We are analyzing genotypes to compare their association with a given disease (Risk and other disease phenotypes). In some cases where a particular genotype is present in the cases but totally absent in controls, when we are trying to do multivariate analysis with SPSS version 19,
A) We are not getting mean and median
B) We are not getting hazards ratio though p-value is given for that particular comparison.
It is because that particular genotype is absent in controls, the statistical software is not displaying hazards ration, mean, median. Our statistician opines that since the data is missing we are not getting those values. Kindly suggest if we can give a value as 0.5 where ever the data is missing just for the sake of calculation.
When we do Fisher exact test using Graph pad, where ever the value is 0 the Graph pad program assigns a value of 0.5 for the sake of calculation. Kindly suggest us a way to overcome this problem for getting Kaplan Meier curves and Multivariate analysis using SPSS V19 software.
[Email removed by admin]
I understand you are dealing with absolutely new type of data which is Censored Data. I think also your censored data should be classified into Right Censored data.
In censored data you cannot calculate mean and median directly, but you need to estimate the model of data then calculate the expected value E(X), V(X) and so on.
You can simply estimate Kaplan Meier Model (nonparametric model), or Cox Model.
Or you can also estimate a more advanced and complicated parametric model (exp, log-normal, log-logistics...) through MLE for censored data then test the goodness of fit test of the estimated model.
In SPSS you can work with the menu: Analyze > Survival to deal with right censored data.
However, you need to read a lot about censored data and survival analysis before you start working with SPSS.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS