Statistical Analysis for Two Birds Counting Methods

[evelingtavera17]

I want to compared the relative abundances of two bird counting methods. I am using the proportions instead of the numbers itself, because i dont want to argue that one method have the same effort or chance to capture the same amount of information, just want to know if the structure and composition of the data are similar. I am going to use several methods to compare the composition and in the case of structure basically the diversities with grades of significance. The problem is that i also want to compare the structure of two samples and choose a good statistical test. Anova and other tests have some asummptions that i dont think im going to fulfill for all my samples. All that kind of tests compare means so i have to use proportions or densities but make a lot of data standarized (arbitrary). For example i have read some papers that compare point counts and mist nets using an analizing unit like species counts per 4 point counts or 100 mistnet/hours. In such cases i have lot of data for each species to compare the means but you have to much arbitrary or unconscious assumptions like "each point or mist net study area at any time have to represent what the second round of count is going to have". I will be very thankful if anyone can help me selecting a simple method to compare the abundances between two samples. For data like this

sp. Sample 1 Sample 2
1 65 30
2 0 1
3 20 3
4 5 0
5 12 13

Because sample 1 and 2 are data from different census techniques i am going to use proportions. Then i want to compare not the means but the proportions of each species so i can have a significance of the similarity or dissimilarity of the tendency of the proportions based on the comparisons of the values between the species. At last i can say if sample 1 have the same results or not from number 2. Starting with this i can use another methods just to corroborate this at a level of comparing different places with different census method ( more than 2 samples) with ordination graphs like deternded correspondence analysis, just to be more easy to visualize and understand the conclusions.

Best Regards,

Eveling T Fernandez
Lima - Peru

[stereologist]

Be careful using ratios. Ratios can often lead to problems. A problem you might consider here is known as Simpson's paradox. It is possible for one set of numbers to have higher ratios than a second set of numbers. But when the overall ratios is checked the opposite is true.

I am not saying ratios are wrong in your case, but the use of ratios has led to many misconceptions in the biological sciences.