LISTSERV at the University of Georgia
Menubar Imagemap
Home Browse Manage Request Manuals Register
Previous messageNext messagePrevious in topicNext in topicPrevious by same authorNext by same authorPrevious page (September 2006)Back to main SPSSX-L pageJoin or leave SPSSX-L (or change settings)ReplyPost a new messageSearchProportional fontNon-proportional font
Date:         Tue, 5 Sep 2006 08:46:50 -0700
Reply-To:     Dominic Lusinchi <dominic@farwestresearch.com>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Dominic Lusinchi <dominic@farwestresearch.com>
Organization: Far West Research
Subject:      Re: Comparing groups for more n more than 30
Comments: To: Manmit Shrimali <MShrimali@marketrx.com>
In-Reply-To:  <9D4BEF69F83C9744AB34E5FE2F78F46DB5E14C@matrix.mrxin.marketrx.com>
Content-Type: text/plain; charset="us-ascii"

Manmit,

You CAN perform statistical tests with samples less than 30. The issue is: what is the shape of the parent distribution that has generated the data? If you can safely assume that it is "normal", you can use any parametric test. If it is not normal because your sample size is small, the Central Limit Theorem does not apply, especially if the parent distribution is very different from a normal distribution. In that case, you can rely on nonparametric tests.

Anytime a statistical test is used you have to check what assumptions must be met for it (the test) to be effectively used.

This answers only part of your query. Others on the list more adept with syntax can help you with the other part: setting a benchmark.

Good luck, Dominic

Dominic Lusinchi Statistician Far West Research Statistical Consulting San Francisco, California 415-664-3032 www.farwestresearch.com

-----Original Message----- From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of Manmit Shrimali Sent: Monday, September 04, 2006 11:19 PM To: SPSSX-L@LISTSERV.UGA.EDU Subject: Comparing groups for more n more than 30

Hello All,

I am facing bit tricky situation. I frequently run statistical significance testing, column means and column proportions in custom table. As a rule of thumb, I understand that we should not do significance testing if n (base) is less than 30. However, SPSS will still generate the p-values and significance testing even being the valid n less than 30. Is there any way, to set benchmark for SPSS, available to set rule of comparing groups and generating stat testing only if the valid n of groups are more than 30?

Any help will be of great help.

Regards,

Manmit


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