@Behnam As your data and results from analysis suggest that there is no interaction effect of Trust on the relationship between participants' Willingness to Use Online Communities and their Actual Usage of Online Communities. An interaction variable (moderator) is supposed to either enhance or buffer the effect of IV on DV. But in your case, it doesn't affect the said relationship at all (insignificant moderation effect). This means that increase or decrease in participants' Trust doesn't affect (enhance/buffer) the relationship between participants' Willingness to Use Online Communities and their Actual Usage of Online Communities at all. This is what your data says and now rest depends on the literature from the domain whether this finding conforms/contradict earlier researches.

Good luck with your work!

It sounds like you used your exploratory sequential design to generate a hypothesis, but that hypothesis was not supported by a subsequent quantitative study.

If you think more generally about the issue of a failed hypothesis, the two possibilities are a false negative (i.e., the effect was there but you somehow missed it, due to operationalization, sampling, etc.) or a true negative (i.e., even if the effect was present it was too small to detect).

So, the first thing I would recommend is to go back to your qualitative data and reread the relevant passages. Among the questions I would ask at that point is how well your quantitative operationalization matched the way that the original participants' discussed this issue (i.e., a possibility for a false negative). Alternatively, you might ask how important this factor was compared to the other things they discussed (i.e., a possibility for a true negative due to a small effect size).

I hope that helps.

David L Morgan Thanks for your contribution to the discussion. It was really helpful.