Dear Sabhrina, the question seems too generic to me. You area dealing with counts, so in general I would say the chi square test is the one. If you want to compare the abundance of female and male bats of the same species in two different conditions, by using the result of a single count, go with the chi square test followed by a cohen's measurement of agreement or something like that.

If you have the result of many counts (I mean repetitions) it is a pity to loose such a variability in the analysis so I would suggest some anaysis taking care of the variance, such as the ANOVA or the pearson 's correlation. They answer slightly different questions: the ANOVA tells you if the number of bats counted per category (age/sex, I guess you are dealing with age in years, or categories such as juvenile/adults) are different in average between two or more conditions (remember to use a post hoc test if you have more than two categories), the correlation tells you how much the number of bats in the different categories varies according to different conditions (comparing i.e. counts of females vs counts of males), and it is valid if you are dealing with exactly paired data (but counts performed in the same place at the same time respect this condition) in at least 5-6 (better 8 or more) different counting stations.

One more consideration, I take from my experience with counts of small European bat species. Bat counts do not vary in a Gaussian way, as most of them are social species, so that the variance is too highly affected by the finding of very numerous aggregations. If you are in such a situation, then I suggest to use non parametric techniques, such as Spearman's rank correlation instead of Pearson's correlation and Kruskal-Wallis' test instead of ANOVA.

Hope this is helpful and you will like the results.

Dear Sabhrina

Like all tests, it depends on the assumption that we want to test.

The chi-square test used to compare several independent groups described by a categorical variable: it is the qualitative equivalent of ANOVA said Pierre Legendre

We can also measure the link between two variables: equivalent

qualitative correlation. Finally we can estimate the discrepancy between an observed distribution and a theoretical distribution.

In the chi-square test, the class sizes are used for the analysis of variance (ANOVA so) we have continuous numeric values. The chi-square is used to investigate whether the distribution of classes and is compatible with a distribution model (often equal distribution, but not always), while ANOVA is used to investigate whether differences in means between samples are significant or not.

ANOVA and CHi^2 tests assume that you have distinct groups. Let's take age as an example. Suppose that you choose 90 day intervals for age. A chi^2 test says that a subject that is 90 days is completely different from a subject that is 91 days old. Where as a subject that is 5 days old is exactly the same as a subject that is 90 days old. I would use a regression analysis. I would keep age continuous and code gender as Male = 0, Female =1 and code reproductive status as normal = 0 and none = 1. 

You start with the measures of your variables. For the pair ratio/interval-ratio/interval, Pearson r, for the pair ordinal-ordinal, Spearman rs, for the pair ordinal-ratio/interval, Spearman rs, for the pair nominal-nominal- phi, V, c, or other chi-square based measures of association.

Hello Sabhrina, I think what Alain and Andrew just wrote supports the fact that, the Chi square test is preferable. There is not much to say based on your question. You can go ahead to use the Chi square test to verify whether or not the interact between the two variables is significant. For a thorough analysis, you can go ahead and code the variables as Andrew did mention in his suggestions above and then do a regression analysis using the SAS software.

Dear Sabhrina, the question seems too generic to me. You area dealing with counts, so in general I would say the chi square test is the one. If you want to compare the abundance of female and male bats of the same species in two different conditions, by using the result of a single count, go with the chi square test followed by a cohen's measurement of agreement or something like that.

If you have the result of many counts (I mean repetitions) it is a pity to loose such a variability in the analysis so I would suggest some anaysis taking care of the variance, such as the ANOVA or the pearson 's correlation. They answer slightly different questions: the ANOVA tells you if the number of bats counted per category (age/sex, I guess you are dealing with age in years, or categories such as juvenile/adults) are different in average between two or more conditions (remember to use a post hoc test if you have more than two categories), the correlation tells you how much the number of bats in the different categories varies according to different conditions (comparing i.e. counts of females vs counts of males), and it is valid if you are dealing with exactly paired data (but counts performed in the same place at the same time respect this condition) in at least 5-6 (better 8 or more) different counting stations.

One more consideration, I take from my experience with counts of small European bat species. Bat counts do not vary in a Gaussian way, as most of them are social species, so that the variance is too highly affected by the finding of very numerous aggregations. If you are in such a situation, then I suggest to use non parametric techniques, such as Spearman's rank correlation instead of Pearson's correlation and Kruskal-Wallis' test instead of ANOVA.

Hope this is helpful and you will like the results.

Please check and try with the attached document (very useful)

Best,

G.Lemperiere

Dear Sabhrina,

Your analysis depends on the type of the variable of bat assemblage, if it a nominal variable (or ordinal) you can use Chi-Square test to test the relation between each variable (Sex, Age) with bat assemblage. if bat assemblage is a continuous variable and approximately has normal distribution or you have a large number of bat, two way analysis of variance is appropriate to test the the effect of sex, age into a bat assemblage.

Since, your variables are categorical data,  I would prefer you go for chi-square tests

the chi square test followed by a cohen's measurement of agreement or something like that