A 2% difference in scores between the most and least dogmatic users means your test complete hogwash.

Your refusal to provide any data beyond averages is very telling. My guess is that there's a very wide distribution of scores, little trend, and very little correlation.