I am trying to find out if there is any significant association between basic education and treating water at household level between two groups of populations differing in their socio-economic status. The lower socio-economic group (LSG) clearly comprises of a higher number of individuals without basic education compared to the higher socio-economic group (HSG) (39.78% vs 20.15%) and water treatment at household level for LSG & HSG were 31.4% vs 78.03%. However when i tried to find the correlation coefficient it does not show any association
Treated water = 2, Did not treat water = 1
Had basic education = 2, Did not have basic education =1
What can be the reasons?
Any good literature on this issue will also be of great help.
Dear Sayani,
The following links may be helpful:
http://www.ncbi.nlm.nih.gov/pubmed/26010901
http://www.ncbi.nlm.nih.gov/pubmed/9093672
http://www.jidonline.org/article/S0022-202X(15)32126-6/abstract
http://onlinelibrary.wiley.com/doi/10.1111/gbb.12213/pdf
Rafik
What types are your variables? Are they dichotomous and are you trying to find Pearson correlation between them?
Yes they are dichotomous variables and am calculating correlation in excel (Pearson)
Thanks Rafik. I looked at the papers. They are all related to genetic studies but I will try to understand. If you find any simple paper discussing this issue then it will be even more helpful.
Calculating Pearson correlation coefficient is not a good idea for your dichotomous variables. If you want a correlation coefficient insistently then you could make use of tetrachoric correlation coefficients to measure association.
To look at difference of associations for those two socio-economic groups you can compare odds ratios (or risk ratios). To do this calculate odds ratio for LSG and HSG and compare them. By making this you control for socio-economic status. Or make a logistic regression say basic education being the dependent variable and others being independent and see what happens.
Dear Sayani,
The following links are related to medical situation (out of the genetic field) and are easy to understand:
http://www.hindawi.com/journals/jpr/2015/608604/
http://europepmc.org/abstract/med/7768372
http://ageing.oxfordjournals.org/content/28/4/367.full.pdf
http://aapgrandrounds.aappublications.org/content/30/3/28
Hoping this will be helpful,
Rafik
If i dont find any significant correlation but odds ratio shows some kind of association, how am i supposed to interpret that? I have not looked at tetrachoric correlation yet. Am talking about normal Pearson's correlation
- Badges
- Science topic