I'm trying to analyse differences in saccadic eye movements (e.g. amplitude, velocity, and latency) between two groups of participants (A and B). My study is small, Group A has 12 participants and Group B has 15 participants. However, there were baseline differences in visual acuity (LogMAR values) based on the Wilxcon Test.
Anyone has any idea how to I adjust or account for differences in baseline visual acuity data to study the differences between the two groups as my visual acuity data is positively correlated with my saccadic amplitude data? Is there a way to account for them in SPSS with non-parametric statistics?? fyi: my data had violated normality as well.
Hi June Wei,
in our previous paper we used linear mixed models (LMM, in SPSS) to analyze saccades parameters.
https://www.ncbi.nlm.nih.gov/pubmed/28619425
I suppose that you have dozens of trials (saccades) per participant but you are analysing only the mean value => 1 mean amplitude per subject. You are losing a lots of information therefore I suggest you to use all trials with the long format of LMM;
You can make acuity a categorical variable (unless you want to cover all acuities) and include that in your model as a fixed factor. Then you can test any interaction between group and acuity. Hope this helps
Thanks a lot for your help! I've read through your paper. Given that you've three dependent variables (sacccade latency, amplitude and peak velocity), I'm curious if you performed a multivariate LMM or just a LMM that is performed individually for each of your dependent variables? are there implications for performing the latter procedures individually for each of dependent variables over the multivariate LMM?
It is correct to perform a linear mixed model (LMM) analysis for each outcome;
The only thing to keep in mind is that peak velocity and amplitude are correlated and their analyses is somehow redundant;
Thanks Antionio, I was able to perform LMM. The one thing I forgotten to mention is that I've got both an anti-saccades and a reflexive saccades paradigms for which both Group A and and Group B performed as well. So technically the study is of a split-plot ANOVA design. I didn't encounter problems with performing LMM when ID was labelled as random effect, and Group as fixed effect i.e to analyse anti-saccades and reflexive saccades independently. I thought that Paradigm was a repeated measures (given that both groups performed both anti and reflexive saccades). However, I encountered an error message stating: 'The levels of the repeated measures effect are not different for each case within a repeated subject'. Was I incorrect to set Paradigm under repeated measures? I'm not sure how to overcome this situation.
Paradigm is one of the conditions used, therefore is a "fixed factor"
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS