LISTSERV at the University of Georgia
Menubar Imagemap
Home Browse Manage Request Manuals Register
Previous (more recent) messageNext (less recent) messagePrevious (more recent) in topicNext (less recent) in topicPrevious (more recent) by same authorNext (less recent) by same authorPrevious page (March 2002)Back to main SPSSX-L pageJoin or leave SPSSX-L (or change settings)ReplyPost a new messageSearchProportional fontNon-proportional font
Date:         Fri, 22 Mar 2002 11:01:46 -0600
Reply-To:     Adam McKee <Adam.Mckee@usm.edu>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Adam McKee <Adam.Mckee@usm.edu>
Subject:      Re: Power and Multiple Regression
Comments: To: Marc Mackert <Marc.Mackert@state.co.us>
Content-Type: text/plain; charset=us-ascii; format=flowed

I've heard of rules of thumb like that; the trouble is they have no basis and there are as many rules of thumb as there are thumbs present during such discussions. So long as you have enough degrees of freedom, you can get a computer to print out some results. Unless your power is adequate, they will be meaningless. The ultimate subjective question is: What degree of beta error is acceptable?

Marc Mackert wrote:

>Hello to all! > >I am a relatively new member to this list and have found it extremely >helpful to read the postings and responses. I now have an occassion to >post my own question and am hoping for similar helpful responses! I am >helping a friend with the statistics for his dissertation, for which he >is using moderated hierarchical multiple regression to look at the >effects of 2 predictor variables and 3 moderator variables on the DV. He >currently has 69 subjects, which if we look at a power analysis is not >sufficient at this time to even find a medium effect size, and the >literature suggests that he will need to be able to detect a small >effect size. I am suggesting that he find more data, yet he is also >hearing from another source that all he needs for this type of test is 4 >cases per variable in the equation. I have never heard of this before, >and it seems like an incredibly small sample size (for example, 20 >participants for this study!) so I was wondering if anyone else has >heard of this 'rule of thumb' and/or what your thoughts are on this. > >Thanks in advance for any help! > >Marc > >Marc J. Mackert, Ph.D. >Quality Assurance Analyst >Administrative Review Division >1580 Logan St. >(303) 894-7510 x24 >

-- -- Adam J. McKee, Ph.D. Director, Mississippi Statistical Analysis Center

University of Southern Mississippi Department of Criminal Justice Box 5127 Hattiesburg, MS 39406-5127

Phone: (601) 266-4760 Fax: (601) 266-4391 E-mail: Adam.Mckee@usm.edu


Back to: Top of message | Previous page | Main SPSSX-L page