hi i have to write my result section for my thesis but I keep on getting stuck because of formatting and issues. for analysis i have have conducted multivariate glm with sex,studyfaculty as between-subject factors and sq,eq,aq,spq as dependent variable. from the analysis i have n significant results ( no p value less than .05). how should i report in case of insignificant results? should i use tables or graphs or is it unnecessary? what else should i report on my result as none of the findings were significant nor the interaction effect. please help!
Good day.
It could help you:
https://www.statology.org/here-is-how-to-interpret-a-p-value-of-0-000/
Regards Marián :)
Dear Marian,
thank you for your response. however, i am not actually looking for answer regarding the p value. what i am looking for is reporting insignificant results in apa format because almost all papers have partial significance and i dont find any paper which shows apa reporting for glm analysis without any significant results. is there any advice you could provide me regarding this?
Hello
Multivariate analysis is often sequel to a series of tests aimed at corroborating the legitimacy of the dataset. Prior to data analysis, data screening needs to be performed for the purpose of investigating missingness, normality and outliers. Then, you can start data analysis, and you start by performing measurement scale analysis with the help of factor analysis. ( This is where you substantiate the reliabity and validity of your scale). Your P-values are there to testify for nomological validity. They estimate the causal relationships connecting the dependent variables to the independent ones. In this particular section in your study, you cannot afford to refer to values obtained from data screening ( outliers, normality...) neither can you restate the outputs from measurement scale analysis simply because those were necessary conditions for the legitimacy of the causal model in your study. Whether or not your hypotheses are supported is nother concern that is not methodological. As for the convenient way of diplaying the hypothesis testing in APA, I suggest to have look at this article
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3431747
Hi, Nabil morchid! first of all thank you for your answer. i already use the questionnaire which has been valdiated so i dont have to go through factor analysis. i had few outliers which i removed prior to data analysis process. what i need here is APA style that displays in my case specifically glm multivariate with single iv and multiple dvs. do you have any paper which does this? i already checked lot of papers but hardly find any.
You may be test one trail and two trail and see the difference according to hypothesis. Significant and not significant are not basic problem. You may explain discussion with supporting literature why this result found.
Sushanta Roy Chowdhary thank you for the suggestion. i already have two-tailed test for every analysis my problem is more related to the actual process of reporting which is of course according to APA format and not the p-value itself. but thank you for your help.
Hi N. Regmi
During the statistical analysis, when your P-value are significant, you can claim that the parameter is effective but if P-value are not significant, you can claim nothing. There could be two senario, one: parameter is not effective and two: parameter is effective but the precision of experiment was not enough to see that.
Good luck
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs