odds ratio reveals strength of association and also does the correlation... cross sectional study with categorical outcome measure and normal socio-demographic independent variables....... looking to figure out the association...........did literature review ... most of them has used both .. but none have given the justification for performing both tests... which one to use and why kindly place your views
Odds ratio is very reliable and informative for a case-control study . Good luck !
but mine is not a case control.......its just a single group population ...........
A good reference on analysis of cross sectional data is:
Fienberg. The analysis of cross classified categorical data. MIT Press 1977.
Using logistic methods enables you to assess strength of individual effects,after adjustment for other variables and is preferable to correlation , Chi sq, or factor analysis methods. For an example of such an analysis see attached paper.
Odds ratio and correlation don't measure the same thing- correlation looks at how much one variable is explained by another (for example, is how much is wieght explaine by height?). Odds ratio compare the odds of a specific outcome in two groups, one exposied an done unexposed to a specific exposure. Here it sounds like you have a categorical outcome, and several explanatory variables, which make sme think you can use odds ratio to measure your association. You would have to think carefully about the type of model you would use, depedning on yoru type of data- if it is cross sectional data, an dyou have several variablesthen binomial regression might be a good choice (although i have not seen the data so don;t quote me on that!)
I agree with Michael.
What type of data are you looking at? Correlation works with continuous data/measurement e.g weight, height, exact age etc. as mentioned by Michael, correlation looks at how much one of the variables is explained by the other variable. Like: How much is weight explained by someones height? It doesn't work well with categorical data like gender, social class, etc.
The question is explained excellently by Mr. Michael and Henry. Different kinds of studies depending upon the nature of variables can be performed exploring which hypotheses. Odds ratio is measures in case-control studies while correlation coefficient is measured in correlational studies.
I also agree with Michael and Henry. They have nicely explained OR and correlation. Which one to use depend upon type of study and your hypothesis. Many published articles used both which is not required. More detail about your study is required to say which one should be used in your case.
in 1984, a large monography was written by Breslow and Day for IARC. they describe in detail what a case-control really is. Everybody has a weird cartoonish version of what a case-control is, but in reality a case-control is an effective sample of a large cohort. If your controls are taken at random from the population where cases arise at the same time you identify the cases, the odds ratios become approximations of the risk rate ratio.
Please watch the Youtube video attached from the UC Berkeley
https://youtu.be/u02hDH4WqLk
odds Ratio is most likely to give you what you are looking for in a more precise manner..compared to a correlation coefficient. A correllation will tell you that there is a significant association between variable X and variable Y..but an odds ration goes further to tell you how variable X and Y is related. it quantifies the relationship.
I guess that odds ratio in latent class analysis the strength of association if significant. It is based on a z score of 1.96. If the odds ratio is larger than 1.96 and the p value is smaller that .05, the estimate indicates that the classes have the same effect in terms of strength.
- Badges
- Science topic
- Similar topics
- Psychology