Hi, I hope you can help me...
I'm attempting to figure out how best to analyze my experiment...
My sample design:
I have two rivers (river 1 and 2). On both rivers, 5 macroinvertebrate samples are randomly taken from a 10x2m station monthly (for 12 months). Each sample is then counted for total macroinvertbrate abundance. After month number 4, a flood occurs on river 1. No such flood occurs on river 2.
My research question:
Did the flood on river 1 have an impact on macroinvertebrate abundance?
Note: initial analysis has demonstrated that (log n+1) macroinvertebrate abundance data for each station, each month, are approximately normally distributed and have equal variances.
My initial thoughts:
I have a BACI type design.
So far I've looked into a three way-repeated measures ANOVA model... (Factors: 'River' (i.e. river 1 (impacted), river 2 (control), 'month' and 'period' (before/after the flood)).
My questions to you:
1. Is this the correct path to be going down?
2. If so, should 'month' be nested within 'period'?
3. To account for samples not being independent, should 'river' be treated as the repeated measures factor (i.e. a random effect)?
4. Is this design balanced, or not? (I think it is, but I'd like to get another opinion).
5. How would things change if my data were not deemed to be normally distributed?
Any suggestions and/or guidance will be very useful,
Please ask if you have any questions or require further detail,
Many thanks in advance,