I have been told that one-way ANOVA is used only in experimental studies. For some reason, I find it useful in comparing errors of two groups in terms of numbers. Can you help clarify this?
It doesn't need to be an experimental study. For two groups you can use a one-way ANOVA, but usually a t-test is applied. One-way ANOVA is mostly used for three or more groups.
Anova may or may not be appropriate for a situation where you are counting errors, if that is what you are doing. It's best to understand the assumptions and appropriate situations for anova or general linear models. It may be the case that a different statistical test or approach may be more appropriate for your situation.
While you describe some of the conditions for using ANOVA to analyze your data have been met, it appears that others have not been. First, with two groups, ANOVA and an independent groups t-test are mathematically identical and will yield exactly the same result. Count data is quantitative and fits the ANOVA/t-test model. However, in order to interpret the significance of the result correctly, both procedures require that the subjects be randomly placed in the two groups.If the assignment is not random, any significant results may be due to pre-existing differences between the subjects in the groups. Good luck and I hope this is helpful.