My sincere apologies if this question has been posted earlier. I am looking for help in terms of how to do Intraclass Correlations and find out how to do within group reality in SPSS.

I am currently pursuing my PhD in Strategy and being 50 years old, I am just not able to understand the formulas given in Bliese & Haverson (1996).

My situation is this. I have done survey on team performance and other team capabilities within my service organization. I have got responses from 210 Team leads) and from 2+ team members under each of these leads. The Team lead has also given their opinion on team performance (7 point likert scale) and the team members have also done the same.

I am now choosing Team as my unit of analysis and hence I am aggregating the individual team member responses to each item to get a mean score at the team level from the team members.

As mentioned earlier, a team can consist of anywhere between 2 and 5 members apart from the team leader with the Average team member / team being around 2,7 or so.

Since i am not planning to to Multi Level modelling, I have decided to go this route based on similar work done by

Groves, K. S., & Feyerherm, A. E. (2011). Leader Cultural Intelligence in Context: Testing the Moderating Effects of Team Cultural Diversity on Leader and Team Performance. Group & Organization Management, 36(5), 535–566.

Now I want to do this with the data in SPSS.

Pls let me know if you know how to do the same.

Thanks in advance.

  • $\begingroup$ You stated what you would like to do but I can find no specific question in your post. $\endgroup$
    – g g
    Commented Aug 13, 2019 at 9:26
  • $\begingroup$ I don't know how to do Intraclass correlations & Within-group reliability in SPSS and I was hoping someone can help me in this regards on how to go about using the data that I have collected or atleast point me to some video / article which explains the process of doing the same $\endgroup$
    – Shankar
    Commented Aug 13, 2019 at 9:48

1 Answer 1


Intraclass correlations in SPSS for a multilevel model are not currently computed directly. However, you can get the needed parts and do it by hand.

The ICC is a proportion of variance:

     Mean Square(Between Groups) - Mean Square(Within)
 Mean Square(Between) + ((#groups-1) * Mean Square(Within))

Use the MIXED procedure, like this:


to get the Mean Squares.

I found a video which looks to me like it uses MIXED to do a multilevel model:



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.