Comparing the means of two groups – the t‐test
| Published date | 01 December 2002 |
| DOI | https://doi.org/10.1108/07378830210452677 |
| Date | 01 December 2002 |
| Pages | 469-471 |
| Author | Joseph Janes |
| Subject Matter | Information & knowledge management,Library & information science |
Comparing the means of
two groups ± the
t
-test
Joseph Janes
Introduction
Gifts are lovely to receive, aren't they? It always
gives one that warm and rosy feeling inside,
knowing that someone cared enough to select
something they think you would want or need
and then went to the trouble to get it, wrap it
up, and present it to you.
Except, of course, when they are real dogs.
Those of us in the library world know only too
well that when a ``gift book'' comes in, it can be
a real boon and a wonderful addition to the
collection, or it can be a pain in the neck that
makes no sense to the institution and its
mission, probably just a tax write-off or a
cast-off more appropriate for the annual
friends' book sale than the stacks.
Stereotypes aside, this is an empirical
question. Is there a difference in usage between
materials received as gifts, as opposed to those
purchased by a library? This is an intriguing and
potentially important question, and one that
was addressed in an article a couple of years
ago. Aside from being an interesting question,
this is also a good demonstration of the use of
one of the most important and most frequently
used tests of statistical inference we have, the
t-test.
The t-test
First, a bit about the test. There are several
flavors of the t-test. The one illustrated here,
and the one we see most often is the t-test for
two independent groups, that is, two groups
that have different membership and are not
matched up in any way. (For example, if your
subjects were pairs of librarians representing
individual public libraries in your county, and
you put one of each pair in group A and the
other in group B, the groups are now related,
and you would use the t-test for two related
groups.) The test is measuring how likely it is
that the difference between the means of the
two groups is due to some real difference
between the groups and not due to random
chance. The larger the difference, the more
likely it is because of something real, but it is
not quite that simple.
The author
Joseph Janes is Assistant Professor at The Information
School of the University of Washington, Seattle, Washington,
and is Founding Director of the Internet Public Library.
E-mail: jwj@u.washington.edu
Keywords
Research methods, Libraries, Statistics
Abstract
This column continues a series on topics in research
methodology, statistics and data analysis techniques for the
library and information sciences. It discusses the
t
-test for
the differences in means between two independent groups.
Electronic access
The research register for this journal is available at
http://www.emeraldinsight.com/researchregisters
The current issue and full text archive of this journal is
available at
http://www.emeraldinsight.com/0737-8831.htm
On research
469
Library Hi Tech
Volume 20 .Number 4 .2002 .pp. 469±471
#MCB UP Limited .ISSN 0737-8831
DOI 10.1108/07378830210452677
To continue reading
Request your trial