minimum ICC for HLM

[Stephen Porter]

Hi all,

I'm looking for some cites as to what the minimum ICC should be to justify multilevel modeling (I've seen .05 and .10). I'm not saying I believe in them, I just need some cites.

Cheers,

Steve

--

Stephen R. Porter
Professor, Department of Leadership, Policy and Adult and Higher Education

​Visiting Scholar, Education and Workforce Development, RTI International​

Box 7801
North Carolina State University
Raleigh, NC 27695
srpo...@ncsu.edu

stephenporter.org

[Stephen DesJardins]

Steve,

It gives me great pleasure to offer the following (probably) unhelpful (given the question) answer of....IT DEPENDS.  As you might expect, the answer to this question is more complex than a simple "rule of thumb" or single number that applies in all cases.  See Kreft & de Leeuw, 1998, pp. 9–10, a copy of which can be found in the link below, for a short explanation: 

http://gifi.stat.ucla.edu/janspubs/1998/books/kreft_deleeuw_B_98.pdf    They have some other cites that may help you out too. 

I was chatting with Julie Posselt a bit ago and she said Raudenbush told her (verbally) a rule of thumb, but that he didn't or wouldn't put it in writing, maybe because of what Kreft/de Leeuw say ("it depends").  She can chime in about that, if she cares to. 

Have a nice day!!  

Steve

--
You received this message because you are subscribed to the Google Groups "Quantitative Higher Education Research" group.
To unsubscribe from this group and stop receiving emails from it, send an email to qher+uns...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Stephen L. DesJardins

Professor, Center for the Study of Higher and Postsecondary Education

Professor, Gerald R. Ford School of Public Policy 

“Don’t hate the black, don’t hate the white. If you get bitten, just hate the bite.”  

Sly Stewart (of Sly and the Family Stone) on The Ed Sullivan show, 1968. 

Contact Information: 

2108-C School of Education Bldg.

University of Michigan

610 E. University Ave.

Ann Arbor, MI  48109-1259

Voice: 734.904.7857

email: sd...@umich.edu

[Stephen Porter]

I agree; this is for a validity paper for an instructional log for teachers that some colleagues and I developed. 

We're arguing that individual items on the log, to be any good, have to exhibit some variation between teachers, so we were thinking of .10 as the minimum benchmark, because that's often where folks will say you need to use HLM. Hence the need for a cite.

Cheers,

Steve

--

Stephen R. Porter
Professor, Department of Leadership, Policy and Adult and Higher Education

​Visiting Scholar, Education and Workforce Development, RTI International​

Box 7801
North Carolina State University
Raleigh, NC 27695
srpo...@ncsu.edu

stephenporter.org

[Stephen DesJardins]

See this simulation study which may help.  Esp. page 6. 

http://www.joophox.net/papers/p090101.pdf

[.]

That Maas and Hox piece is a nice one. Here's something Ron put together in response to a similar query we had from one of our readers of “An Intro to MLM". Maybe this will help, too. 

Scott

-----
Scott L. Thomas, Dean 
School of Educational Studies 
Claremont Graduate University 
Editor, Journal of Higher Education 
http://www.cgu.edu/pages/5659.asp