted for much of this gain. Categories A and D serve to isolate the short-term learning effects. This is the outcome of having students taking both tests on the same class machine. The percentage increase for both categories was very similar, 38% and 39% respectively. This results in measured short-term learning effects of 38.5% (the simple average of the two). By subtracting the control groups (groups A and D) average improvement from the other groups, (groups B and C), the impact of the change in processor speed is isolated. The resulting isolated impacts of the changes on processor speed were an improvement in the group B category productivity (the group that moved from a slower to computer to a faster one) of more than 9%. The group that moved from a faster computer to a slower one, group C, suffered more than a 6% decrease in productivity.

Table 4: Improvement in Number of Correct Questions
 

Test 2 Slow

Test 2 Fast

Test 1 Slow

38%

Mean = 6.7

St Dev. = 8.85

48%

Mean = 7.4

St Dev. = 7.94

Test 1 Fast

33%

Mean = 5.2

St Dev. = 7.45

39%

Mean = 6.7

St Dev. = 6.95

These results were tested for significance at the .01 level using the t-Test: Paired Two Sample for Means. The critical value of the Student t statistic for the group with the smallest number of observations was 2.738. Since all calculated Student t values were greater than the critical value for the smallest group, all differences were found to be significant. Null Hypothesis 1 is rejected. A change in microprocessor speed does make a difference in the quantity of tasks accomplished as measured by the number of correct answers. Table 5 contains the calculated Student t and associated p-value for each group.

Table 5: Student t and p-value (Correct)
 

Test 2 Slow

Test 2 Fast

Test 1 Slow

t=6.156

p<.001

T=5.401

p<.001

Test 1 Fast

t=4.127

p<.001

T=4.269

p<.001

One explanation for an increase in correct answers is that subjects attempted more total questions due to their familiarity with the type of questions on the test and thus increased proportionally the number of correct questions answered. Table 6 indicates the change in percentage of questions attempted for each group. The average percentage change in the number of questions attempted by the two control groups, groups A and D, is 36.2%. After accounting for the learning effect demonstrated by the control groups, the students moving to a faster computer from a slower one, group B, attempted 4.4% more questions. The group that moved from a faster computer to a slower one, group C, attempted 7.1% fewer questions. The change in the number of questions attempted by the experimental groups can be attributed to the change in microcomputer processor speed.

Table 6: Percent Improvement in Number of Questions Attempted
 

Test 2 Slow

Test 2 Fast

Test 1 Slow

36%

Mean = 6.9

St. Dev. = 7.33

41%

Mean = 7.5

St. Dev. = 8.34

Test 1 Fast

37%

Mean = 5.3

St. Dev. = 7.44

39%

Mean = 7.1

St. Dev. = 9.23

These differences were tested for significance at the .01 level using the t-Test: Paired Two Sample for Means. The critical value of Student t statistic for the group with the smallest number of observations was 2.738. Since all calculated Student t values were greater than the critical value for the smallest group, all differences were found to be significant. Therefore, Null Hypothesis 2 is rejected, a change in microprocessor speed does make a difference quantity of tasks accomplished as measured by the number of tasks attempted. Table 7 contains the calculated Student t and associated p-value for each group.

Table 7: Student t and p-value (Attempted)
 

Test 2 Slow

Test 2 Fast

Test 1 Slow

t=6.312

p<.001

T=5.135

p<.001

Test 1 Fast

t=4.461

p<.001

t=4.395

p<.001

If students learned how to solve additional problems while taking test 2, the quality of work should have improved regardless of microcomputer processor speed. That is, the change in the number of correct answers should change relative to the processor speed, not the number of questions attempted. Table 8 shows that there was little improvement in any group when considering the number of correct questions divided by the total number of questions attempted. There also was little change between groups indicating that the results appeared to be independent of microcomputer processor speed.

Table 8: Percent Improvement in Number of Correct/Attempted Questions
 

Test 2 Slow

Test 2 Fast

Test 1 Slow

3%

Mean = 2.6%

St. Dev. = 10.96

4%

Mean = 3.4%

St. Dev. = 11.73

Test 1 Fast

3%

Mean = 2.9%

St. Dev. = 17.97

4%

Mean = 3.4%

St. Dev. = 13.27%

These differences were tested for significance at the .01 level using the t-Test: Paired Two Sample for Means. The critical value of Student t statistic for the group with the smallest number of observations was 2.738. Since all calculated Student t values were less than the critical value for the smallest group, none of the differences were found to be significant. Therefore, Null Hypothesis 3 cannot be rejected. There is no evidence that microprocessor speed has any effect on the quality of tasks accomplished as measured by dividing number of tasks attempted into the number correct. Table 9 contains the calculated Student t and associated p-value for each group.

Table 9: Student t and p-value (Correct/attempted)
 

Test 2 Slow

Test 2 Fast

Test 1 Slow

t=1.827

p=.075

t=1.808

p=.080

Test 1 Fast

t=0.678

p=.502

t=1.502

p=.143



 < Day Day Up > 



Advanced Topics in End User Computing (Vol. 3)
Advanced Topics in End User Computing, Vol. 3
ISBN: 1591402573
EAN: 2147483647
Year: 2003
Pages: 191

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net