Date: Wed, 28 Sep 2005 12:02:42 -0700
Reply-To: Jean Campbell <campbel@u.arizona.edu>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: Jean Campbell <campbel@u.arizona.edu>
Subject: Re: Recoding
In-Reply-To: <14CC168906C71545B71C41A209D67ECE6330D9@beech-ex.beechacres.local>
Content-Type: text/plain; charset="US-ASCII"
Egon,
Thanks for your offer of assistance, I hope I can be clearer.
My goal is to recode all of my profession data into one of the following
categories:
1. Healthcare-medical
2. Healthcare-behavioral
3. Other human services
These are mutually exclusive categories, both conceptually and by virtue of
how the data was collected. The first two, above, are from a forced-choice
list presented to the person filling it in, the third category comes from a
combination of other forced choices ("police officer," "manager/hr,
"coach/fitness") AND a write in "other" option on the same forced choice
list where the person identified a specific profession that is "human
services, as so....
1 healthcare-medical
2 healthcare-behav
3 social worker
4 manager/hr
5 student
6 retired
Etc...
12
13 Other (specify _________)
In SPSS, this means there are two variables: one that represents any of the
12 possible professions (numeric) and one that represents a write-in
response (string). In the interest of clarity, these are variables A and B.
So I want to have everyone who choose "healthcare-medical" and everyone who
wrote in something I adjudicate as "healthcare-medical" to be recoded to
"1." And so forth with "behavioral" as "2" and "human services" (which I am
choosing from various other forced choice and write-in categories) as "3".
Thus, I want to recode variables A and B need to be combined in one recoded
variable as 1, 2, or 3.
Does this help?
Jean
Jean Campbell, MPA
campbel@u.arizona.edu
520/626-1085
-----Original Message-----
From: Kraan, Egon [mailto:ekraan@BeechAcres.org]
Sent: Wednesday, September 28, 2005 11:45 AM
To: Jean Campbell
Subject: RE: Recoding
Depends on what your two variables are. There would be many different
possibilities to what the data contained in the two variables are, which
would lead to handling the data in different ways. Please be more
specific about the data in your cells. For example, is the string
variable the job category, and the numeric variable the level of
experience, i.e. programmer 1, programmer 2, programmer 3, etc... or
does the numeric variable reflect the same information as the string
variable, where lawyer = 1, programmer = 2, driver = 3, etc....
Please be more specific about your data and the result you want to see.
Thanks
Egon
======================================
Egon Kraan
Beech Acres
Evaluation & Research
ekraan@BeechAcres.org
http://www.beechacres.org
-----Original Message-----
From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of
Jean Campbell
Sent: Wednesday, September 28, 2005 2:11 PM
To: SPSSX-L@LISTSERV.UGA.EDU
Subject: Recoding
Hello listers,
I am a fairly new user without too much statistical training, and am
trying
to do some recoding. Would appreciate any help on solving this
seemingly
simple problem, have found no reference in my manual.
I have two separate variables that list a person's profession, one is a
string variable and the other is numeric. I want to recode both
variables
into one new variable "profession." But so far I can only recode one
variable at a time into a new variable...
Any help?
Jean
Jean Campbell, MPA
campbel@u.arizona.edu
520/626-1085
-----Original Message-----
From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of
Art
Kendall
Sent: Wednesday, September 28, 2005 10:44 AM
To: SPSSX-L@LISTSERV.UGA.EDU
Subject: Re: SPSS vs. Excel?
If you are really only doing basic descriptive statistics and no ANOVA,
Regression, random number generation, etc. Then you may be able to get
by with a spreadsheet. Various statistical discussion lists have
covered using Excel for statistics. There are numerical analysis and
other problems
here are a few of the views.
http://www.npl.co.uk/ssfm/ssfm1/validate/testing/excel.html
Several items on software that includes Excel or Excel itself.
http://www.pages.drexel.edu/~bdm25/publications.html
<http://www.pages.drexel.edu/%7Ebdm25/publications.html>
http://www.stat.uiowa.edu/~jcryer/JSMTalk2001.pdf
<http://www.stat.uiowa.edu/%7Ejcryer/JSMTalk2001.pdf>
I advise my clients to use a statistical package for statistics. Spread
sheets (QuattroPro, Lotus, Excel) are all great for use as
spreadsheets. As statistical packages think of "please pass the hammer,
I want to drive in this nail with the funny spiral ridge around it". If
they are doing work with real life consequences (public policy
oversight, etc.) I recommend SPSS as the primary package for several
reasons. The packages are very solid mathematically. SPSS stands out
because of the human factors aspects. However, there are occasionally
procedures that are not in SPSS, so other packages might be necessary
after the data has been cleaned and prepped. It is easy to pass data
from SPSS to other packages and to limited-purpose programs at the
frontiers of stat. Cleaning, prepping, and exploring the data is
usually 80 to 95% of the analyst's time, and SPSS really stands out in
these things. If analysts' time costs money this becomes very
important.
The GUI shortens the time to write syntax (process commands and
documentation). The syntax facilitates the developmental process of
revising the analysis until it does what is needed. The syntax
facilitates redoing portions of the work as the analyst, supervisors,
and quality assurance reviewers develop their views and understanding of
the data. The syntax helps with getting help by showing what was done
It facilitates sharing the data and metadata as is required by the
ethical standards of several disciplines and regulations for grants in
the US.
Hope this helps.
Art
Art@DrKendall.org
Social Research Consultants
University Park, MD USA Inside the Washington, DC beltway.
(301) 864-5570
Michael Reed wrote:
> I'm a consultant who works mostly with basic descriptive statistics.
I've
> used MS-Excel for the past few years. Questions: (1) is purchasing
> SPSS 14.0
> worthy my while? (2) is there any way to legally obtain SPSS for less
> than
> the $1,500 going price? Thanks, Michael
>
>
