Do not be discouraged. We just need more information to be helpful to you.

Given the information that you have shared thus far, it is difficult to provide adequate substantive advice regarding your data analysis. Without knowing the specific question(s) you are are trying to answer, the specific design details such as sampling method, sample size(s) for the entire sample or subgroups, the independent and dependent variables involved, one can only assume what you are trying to accomplish.

If you cannot share details publicly, you may need to approach someone near you who has sufficient subject matter and statistical knowledge to assist you. Or, if possible, revise and expand your question here to include the details needed for others to assist you. I hope that you will find answers one way or another.

I think the main cause of reaching non-significant result is your selected sample size which is not probably appropriate. I suggest you to determine your sample size with Minitab, R or PASS softwares and then analyze the results based on power of your test.

Thank you for the replies. As a student, I'm required to work with the sample that I'm given, so I can't determine my own sample size, though I'm sure that would have made a difference.

John, thank you also for replying. It's a 2-way Independent Measures ANOVA. Ethnicity is the moderating IV, as a dichotomous variable. Mindfulness is the IV and Identity exploration is the DV. The groups are as follows:

Between-Subjects Factors

Caucasian or Non-Caucasian

Caucasian 1177

Non-Caucasian 734

Level of Mindfulness

Low 901

High 1010

According to previous research, high levels of mindfulness should have resulted in significantly high levels of identity exploration. I was trying to determine if ethnicity had a further influence on that. Does that provide enough information?

Thank you!! Leah

Leah,

Thanks for the additional information.

I take it from your description that your analysis has three effects: a main effect for Mindfulness, another main effect for Ethnicity, and one two-way interaction (Mindfulness x Ethnicity). Further, you expected to find a statistically significant difference (that is with a p value < .05 for the means on Identity Exploration favoring the high mindfulness group.

The main effect for Ethnicity simply tells you whether there is a significant mean difference in Identity Exploration between Caucasians and Non-Caucasians. The moderation effect, if statistically significant, indicates that the difference between means of Identity exploration for the two Mindfulness groups of one Ethnicity differ from the difference between means of Identity exploration for the two Mindfulness groups of the other ethnicity. So, in summary, your output results should have 4 effects listed: Mindfulness, Ethnicity, Mindfulness x Ethnicity, and an Error term.

Each of the first three effects (Mindfulness, Ethnicity, Mindfulness x Ethnicity) will have an associated F ratio and a p value. Significant p values will be .05 or smaller. In your case, if the interaction term is significant, you have a moderating effect. That is true whether the main effects (Mindfulness and Ethnicity) are statistically significant or not.

If the analyses you have completed so far do not include the two-way interaction, you will need to explicitly add the interaction in the model statements that you have written, or you may need to specify that the design is a fully factorial design. The choice depends upon the particular software you are using.

I hope this points you in the right direction.