Statalist The Stata Listserver

[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Intro social science stats book

From   "Richard Atkins" <>
To   <>
Subject   Re: st: Intro social science stats book
Date   Mon, 03 Jul 2006 13:02:24 +0100

>Why do so many books provide SPSS examples?

I used to teach research methods and stats to psychology students in
the UK and the use of SPSS was so widespread at both under- and
post-graduate levels that there was no real competition. In fact most of
my colleagues teaching research methods knew no other statistical
packages with the possible exception of microtab.

I think the reasons for this were largely historical. Before the
company rebadged itself and its product as SPSS (as a name rather than
an acronym) it was S.P.S.S. which stood for Statistical Package for the
Social Sciences. I suspect the name change reflected a broadening of the
target market but the program had previously been marketed specifically
at social scientists and has been used in social science courses for
years. In the UK, the use of SPSS is so widespread that all the
undergraduate students want it on their CV. Nearly all postgraduate
social science (or at least psychology and sociology) courses use SPSS
so if you don't teach SPSS at undergraduate level the students complain
they are being disadvantaged.

Apart from tradition, there is also entropy.  In many institutions the
staff who teach stats to psychology students are not statisticians but
psychologists for whom statistics is at best a peripheral interest. I
expect they would generally resist if not resent the idea of having to
learn a new stats package. My experience of discussing the merits of
other stats packages with colleagues has been that their eyes glaze over
as you rattle off the advantages of programs like Stata. Such benefits
are of no consequence to people who teach a minimalistic,
painting-by-numbers way of doing statistics and whose requirements of a
stats program is basically the same as the students: that you click a
few times with the mouse and then it "does the stats for you".

Richard J. Atkins

>>> 07/01/06 4:39 pm >>>

>Date: Fri, 30 Jun 2006 10:55:45 -0700 (PDT)
>From: M Hollis <>
>Subject: st: Intro social science stats book
>This fall I'll be teaching an intro stats course to
>undergraduates in sociology and other social sciences.
>  I'm beginning to search around for the elusive
>perfect textbook and I'm wondering if folks have
>suggestions.  I'm looking for a book that emphasizes
>intuition about statistics and their proper use rather
>than a mathematically-oriented one that focuses on
>deriving the formulas.

Having taught this class since 1983, I can report the disappointing 
observation that the available books seem to have become worse and 
worse over time---- or perhaps I have become more and more demanding 
:-}  I think the class of the market here is Agresti and Finlay's 
book, but it's probably a bit too hard for undergraduates in the 
social sciences at most schools. Two very popular books I would not 
recommend are Healey's Statistics: A Tool for Social Research (too 
much emphasis on formulae and computation, not very conceptual, poor 
homework problems), and Levin and Fox's Elementary Statistics (Hall 
of Shame quote from 10th ed., p. 232: "Put simply, P is the exact 
probability that the null hypothesis is true in light of the sample 
data....)   This fall, I'm trying Nachmias-Frankfort and 
Leon-Guerrero's Statistics for a Diverse Society which, despite the 
rather touch-feely title, appears to have some real virtues, e.g., a 
focus on understanding relationships as opposed to testing 
uninteresting hypotheses.  It postpones sampling distributions and 
hypo. testing to the end of the book, which I am convinced is the 
right way to do things.

>  In my ideal world, the book
>would provide Stata examples, but that seems to be a
>rare occurance. 

I think that Rich Williams' answer of *why* is right on target, but I 
wouldn't worry about this too much.  I think there are pedagogical 
reasons not to bother with statistical packages in the introductory 
class, at least not in the context of a 3-credit course. I have even 
seen at least one research article supporting this view, although I 
wouldn't say I'm totally convinced.


Mike Lacy
Fort Collins CO USA
(970) 491-6721 office

*   For searches and help try:
*   For searches and help try:

© Copyright 1996–2021 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index