Is there a statistics expert in the MN house tonight?

[Blandmum]

If so I have a question.

If a student has carried out an ecolocial study of two different sites, with two different subsites should she have used a students t test.

I thought that for 2+ variables you should use an ANOVA

Help?

Not a statisrician, am an epidemiologist on maternity leave but as I recall my master's stats yes you anova because the t-test performed repeatedly would increase your chance of a false test result. Erm, becauae the t test probability is usually set to 95% abd so to do it on 4 sets of data would need to compare. But onnly anova if the 4 subsites have similar variance.

sorry got distraatced, should say in there that the repeating testsincrement your chances of hitting that 5% error. and i hope i understood enough right from the desc you gave - and am assuming the data is normaly distributed.

god now i remember why stats gave me such a headache.

[opens another beer]