Factor scores versus mean score for clustering

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Factor scores versus mean score for clustering

sahiba.4a
Dear group members,

I wish to apply cluster analysis on the Big Five Personality traits and further profile these clusters on the basis of their levels of fashion consciousness and status consumption.

The Big Five traits are measured using 20 statements (4 statements for each trait) measured on a 5-point Likert scale. Similarly, 5 statements are used to measure fashion consciousness and status consumption each on a 5-point Likert scale.

I wanted to know, if I can use mean scores for these traits in place of factor scores as the input for cluster analysis?

Further, can I use mean scores of fashion consciousness and status consumption for profiling the clusters instead of individual item scores?
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Factor scores versus mean score for clustering

Kirill Orlov
Where are you going to get factor scores from? You didn't say anything of plans of a factor analysis.

===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Factor scores versus mean score for clustering

Peter Spangler-2
By factors, do you mean the five personalities produced by the Big Five Inventory? 

Sent from my iPhone

On Jun 1, 2017, at 9:45 AM, Kirill Orlov <[hidden email]> wrote:

Where are you going to get factor scores from? You didn't say anything of plans of a factor analysis.

===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Factor scores versus mean score for clustering

sahiba.4a
yes
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Factor scores versus mean score for clustering

David Marso
Administrator
In reply to this post by sahiba.4a
Doesn't this all really have to do with YOUR research hypotheses?
What is you sample size?  That would be an argument as to whether FA is appropriate.
What are the correlations among these 20 statements ?
Have you looked at their reliability coefficients.
Are these roughly block diagonal indicating that your 5 traits are not cross contaminated.
Have you tried a 5 factor extraction with various rotations to surmise the above?
Do you have simple structure or is it a flustercluck of doublets and otherwise uninterpretable chaos?

If you have a strong theory about the factor structure why not go all the way and test the beast using CFA -Confirmatory Factor Analysis-  ?
Are these constructs related to other things you wish to predict?
Do you think these 5 things might be part of a higher order construct?
If so CFA/SEM might be a good path to happy place.

YMMV.  Maybe time to approach a senior advisor in your field as to what are  acceptable analytical approaches.
You could always try all of the below and see where things stick together and where they fall apart.
Educate yourself as to the pros and cons of the various things you attempt.


quote author="sahiba.4a">
Dear group members,

I wish to apply cluster analysis on the Big Five Personality traits and further profile these clusters on the basis of their levels of fashion consciousness and status consumption.

The Big Five traits are measured using 20 statements (4 statements for each trait) measured on a 5-point Likert scale. Similarly, 5 statements are used to measure fashion consciousness and status consumption each on a 5-point Likert scale.

I wanted to know, if I can use mean scores for these traits in place of factor scores as the input for cluster analysis?

Further, can I use mean scores of fashion consciousness and status consumption for profiling the clusters instead of individual item scores?

Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me.
---
"Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis."
Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?"
Loading...