need help

rodp

New Member
#1
i am currrently enrolled in a masters progam and am completing a reserarch proposal. very weak in stats.

I want to measure uncertatinty over three time periods. data will be obtained from a likehart scale. would multiple anova be approprite.

I also want to determine if age or gender would have an effect on uncertaitny over the same time periods. could i use regression analysis.

any help would be great!
 

JohnM

TS Contributor
#2
rodp,

If you could give us a few succinct, complete research hypothesis statements, i.e., what questions do you have right now, and what answers do you want your thesis to deliver, then we can provide guidance on the appropriate methods.

Hope to hear back from you - we're happy to help out.

John
 

rodp

New Member
#3
hypothesis

1. Do subjects with ARVD who receive an AICD have different levels of uncertainty pre and post AICD insertion and a three and six months post AICD insertion?
Null hypothesis: There will be no difference in levels of uncertainty measured at t1, t2, t3, t4.
Alternate hypothesis: There will be a difference in the uncertainty levels measured at t1, t2, t3, t4.

2. Are the measured levels of uncertainty at t1, t2, 3 and t4 different for gender and age?
Null hypothesis: there are no differences in the levels of uncertainty experienced by subjects for gender or age.
Alternative Hypothesis: there are differences in the levels of uncertainty experienced by subjects with ARVD requiring an AICD for gender and age.


This is the hypothesis that I am looking to test!
Rodp
 

JohnM

TS Contributor
#4
There are probably many ways to "slice and dice" tis, but here's what I would do - your thesis advisor may have other viewpoints.....

I would use a repeated-measures ANOVA to do an overall comparison among the time periods, and I would do a dependent-samples t-test to compare t1 vs t2 (pre vs post), since they are the closest in proximity to the AICD insertion.

If the repeated-measures (overall) ANOVA turns up significant, I would follow-up with a trend analysis to see what the trend looks like.

I would treat gender and age as separate hypotheses:

- compare male vs female uncertainty on each time period and examine the patterns of differences, if any

- try regression (do a simple x-y plot first) on the age vs uncertainty, since they are both continuous measures, unlike gender, which is categorical