Date: Mon, 21 Jan 2008 21:05:08 -0500
Reply-To: Richard Ristow <firstname.lastname@example.org>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: Richard Ristow <email@example.com>
Subject: Re: Dummy Coding and Interpreting Regression Analysis
Content-Type: text/plain; charset="us-ascii"; format=flowed
At 03:36 PM 1/18/2008, Justin Meyer wrote:
>I am working to determine if a subjective rating of schools'
>implementation is a predictor of posttest score for students in
>those schools. For the subjective rating of schools' implementation,
>schools are rated as tier 1, 2, or 3, with 1 being the best
>implementation and 3 being the worst. I am using a regression
>analysis, entering all of the variables at the same time. Because
>the tier status consists of three possible responses, I dummy coded
>it into two variables. The first variable is 1 for "Tier 2", 0 for
>"not Tier 2". The second variable is 1 for "Tier 3", and 0 for "not
>Tier 3". Is this the correct way to code this variable for a
It is certainly correct; there are some variations, that are also correct.
>I found a b (unstandardized coefficient) of -10.269 for Tier 2 and
>21.171 for Tier 3. Does this mean that, when all other variables are
>equal, students in Tier 2 score an average of 10 points less on the
>posttest and students in Tier 3 score an average of 21 points more
>on the posttest when compared to Tier 1?
It means exactly that. However, don't forget that these estimates
should be expressed as confidence intervals: "95% confidence interval
is...", and give the range. The 95% confidence interval is the
estimated value, +/- twice the standard error of estimate.
>The output from the regression, except for charts, is pasted below:
Thank you for that. However, you may see that it came through with so
many line-breaks added, that one can't do much with it.
So, does this get you farther?
-Best of luck,
To manage your subscription to SPSSX-L, send a message to
LISTSERV@LISTSERV.UGA.EDU (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
For a list of commands to manage subscriptions, send the command