Evaluation Cookbook

cookbook logo
Route map: Home=>Publications=>Evaluation Cookbook=>Information Page
Statistics Questions
 Main   Links page   References   Add comment 

>Info Pages

Questions to ask yourself before you invest too much effort in analysing your data.

1. What will a statistical analysis tell you that you couldn't establish in other ways?

2. Have you thought about your sample and how it was selected? Are the number of missing responses satisfactory? Why are they missing?

3. Does your data include outliers or unusual observations? Why? Will they have an undue influence on your conclusions?

4. Is the size of your sample sufficient to make a full analysis worthwhile?

5. Have you used graphical methods to explore your data and get a good feel for it?

6. Are there any confounding variables?

7. Do you know what type of data you have collected? (Nominal? Ordinal? Interval? Ratio?)

8. Have you considered using confidence intervals rather than/in addition to an hypothesis test?

9. Do you know what conditions are assumed for the test you are considering to be valid? Do your data meet those conditions?

10. Just how reliable are your data anyway?

Nora Mogey
Coordinator, LTDI,
Heriot-Watt University.

versions for printing are available

[ Main | Links | Add comment ]
[ Printable version ]
[ Contents | Information pages ]

© All rights reserved LTDI and content authors.
Last modified: 25 March 1999.