We run multiple regression. Results are quite encouraging. Output shows that 3 out of 5 independent variables have significant effect on the dependent variable. There is no multicollinearity as well as auto-correlation problems. Unfortunately the F statistic is not significant at 5% significance level. Doesn't it mean that the regression model is not fit? What can we do with it?
Dear Aminul,
If the F-statistic is n.s. it means that your model does not work. You have to delete or add some of your independents and then check the F statistics.
Dear Aminul,
If the F-statistic is n.s. it means that your model does not work. You have to delete or add some of your independents and then check the F statistics.
When the model does not work you have to change it and find a solution that works
Thank you very much Beatrice and Raid Amin for your feedback. It was indeed very helpful.
I also obtained a similar result, when I related weed density to soil type and crop management factors. Thanks for this issue. It has helped me a lot.
You need to add additional independent variable and check for the significance
There is No problem if some of the indepent variables is not significant. This is the results and you need to compare them with the results of other researcher. Additionally, you can use stepwise regression analysis to obtain the final model with the significant variables.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA