0
$\begingroup$

Any and all help would be much appreciated! I've detailed a bit of information below, but if I left anything out please let me know

My hypotheses are (info: NA= novel asssembly; S= selections or assembly piece):

1) H0: NA, and S will not affect the completion rates of the assembly tasks. H1: NA, and S will affect the completion rates of the assembly tasks.

2) H0: Time to complete an assembly between, NA, and S will be the same H1: Time to complete an assembly between, NA and S will be different.

3) H0: Completion rates for color-coded and non color-coded novel assemblies will be the same. H1: Completion Rates for color-coded and non color-coded novel assemblies will be different.

4) H0: Time to complete an assembly between color-coded and non color-coded novel assemblies will be the same. H1: Time to complete an assembly between color-coded and non color-coded novel assemblies will be different.

In my experiment, I have 3 independent variables at 4x3x2: - Assembly difficulty (low, medium, high, very high) which I believe is an ordinal variable - % increase in assembly pieces (0%, 25%, 50%) which I believe is an ordinal variable - Color coded assemblies (color coded, not color coded) which I believe is a nominal variable - I also gave participants a pre-test to measure their spatial rotation ability, I think this may be the 4th IV with 1 level, but I'm not sure what to type of variable it is.

I ran the study as a within-subects 4x3x2. I measured two dependent variables, completion time (coded in seconds) and completion rate which was a binary variable (success/failure). I used a complete counterbalanced design, similar to the one found here

At this point it feels like there are a ton of different methods I could use to analyze the data. I have some general ideas, but I don't feel like I know enough that I feel confident running the analyses and writing up the results. I'll bullet out what I'm thinking because bullets are easier:

  • ANOVA for Assembly difficulty x completion time / completion rate

  • ANOVA for % increase in assembly pieces x completion time / completion rate

  • ANOVA for color coded/not color coded x completion time / completion rate

  • rank all of the scores from high -> low and see if there is a correlation with pre-test scores and completion times/rates

I have an ok understanding of R (thanks to the seemingly infinite resources on the internet) and plan on using that as my tool for analyses.

I know there are quite a few tests that measure whether or not your data satisfies the assumptions required for particular methods. We never got too deep in these at school thus that's an area I know even less about. Any help in that department would be awesome.

$\endgroup$
0
$\begingroup$

First of all, why do analyses for both completion time and rate, increasing the problem of multiple testing, if they are linearly related? Secondly, the above-mentioned approaches are valid as long as you are not interested in the interaction between the two factors: e.g. what if the effect of assembly difficulty depends on whether there is color-coding (and it probably does). To address that, you can try a two-way or three-way (to include % increase) ANOVA. Pre-test scores can be used as a covariate in the analysis but whether you can partition all these effects depends on the sample size. Here's a good discussion of model assumptions https://stackoverflow.com/questions/2933253/homoscedascity-test-for-two-way-anova

| cite | improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.