Checking Consistently Similar Responses

Moderators: statman, Analyst Techy, andris, Fierce, GerineL, Smash

Posts: 13
Joined: Fri Apr 04, 2008 8:21 pm
Location: Pontypridd, Wales

Checking Consistently Similar Responses

Postby skinfreak » Mon Dec 08, 2008 10:34 pm

I windered if SPSS has a way of checking the consistency of data i.e. do the same respondents answer high or low consistently in an effort to establish if there is a bias towards the positive or negative.

The data indicates consistently high responses but I would like to prove it if possible, but am just not sure how.

Posts: 2760
Joined: Tue Jun 12, 2007 12:08 pm
Location: Florida, USA

Postby statman » Tue Dec 09, 2008 2:13 pm

Not exactly since this generally is an analyst determination of whether or not extremes are viable but there are tests that can be run or distribution stats to help. There also are higher order stats like cluster analysis and/or association tests available.

My suggestion is to 1st plot the responses looking for outliers (generally defined as > +/-2 standard deviation but based on the type of data) and look at you skewness & kurtosis stats overall then by key demographics segments viewed side-by-side.

All this is preceded by an emphasis that the type of data (scaly, ordinal, nominal...) always affects what one accepts/looks for.

See the note below

NOTE: Please read the Posting Guidelines and always tell us your OS, the SPSS version and information about your study and data!

Statistical Services

Who is online

Users browsing this forum: No registered users and 1 guest