On Mon, 27 Oct 2014 08:19:29 -0700, Jim Clark wrote:
Hi
Important to note that Muenchen's site examines all areas that
use statistics.
I agree. But one could argue that because not all psych majors
or psych grad students continue into academic psychology,
Muenchen's analyses are relevant Even Ph.D.s who leave
academia may want to know whether knowing SPSS is a still
a "marketable" skill. The greater the "penetration" of IBM SPSS
into the business community, especially with its "predictive
analytics" strategy, the stronger the argument for teaching SPSS
even for those who no longer stay in psychology.
SPSS still dominates, I believe, in Psychology and the Social
Sciences generally, at least judging by the software used by
recent graduates and stats texts published with psychology in mind.
As I mention below, SPSS leads in mentions in scholarly articles
(see Muenchen's Figure 2b) though it peaked years ago and is
now on the decline. Second place SAS shows a similar pattern.
One could do a similar search in PsycInfo and other databases
to see how often SPSS and other packages are used and this
would strengthen the argument for the continued use of SPSS
in psychology courses. However, Figure 2c (which eliminates
SPSS and SAS) clearly shows that R is rapidly growing in mentions
in the scholarly literature. This is the handwriting on the wall.
As for the cost of SPSS, what is the cost of many researchers
having to learn R or whatever might be available cheaper?
If it were only a decision that researchers and teachers made, then
the cost-benefit of continuing to use SPSS relative to R would be
clear-cut. However, last time I checked, most administrators weigh
hard costs like several hundred dollars per individual license more
highly than the cost of learning new software -- just remember what
happens when there is a changed from one learning management
system to another (e.g., from Blackboard to Sakia) or whenever
Microsoft "upgrades" windows or its software. I know that
R is not that easy to learn, especially if one knows other packages
very well (there are pronounced proactive interference effects).
But as graduate programs continue to adopt R and new faculty
with R skills come in, the senior faculty's complaints about how hard
R is to learn start to become less and less relevant.
Local factors (what is everyone else who is using stat software using
at your institution) are important factors. If the institution has to
cut
costs, it may just decide by fiat. Over on the SPSSX-l mailing list
one institution went from SPSS to Stata and one of the SPSS mavens
now had to develop equivalent skills in Stata. That's a task no one
looks forward to.
-Mike Palij
New York University
[email protected]
-----Original Message----
On Monday, October 27, 2014 10:00 AM, Mike Palij wrote:
On Mon, 27 Oct 2014 05:30:30 -0700, Larry Daily wrote:
Hello, fellow TIPSters!
Every year I have to write a technology funding request to maintain our
access to SPSS. This year there is a lot of concern about the
ever-rising cost of SPSS. One of our concerns is that we want to
prepare students for what they will face in grad school, thus my
question: for those of you who teach in grad programs, what statistical
software do you have available to your students? Do you consider it a
benefit if a student has had SPSS experience as an undergrad?
Any feedback on the issue would be much apprecitated.
I don't know of any recent surveys of statistical software used by
graduate
programs but Robert Muenchen does do a survey of statistical/data
analysis
software used in business and he compares a large number of software
packages
and uses different metrics to measure "popularity" (e.g., how often do
job
listing specify a specific package, how often is a specific package
identified
in a the journal literature, etc.). His review is available on the web
and can
be read here (up to date for 2014):
http://r4stats.com/articles/popularity/
On the measures Muenchen uses, SPSS does relatively well but other
statistical
packages (e.g., SAS) may do better but "R"
appears to be growing in popularity and even exceeds SPSS in some areas
(see
Figure 1a on number of jobs mentioning a specific package). See Figure
2b for
how SPSS has been doing over the past decades relative to other popular
software.
If nothing else, it's an interesting read.
---
You are currently subscribed to tips as: [email protected].
To unsubscribe click here:
http://fsulist.frostburg.edu/u?id=13090.68da6e6e5325aa33287ff385b70df5d5&n=T&l=tips&o=39666
or send a blank email to
leave-39666-13090.68da6e6e5325aa33287ff385b70df...@fsulist.frostburg.edu
