I recommend that you show the full model, which includes all variables in your regression model, and the best model, which includes only important variables. You can select a significance level to select those important variables (e.g., .05 level).
To be parsimonious, displaying these two models will be fine unless you have theoretical or methodological reasons to show some other regression models.
Also, show adjusted R square for each of these two models, instead of regular R square because regular R square value become larger with more variables included in a regression model. In other words, to compare the best model with the full model, the adjusted R square is better than the regular R square.
A plenty of articles should contain tables showing the backward regression model. But, you can take a look at mine. I have included only standardized regression coefficients. But, another way is to include unstandardized regression coefficients with standard errors. Also, I showed variation inflation factor (VIF) because collinearity was an issue in my regression models. If collinearity is not a concern in your model, you may not need to present VIF in your table(s).
See Table 4 on page 112in my article:
Don Soo Chon. (2010). Medical resources and national homicide rates: A cross-national assessment. International Journal of Comparative and Applied Criminal Justice, 34, 97-118.
- Badges
- Science topic
- Similar topics
- Mathematics