0
$\begingroup$

I've tried using google sheets/excel, but I can't figure how to run a 2 factor without repetition ANOVA (if that's the appropriate test for this data set). I think it's the way I set the spreadsheet up... I'm thinking I'll have to dummy code in R for the different variables?? I describe my hypothesis below. https://docs.google.com/spreadsheets/d/1czrXmJuxvwalDwNDLHIvt64QqNb8cSTahEEK8DIiQPM/edit?usp=sharing

the study: "right and left visual hemifields and duration of the presented stimulus were manipulated using a brief lexical decision task (differentiating words from non-words) using emotional and neutral words. (and non-words but we are ignoring that in terms of data analysis) To measure lexical decision, participant’s reaction time and accuracy of each stimulus response were analyzed." I was thinking 1 for emotional and left, 2 for emotional and right, 3 for neutral and left, and 4 for neutral and right The data spread sheet is kind of confusing, but each row is a participant, each column is a specific word and presentation, and each cell gives that subjects reaction time in response to that specific stimulus. I just need suggestions on how I should run the data with the appropriate tests... doesn't matter if it's in excel, google sheets, SPSS, or code in R. I just keep ending up running it with 84 different variables, but the 84 is the 84 presentations of different stimuli (words on the screen, appearing in either right or left hemi) for each of the 26 participants. Any ideas?

|cite|improve this question
$\endgroup$
1
$\begingroup$

First, you have two independent variables: Emotionality (neutral or emotional) and side (right or left). Those variables have to be in your spreadsheet somewhere.

Second, R and SPSS both do the dummy coding for you, but questions about how to code are off-topic here.

Third, since each participant was assessed multiple times (with different words) your data are not independent and, therefore, ANOVA is not appropriate. You probably want some sort of multi-level model. These are complex. I suggest you hire a consultant.

|cite|improve this answer
$\endgroup$
  • 2
    $\begingroup$ "questions about how to code are off-topic here." This is true for the programming meaning of the word "code," questions about how to code categorical data in different way (reference coding, orthogonal coding, effect coding, etc.) are totes on topic. (Which you know, but some readers might not :) $\endgroup$ – Alexis Dec 6 '18 at 2:33

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.