Figure presents a dependence of spore distribution on rainfall, classified in four categories. All observations revealed not normal distribution and they were based on five years of measurements. This figure I would like to strengthen using an appropriate statistical test to show statistically significant differences between different amounts of rain that fell within 24 hour...Do you think that Friedman Anova & Kendall's concordance is a right choice? Or is there any other, more suitable test for such analysis? .If there were two categories, instead of four, should I use then a sign test?
If not mistakes, better to consider Repeated Measure ANOVA in order to be able to evaluate the difference between the values at different time intervals. For normality, you can normalize your data using Log scale.
Thank you Abdualrahman for your answer; I would prefer not to transform my data as from my previous observations transformation removed several percent of the data including important scores.
I think because your dependent variable is continuous, you can perform a repeated measure with hour as repeated measurement and rainfall as groups. As I saw in the graph, because of interactions between time and rainfall you can perform an orthogonal polynomial test too.
There seems to be a very strong circadian rhythm. I would model this, making the amplitude (and possibly the phase) dependent on the rain intensity.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis