Posted by
Wozniak, Paul A on
Sep 01, 2006; 6:46pm
URL: http://spssx-discussion.165.s1.nabble.com/change-of-perceptions-tp1070707p1070710.html
I've done a fair amount of work into age-related attitude effects on
surveys. They occur so commonly that I'd be highly suspicious if you had
not found an association with age. Typically I find a concave bowl
distribution with peak at the pre-24 year olds, sinking to the
misery-laden 39-49 lifestage, and rising continuously to post-85 year
olds (as high as our sample data goes).
One helpful visualization is to do error bars for age (assuming you have
20 or more per year). You can also do this for age categories of, say,
five years.
There's a good deal of psychology literature that lends substantial
support to this aging effect (e.g., happiness studies), just google
some.
As for finding the turning point, my SWAG would be 50-55 years old.
Also, if you have continuous age data, you're very fortunate. Why
categorize except to simplify reporting on bar charts?
Paul Wozniak
Customer Research Analyst
Wisconsin Public Service
Green Bay WI 54302
920 433 1630
[hidden email]
INCLUSION QUOTE: "The outcome of any serious research can only be to
make two questions grow where only one grew before."
Thorstein Veblen, Wisconsin philosopher, 1857-1929
-----Original Message-----
From: SPSSX(r) Discussion [mailto:
[hidden email]] On Behalf Of
Spousta Jan
Sent: Friday, September 01, 2006 4:31 AM
To:
[hidden email]
Subject: Re: change of perceptions
Hi Veena,
1) It is possible that there is no clear "turning point" - rather, the
changes may be continuous, eg. every year of age may change the
perception by 0.01 p. Try to plot the means of the Likert scale against
the age groups say 10 years wide and look for possible turning points.
2) An useful method for detecting breakpoints of this type are decision
trees - run a tree (C5, CART, CHAID or whatever you like) with the
likert scale as dependent/target variable and age as the predictor.
HTH
Jan
-----Original Message-----
From: SPSSX(r) Discussion [mailto:
[hidden email]] On Behalf Of
DR VEENA Joshi
Sent: Friday, September 01, 2006 11:22 AM
To:
[hidden email]
Subject: change of perceptions
Hi,
We have done a survey on perceptions about healthcare system using
likert scale (1 to 5). One of the demographic variable is age which is
continuous.
I categorized age into 3 groups <40, 40 - 60 and >60 and found that
respondents perceptions/attitude differ age groupwise.
I would like to find the truning point at which the difference exists.
How do I get this?
I am trying to get small age groups say 5 years, but the age ranges from
18 to 90 years.
pl. advice.
veena