Confused regarding interaction plot

#1
Hi guys,

I am having trouble understanding an interaction result which does not seem to make sense. I have double-checked everything is coded correctly, but every time I run the analysis, I get the same spike in women participants, which does not correspond with the actual numbers reported in the descriptives.

If anyone could point me in the right direction in what I may be doing wrong, I would greatly appreciate it.

Best regards.
 

Attachments

Dason

Ambassador to the humans
#2
I think you messed up your labels either importing the data or in displaying it here. All of the given group means correspond to a column in your data but not always to the one you have it labeled as.
 
#3
I think you messed up your labels either importing the data or in displaying it here. All of the given group means correspond to a column in your data but not always to the one you have it labeled as.

Thanks for getting back to me. I've checked my labels and still can't work out where I am going wrong, I've attached a screenshot.

I'm carrying out a repeated measures within-participant ANOVA
 

Attachments

#4
I'm still really stuck with this guys, I'm not sure how I can illustrate any more that I am making an effort, I have even tried to recode things the other way around (screenshot attached) but I am totally stumped as it's still showing females having far higher levels of recall than they should...
 

Attachments

hlsmith

Not a robit
#5
Yeah you have some coding of groups issue. Just start over, recreate data and code. That may help find your issue. What program are you using?

On a side note, i dont get how you are trying to pull off repeated ANOVA with interaction with 36 variables?
 
#6
Yeah you have some coding of groups issue. Just start over, recreate data and code. That may help find your issue. What program are you using?

On a side note, I don't get how you are trying to pull off repeated ANOVA with interaction with 36 variables?
That is what is confusing me... I was told by the tutor to use this data and run a 3x2 factorial repeated measures ANOVA, but it doesn't seem to be the best analysis to use. I have double-checked with them and they have re-iterated that its definitely repeated measures.
 
#7
Just to give an explanation in regard to the question; it is:

A study was conducted to investigate biases in face recognition. Researchers examined participants’ ability to recognise faces of male and female models from three different ethnic groups (Asian, Caucasian, and Black Caribbean). Six participants were presented with equal numbers of male and female faces from each ethnic group. After a one hour delay, they were tested on their ability to recognise these same faces. The data below represent the total number of correctly recalled faces for each participant in each experimental condition. You will need to set-up an SPSS data file and enter these data for analysis.

I don't know if the tutor has given me the wrong data, but I would have guessed that there would be a before/after category, but this has not been provided. Thus I'm stumped...
 
#8
If anyone can give me any further heads up before I go back to the tutor I'de appreciate it, I've already had 'the look' from them telling me not to ask any more questions, and I don't want to go accusing them of suggesting the wrong analysis (as suggested by hismith) before I know for sure. But I still can't seem to make these data fit!
 
#9
Well, thanks to those who replied I greatly appreciate it. I've tried recoding everything and re-entering it all as suggested but I am getting the same results. It's very frustrating, I've spent the last 8 hours trying to get my head around this, but I just can't make it fit. I thought I was getting somewhere in my understanding of SPSS but I just can't seem to understand this particular issue. Thanks anyway.
 
#10
How many faces did they review? is it the gender and race in the image, seems so? Why not have a binary variable for correct or not and have guesses nested in guessers? Then that would be similar to repeated reg, but would be multilevel logistic regression.