How to establish correlation for test results?

#1
Hello,

I want to test the correlation of test results with articles read. I have one test and 3 articles. I would like to see correlation of each article to test result and then all articles to test result. Hypothesis is that if you read articles, you did better on test.
I was thinking of Spearman's rank correlation. How would you go about it? Please help :)

Many thanks!
 

Attachments

hlsmith

Less is more. Stay pure. Stay poor.
#2
Not sure how much data you have and where the scores are distributed at, but a beta regression with each article entered as a binary predictor and their interactions also in the model - could get at this.
 

Karabiner

TS Contributor
#3
Hello,

I want to test the correlation of test results with articles read. I have one test and 3 articles. I would like to see correlation of each article to test result and then all articles to test result. Hypothesis is that if you read articles, you did better on test.
I was thinking of Spearman's rank correlation. How would you go about it? Please help :)

Many thanks!
Could you please so kind as to tell us what you mean by test result? Test passed, or score?

With kind regards

Karabiner
 

Karabiner

TS Contributor
#5
So for each paper, you can see whether there's a significant association with score, either by doing Spearman rank correlation,
or by using Mann.Whitney U-test ("paper A read yes/no as grouping variable, score as dependent variable).
And you can correlate number of papers read with score (Spearman rank correaltion).

With kind regards

Karabiner
 
#6
So for each paper, you can see whether there's a significant association with score, either by doing Spearman rank correlation,
or by using Mann.Whitney U-test ("paper A read yes/no as grouping variable, score as dependent variable).
And you can correlate number of papers read with score (Spearman rank correaltion).

With kind regards

Karabiner
Thank you so much! This is tremendously helpful! :D