1. ## Which Tests?

Hello!

I am considering a study I'll note below, but have limited statistical experience. Accordingly I am seeking insights on which tests are appropriate given the data and what I am looking to test.

In this potential study I would examine three variables, let us call them variable A, variable B, and variable C. All are interval variables. My prediction is that higher scores of variable B are related to higher scores of variable C (previous regression analyses have shown this to be true). Further, I predict that the magnitude of variable A will impact the effect of variable B on variable C. That is, I predict that subjects with low scores on variable A will see a greater impact from variable B on variable C than subjects who have high scores on variable A.

The literature has indicated that there are other factors besides variable B which impact variable C. Accordingly I'm looking to incorporate these other factors into the analysis in order to better isolate the impact of variable B on variable C.

Thus far I've been advised to utilize the hierarchical regression method as an exploratory analysis to see if adding a third variable will change the predictive power of the model. In phase two I would then incorporate the variables deemed significant in the final model. I am unclear on how to account for the other factors the literature has noted that may impact variable C in this analysis, however.

If you have any insights on the information above or have suggested readings to help myself better understand how to approach this type of problem, I would greatly appreciate your input.

Thanks!

2. ## Re: Which Tests?

In this potential study I would examine three variables, let us call them variable A, variable B, and variable C.
Such abstractions usually make things more difficult.
what are the study objectives, what will be the actual
variables variables and measurement, and how large
will the sample size be?

With kind regards

K.

3. ## Re: Which Tests?

Hello,

Sorry for the terse language, I'll be more clear here. I am intending to measure the impact satisfaction level related to living on campus in a university setting has on student GPA. On campus living satisfaction is to be measured by a survey distributed annually to the on campus student body. The third variable in question would be student dropout proneness, as measured by the College Student Inventory. My prediction is that students who are considered dropout prone, or have a high level on the dropout proneness measure, will have a higher GPA when they also have had a positive on campus living experience. Conversely, I predict that students who are not considered dropout prone will not have their GPA moderated at all by on campus living satisfaction level.

The literature has shown certain variables to be predictive of college GPA. The top two are generally college entrance exam scores (ACT/SAT) and high school GPA. My hope is that by incorporating these into the statistical analysis I will be able to better isolate an on campus experience effect.

One potential hurdle is a possible correlation between college entrance exam scores/high school GPA and dropout proneness. While this is conjecture because I have yet to search the literature on the topic, I would guess that students who are dropout prone have habits or environmental factors that likely cause both dropout proneness to be high and entrance exam/high school GPA scores to be low. This may be a potential hurdle.

The sample for the on campus experience survey is typically 300 to 400 and can be expected to be similar for this purpose. This would be the variable that would set the sample size.

I hope this makes my intentions more clear, and I appreciate any input or advice.

Thanks!

 Tweet

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts