Hello. I am trying to norm some visual stimuli at the moment. I collected the data (people were rating faces on 4 dimensions) and I am now trying to write the results. I am not sure what analysis to use though. Has anyone a similar experience?

What do you mean? It depends what the data is all about i.e., what question are you asking?

Do you mean you wanted to see how statistically different peoples choices are from each other? If so, that's a Friedman's test. If you're looking to see how much peoples' opinions agree, thats the kelly's concordance test. etc etc..

Need to know a bit more about your data...i.e., is it parameteric, and what do you want to do with it?

I had 40 people rating 160 faces on 4 dimensions (attractiveness, arousal, distinctiveness, familiarity) on 9 point Likert scales.. The reason is that I want to select a highly homogeneous group of these faces (concerning the 4 above dimensions) to use in my further experiments. The thing is that every published article refers to this procedure by just saying: The stimuli had been normed by 40 independent judges. So I don’t really know how people do it and I do not have a prior experience with norming. Do you have any idea?

This "normed" phrase seems strange to me. It seems they are either refering to the fact that they have asked people to rate the pictures and taken the average, or they have morphed them. It sounds like the former.

Right, so collect together your scores in a spreadsheet, and wack it into SPSS and work out which faces are coming up as being consistanly homogenious. I would use a correlation to start with.

What does your supervisor suggest? It seems a little odd to me, I work with face images in the area of emotional expression, and I haven't come across this term in the literature.

I am dealing with memory for faces (face recognition). So I want to make sure that people wont remember a face because is significantly more distinctive, or older or more arousing, or more attractive than the rest, so I want to choose the 128 more "similar" faces out of the 160 that people rated. Then I will add distinctive features to them using photoshop (scars, moles, facial hair etc) to see the effect of this addition on identification performance. I dont have a supervisor; its a long story but basically I am doing the PhD by myself.

What method do you use in your experiments?

Wow, doing the PhD by yourself? That sounds intense. Well, my work is face perception really - so looking at how visual attention modulates perception of emotional faces. I don't use averged faces, I use either Ekman's faces, NimStim faces or faces from the Karolinska institute. These faces have been rated very thouroughly and come with the stats attached! It might be worth looking into.

Where are you at university?

Sending my drafts to editors and receiving feedback is a good supervision though . Well, thats what I want to know. How I can rate the faces by myself. I chose the faces from a website, I cropped them and made sure everyone is looking straight to the camera, wear the same uniform, I removed blemishes, facial hair, moles etc if they had any... so I think they are ok now. But how do I rate them properly?

Can you give me the references with the "stats attached" please?

The stuff you do is of great interest to me! What I am trying to find out at the moment is whether people remember better faces with distinctive features when we force them to pay attention to the "whole" instead of the parts. Like asking: How intelligent you think is this person? So people's attention switches from the parts (and therefore from the distinctive feature) to the whole.

I have also two levels of emotional arousal (high and low) and found differences in response times (people responded to the "negative", scariest faces faster which is a replication of a thousand studies of course). Anyway... sorry I got passionate; its strange talking to someone about my PhD. Being 100% honest noone knows what I do A last thing....

Basically I got all of my stimuli from my supervisor, but you can access most of them online I believe. The best ones are from the Karlinska institute. These are completely controlled for facial expression, clothing, luminance etc etc...

If you want to assess your own images you are going to need a huge sample of people to rate them. Then, you need to do some stats to find out whether the averages are statistically similar.

I think that the amount of people you will need to validate your stimuli is going to make using the pictures you have a bit difficult. I would suggest emailing a few of the key researchers on the papers you have been reading and seeing ifthey will email you their stimuli - you'd be surprised at how willing and generous most people are.

So are you attached to a department?

B

"my drafts to editors and receiving feedback is a good supervision"

... and coming on here and getting advice. i hope bobby is going to get a credit in your thesis

Dear aliby, I had big problems in my department after my supervisor left my Uni to go to another continent! I would either change topic and do what the staff in my department wanted me to do (which is not what I am interested in) or I would do the PhD by myself. And I think I am doing very well since I have already published a paper with only my name on!

Bobby thanks, I think I will follow your advice and mail some authors! Yes, I am attached to a department but I wouldnt like to name it. I have a completely different view of psychology than people in my department and it is really hard working in an environment where everyone has a completely different approach! What about you?

Morning Scamp. Well done on the paper, that is impressive. Are you only in your first year?!

I'm at UCL. Love the department here, but I have to work pretty hard. But its a really rewarding atmosphere.

I'm interested to know about your views on Psychology, could you tell me how you and your department think differently? Hope I'm not being to nosey - I too could talk about Psychology and the academic situation all day!

Do you at least have a research group where you are that does something similar?