Explanation of post-hoc tests for 1-way ANOVA

[MCMILLEN, Pete (WELLHO)]

Hi all

I've tried but so far failed to pin down a comprehensive explanation of appropriate post-hoc tests to call for 1-way ANOVA. (Apologies for implied laziness - after almost 5 years off on a professional tangent I've lost touch with old stats/SPSS networks and references).

Dispensing for the moment my violation of the underlying requirement for interval-scaled dependent variable (I have ordinal 11-point agreement-scale data - means for a range of prison "quality of life" themes that aggregate common items, which I have assumed sufficiently disaggregated to reasonably justify assumed interval-scaled), my data satisfies the following:

• Somewhat small total (pilot survey) sample: n=229
• Unequal sub-sample sizes for independent (categorical) factor (=12 prison units), ranging from 11 to 45
• Homogeneity of variance across levels of independent
• Null hypothesis of equality of categorical variable means rejected

Would appreciate insight to which of the many post-hoc tests are appropriate/best/standard and why, conversely which ones are not recommended and why?

Given 12 categories for the independent I expect use of contrasts as anything but a later step ill-advised? However, graphical illustration of category means across the various themes reveals a reasonably consistent pattern of "best" and "worst" prison units so I guess I could focus attention on a smaller sub-set at the extremes and test via contrasts?

Thanks in advance of assistance.

Pete McMillen
Senior Research Adviser
Strategy, Policy & Planning
Department of Corrections | Ara Poutama Aotearoa
Mayfair House | 44-52 The Terrace | Private Box 1206 | Wellington
DDI [phone] | Ext. 68052

The information in this message is the property of the
New Zealand Department of Corrections. It is intended
only for the person or entity to which it is addressed
and may contain confidential or privileged material.
Any review, storage, copying, editing, summarising,
transmission, retransmission, dissemination or other
use of, by any means, in whole or part, or taking any
action in reliance upon, this information by persons
or entities other than intended recipient are prohibited.
If you received this in error, please contact the sender
and delete the material from all computers.