Hey
Happy New Year Greetings!
I have species presence-absence data of macro-invertebrates of an Himalayan stream from 7 different sites and two season. What is the best way to represent this data graphically (may be using R) and what are the statistical methods we can employ for the same?
Do you want to answer a specific question? At least, you can create a lot of different figures, possibly none of which would answer your question. 7 samples is not that many so I assume you question is exploratory?
Wim Kaijser Thank you for reply. We have found 34 genera in this stream. This is a 17 km long Himalayan stream and the terrain is very harsh and we were able to sample from only 7 sites. So we want to study how the composition of macro-invertebrates changing among these sites. also we have Physic-chemical parameters for these sites.
What first comes to my mind is an CCA, RDA or nMDS (https://sites.google.com/site/mb3gustame/). This can be performed in R using the vegan package (https://www.davidzeleny.net/anadat-r/doku.php/en:rda_cca). The results can be very insightful.
Nonethelles, one might look at it simpler in the form of a scatterplot and regression line if there is an upstream downstream gradient in the stream, i.e. distance from source (km) in relation to the fraction of gatherers/shredders, species number or abundance. With the application of a propper GLM.
I would advise against placing strong emphasizes on the p-value and more on the estimates (coefficients) and intervals resulting from the GLM keeping it descriptive, rather than conclusive. Your study is exploratory (hypothesis generating) not adressing a specific hypothesis, selecting the model beforehand. Moreover, spatial-autocorrelation is present as you sampling from a single stream, samples are biased to easy accessable locations and"non-random", cofounding variables are present and therefore you likely "over interprete" the autcome.
Hope this helps
- Badges
- Science topic
- Similar topics
- Counseling Psychology
- Confusion