It will depend on  exactly what data you colected for each condition  and the nature of the data.  Assuming the measurements were taken concurrently for the three groups,  then if you have pre and post-treatment  data that is interval or better and is (or can be transformed to) something approaching a normallly ditribution then you  can probably conduct an ANOVA  to compare the change scores (difference between pre and post measures) between each condition. If your data is categorical then a Chi Square might be applicable. Hope that helps!

Statistics is not applied after the experiment. Statistics is applied before the experiment. Thinking about the scientific aim, formulating a precise question, finding an optimal experimental design, understanding the data that is measured: *this* is statistics.

What comes later is simple, as it follows from the thoughts invested in planning the experiment and the understanding of the nature of the data. Exploratory analysis might be an exception, but this is worth only when the data experimental design was good.

As you have not given much helpful details, the only sensible next step in a statistical analysis is to create a good (clear & instructive) plot of the data in relation to the experimental factors.

It is very possible that what you find as a result, after the application of some statistic based on the analysis of differences, is not valid, to generalize. Basically, and as Jochen once again mentions, since the design of statistical procedures are a priori. This allows to know the influence of the maximum of variables in function of diminishing latent ones. This allows you to increase the accuracy of your response.

Thanx all.....

I consulted a statician recently. He read my hypothesis and told me to apply anova in F distribution. I will share my result with all if you.

Thanx again.....

REGARDS