Question about Methodology

#1
Hi everyone! I'm new here, and wanted some advice on a methodology I'm using for a study of mine. Due to some confidentiality requirements, I can't go into specifics, so I hope you'll bear with me as I explain this in a dessert-themed analogy.

Let's say I'm studying cake recipes, and I'm looking at several different several different ingredients in a recipe (variables include Eggs, Sugar, Salt, Baking Powder, etc, coded as 0 for absent and 1 for present). There is also a variable for the type of cake being made (cupcake, bundt cake, etc, each coded with a unique number). I also have Success as a variable, with 0 being that the cake did not bake correctly, and 1 being that it did bake correctly. I want to analyze which ingredients most significantly affected outcomes of success. So far I've been using one way ANOVAs for Success with each variable as a factor. Is this correct, or is there a better analysis I could be using? Note that for the ingredient variables, they are not all completely independent of each other (for example, Baking Powder is more effective with Salt as an ingredient).
 
#2
Hei

Since you're working with binary data, it would be more suitable to use logistic regression or discriminant function analysis. An ANOVA is used for continuous data.
Hope that helps a bit
 

hlsmith

Omega Contributor
#3
If variables are associated with each other (correlated), you may also want to examine Variance inflation Factors (VIF).