Hi Ken
To your first question, you can't exactly test the moderating effect
on a relationship which is not there, so yes go ahead and discard
hypothesis 2. I had that exact problem with my MSc dissertation.
Secondly if your data is skewed you cannot use a parametric test (the
mean would be wrong, and parametric tests use the mean) so you would
need to transform your data before testing (Andy Field suggests
reverse scoring first, then trying any of the standard
transformations). If you can't get it to normal distribution you can
only use nonparametric correlations. Whether you should try this I'm
not sure - I know parametric tests are a lot more powerful, and
depending of the multicolinearity of the data you may be able to do
linear regression tests.
Hope this helps a bit.
--
Many thanks
Ben Haysom-Newport BSc (Hons), MBPsS, MSc
On 11 July 2011 16:26, Ken <[log in to unmask]> wrote:
> Hi,
>
> My thesis concerns functional diversity within teams, which is is defined as differences in knowledge and perspectives that members bring to a group/team through their functional backgrounds - for instance, you may have a number of engineers, economists, and lawyers all belonging to the same team, providing different inputs to achieve a teams' objectives.
>
> In broad brush strokes, the thesis focuses on the relationship between such functional diversity within teams and a team member's trust of a particular work group/team that they are a member of. One of the main aims of the thesis is to show that these relationships will be moderated by a team members beliefs in relation to functional diversity (For example, a respondents' beliefs that functional diversity will be of benefit to a team may consequently lead to high respondent levels of trust with regard to their particular team).
>
> The two hypotheses that I'm concerned about are :
>
> Hypothesis 1: There will be a direct correlation between the level of functional diversity (0 being lowest diversity, 1 being highest) within a respondents team and the level of trust that the respondent has for that team.
>
> Hypothesis 2: The relationship between team functional diversity and trust of the team will be moderated by respondents "Beliefs" about functional diversity (i.e. whether they see it as being a good or a bad thing)
>
> My main questions in relation to the above hypotheses are the following:
>
> I have found no significant relationship between functional diversity and trust. If there is no significant relationship between the two of these, can I still test hypothesis 2, or should I just discard hypothesis 2 based on the fact that there is no direct relationship between functional diversity and trust (hypothesis 1).
>
> Another concern I have is that the responses for "Functional Diversity beliefs" are negatively skewed. Practically all (142) of the responses that I received range from "undecided" (only a few of these) to "Strongly Agree" on a 7 point Likert scale, with only 3 responses on the "Disagree" side of the scale. I wonder will this affect my results badly, or will I need to "transform" the responses to this scale to make it more "normal" and also to check for direct correlations.
>
> Also, please note that all of the scales mentioned above are 7 point likert type scales.
>
> thanks,
> Ken
>
>
|