I would like to ask whether Indicator Species Analysis can be used in a scenario where two classification groups are of differing sizes. I would like to identify indicator species in the central zone of a wetland (13 replicates) and indicator species in the littoral zone of the same wetland (52 replicates). When I run the ISA it will be Centre group vs. Littoral group but the group sizes are widely different. Can I use ISA in this case?
Yes it does. Bigger group size indicate the preferential micro-environment for that species, and hot spot for biodiversity...
I don't think the 13 and 52 are 'replicates'. If you sample the same area, then you have sequential samples or pseudo-replicates. In ecology, replicates are usually similar ecosystems that are distant from each other. But you can still obtain mean values of species within the different zones of the one wetland (replicate) and compare these with perhaps a t-test :)
Yes, ISA can be performed although the size is different. Of course, the analysis from the littoral zone due to larger group size would be more significant comapred to the central zone, if statistical analysis is done.
As Andrew replied above, it looks like you only have one site. So, your study is not replicated. The samples that you are referring to as replicate samples are simply samples used to estimate mean conditions with some level of uncertainty (standard deviation) of the site mean. Replication and pseudoreplication depends on the study question. For example, in some studies samples taken within a system (e.g. a lake) could be consider replicates, but samples from several lakes within a catchment could be your level of replication, or several catchments with a region could be your level of replication. Again, what is considered a replicate depends on the question being addressed.
Hi, Thank you for all your comments. I actually had three wetland sites but I would like to look at each wetland site individually. For each monthly sampling event (there were 13 in total), I sampled once from the centre and 4 times from the littoral zone. So the littoral samples are replicates but the centre are not. One approach that was suggested to me was to explore how sensitive my dataset was to sample size by taking a random subset of 13 replicates from the littoral zone and then doing an ISA on those 26 replicates (perhaps repeat this process a few times to include different subsets of the data). This seems like a suitable approach?
- Badges
- Science topic
- Similar topics
- Biological Science
- Kinesiology
- Running