tesing no difference

[vilas]

I am wondering if we can test no difference between groups. For example, can we statistically test this hypothesis: there will be no difference between group 1 and group 2 in terms of Y (dependent variable)? This sounds to me testing "null hypothesis" which is not possible statistically. Any comments will be greatly appreciated. Thanks in advance!

[Maguin, Eugene]

I think you are interested in testing whether the two means do not differ by
more than a certain amount--a test of equivalence. This problem comes up,
for instance, in testing for drug equivalence, i.e., is the effects of drug
1 equivalent to the effects of drug 2. I'd expect a big literature on this
in medicine/pharmacy. I remember looking at an article, the citation to
which I do not recall, in Psych Bulletin. As I recall, the idea is to show
whether the observed difference in means is less than an upper point and
less than a lower point by means of two one-tailed t-tests, each centered on
either the upper point or the lower point. That is, one-tailed confidence
interval for the upper point does not include the observed difference in
means AND that the one-tailed confidence interval for the lower point does
not include the observed difference in means.

Gene Maguin

>>I am wondering if we can test no difference between groups. For example,
can we statistically test this hypothesis: there will be no difference
between group 1 and group 2 in terms of Y (dependent variable)? This sounds
to me testing "null hypothesis" which is not possible statistically. Any
comments will be greatly appreciated. Thanks in advance!

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[la volta statistics]

Hi Vilas

 

Yes, it is possible. The tests are called equivalence or non-inferiority tests and are widely used in the pharmaceutical fields.

You can find a good description of such tests for instants by IRVING KH & TOSHIHIKO M; DESIGN ISSUES IN NONINFERIORITY/EQUIVALENCE TRIALS Drug Information Journal, Vol. 33, pp. 1205–1218, 1999

 

http://www.diahome.org/DIAHome/resources/content.aspx?type=eopdf&file=%2fproductfiles%2f8357%2fdiaj_12277%2Epdf

 

 

Hope this helps,

Christian

 

**********************************
la volta statistics

Christian Schmidhauser, Dr.phil.II
Weinbergstrasse 108
CH-8006 Zürich
Tel: [phone]
Fax: [phone]
email: [hidden email] 
Web: www.lavolta.ch

---

Von: SPSSX(r) Discussion [mailto:[hidden email]] Im Auftrag von vilas
Gesendet: Montag, 14. Dezember 2009 13:22
An: [hidden email]
Betreff: tesing no difference

 

I am wondering if we can test no difference between groups. For example, can we statistically test this hypothesis: there will be no difference between group 1 and group 2 in terms of Y (dependent variable)? This sounds to me testing "null hypothesis" which is not possible statistically. Any comments will be greatly appreciated. Thanks in advance!

---

View this message in context: tesing no difference
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

[Steve Simon, P.Mean Consulting]

vilas wrote:
> I am wondering if we can test no difference between groups. For example,
> can we statistically test this hypothesis: there will be no difference
> between group 1 and group 2 in terms of Y (dependent variable)? This
> sounds to me testing "null hypothesis" which is not possible
> statistically. Any comments will be greatly appreciated. Thanks in advance!

There are several approaches, and the two one-sided t-tests proposed by
Gene Maguin represents one approach. A simpler approach is to define a
range of clinical indifference and accept the hyopthesis of no
difference between the two groups whenever the 95% confidence interval
for the difference in means was entirely inside the range of clinical
indifference.
--
Steve Simon, Standard Disclaimer
Two free webinars coming soon!
"What do all these numbers mean? Odds ratios,
relative risks, and number needed to treat"
Thursday, December 17, 2009, 11am-noon, CST.
"The first three steps in a descriptive
data analysis, with examples in PASW/SPSS"
Thursday, January 21, 2010, 11am-noon, CST.
Details at www.pmean.com/webinars

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD