This is how i interpreted it:
Results of a binary logistic regression analysis to assess the effect of demographic factors such as farmers education, farmland size, location of farmland in the catchment, and land use type on the likelihood that a farmer would adopt an SWC measure (coded as Yes) or not adopt an SWC measure (coded as No) on his/her farm are presented in the Table below. The full model containing all predictors was statistically significant, x2(df = 4, n = 73, Yes = 56, No = 17) = 9.723, p = 0.045, indicating that the model was able to distinguish between farmers who are likely to adopt or not adopt SWC measures on their farm. Overall success rate of the model was 66%. Based on the odd ratios, farmers with formal education were 13 times more likely than those with no formal education to implement SWC measures on their farm.
Hello,
You are in general ok, for details check the material above.
https://statistics.laerd.com/spss-tutorials/binomial-logistic-regression-using-spss-statistics.php
Good luck!
Here is another web-page that explains the output from the SPSS LOGISTIC REGRESSION command. (Your png file name included spss, so I assume that's the software you were using.)
- https://stats.idre.ucla.edu/spss/output/logistic-regression/
But there is another issue you should be thinking about. You said that you "Yes = 56, No = 17". One rule of thumb based on simulations is that in order to avoid over-fitting, one should have at least 15 events-per-variable (EPV). And "event" in this case would be a NO response (because there are fewer NOs). So by that rule of thumb, you should have no more than 1 explanatory variable in your model. See the nice article by Babyak (2004) for more info. HTH.
- https://people.duke.edu/~mababyak/papers/babyakregression.pdf
Hmmm, not really. For example, " Based on the odd ratios, farmers with formal education were 13 times more likely than those with no formal education to implement SWC measures on their farm. " is not right. This value is for education after conditioning on the other variables in the model. It is important to stress this. That 13 can go up or down depending on what else is included in the model. If you want to make a statement about the association between these two unconditionally, just include education in the model. If you want to talk about the causal impacts of education there are other issues.
While interpreting a binary logistics regression output please ensure you talk about the odds ratios for each variable you have considered in your model just like how Daniel Wright stipulated and not forgetting to bring out the R-square value and also discussing the p-value for the variable, and what method did you use to determine the variables to include and exclude while coming up with this model? Hope you did not exclude very important variables in your analysis.
Please perform a model check/test by following this tutorial (Hosmer & Lemeshow test.)
https://www.youtube.com/watch?v=iZoaXETWAN4
The best I can think of is to paraphrase Kelvyn Jones . The best way to report your research is to clearly answer your research questions. Best, David Booth
- Badges
- Science topic
- Similar topics
- Correction