Question

a statistics professor in an attempt to make students more aware of their study habits, began requiring students to maintain a daily activity log for the class. The log consisted simply of recording the hours and a brief summary of what was done for the class on a daily basis, such as reading the text, working on homework, copying notes. For the year previous when the log was not completed, the professor observed 22 of 137 students received an F in the course. For the year the activity logs were required, the professor observed 9 of 122 student received an F in the course.

a) Based on the data does it seem reasonable to assume the student logs are working to reduce the number of students who fail the course?

b) Briefly discuss the potential confounding factors and the validity of making the claim the the student logs is reducing failure in the course.

Answer 1

(a)

We have to perform two sample proportion test.

We have to test for null hypothesis

against the alternative hypothesis

Our test statistics is given by

Here,

First sample size

Second sample size

First sample proportion

Second sample proportion

Corresponding    [Using R-code '1-pnorm(2.148467)']

We reject our null hypothesis if , level of significance.

We generally test for level of significance 0.10, 0.05 or something like these.

So, we can reject our null hypothesis.

Hence, based on the given data we can conclude that the students logs are working to reduce the number of students who fail the course.

(b)

Potential confounding factors may be as follows.

• Number of students has reduced (by 15) in the current year with respect to previous year. There is a possibility that most of the students who received F in the last year left the course which made the proportion of getting grade F lower.
• If the classes of this two years consist of different bunch of students, there may be a possibility that students in the current year are more intelligent or are more aware about their study.
• There is a possibility of difference in toughness in level of question in these two years.
• There is a possibility that to prove his method as an effective one in front of others, the professor either made questions easier or performed easier checking while evaluating copies.

If any of above confounding factors impose enough importance in diverting our observed data, then there is no validity of making our claim as we concluded in part (a).