As Mirko has noted, your percentage variable could be treated as a continuous explanatory variable. But given that there are only 3 possible values, I would be inclined to treat it as categorical. And if all explanatory variables, I would use a factorial ANOVA model. It is equivalent to a regression model on the same data, but the output from ANOVA may be a bit more convenient for what you are trying to do. Are any interactions expected a priori? What software do you use? HTH.

You can dummy code the sex and use a simple regression/ANOVA model

Thank you everyone. Now I got an idea how to deal with my data.

Bruce Weaver I will analyze in R.

I am not a frequent useR, but Sal Mangiafico is. You may find his resources at the link below helpful.

• http://rcompanion.org/handbook/G_09.html

I Think it depends on your research question

If you want to find any differences among categories

you should do ANOVA Test

If you want to find out what independent factors may cause one independent variable

you should chose regression model

It’s my Limitted Knowledge

What is your sample size? How many individual insects do you intend to examine? As others have mentioned ANOVA would be the best choice, but you will need many more observations in your 3 x 3 x 2 factorial design. The minimun to obtain cell means is n=36, but that is only n=2 in each cell.

If you have all possible combinations of the factors represented in your data set, then ANOVA is the way to go. You will only be able to include interaction terms in the model if your treatment combinations are replicated. If you have an incomplete factorial, you will need to use contrasts to compare groups so you are not confounding any effects.