Cross Validated is a question and answer site for people interested in statistics, machine learning, data analysis, data mining, and data visualization. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I'm a clinical researcher and I am comparing children who had fluid in the abdomen (ascites due to liver disease). I have 4 groups based on the kind of fluid analysis and in each group I have incidence of renal failure, bleed, altered sensorium, etc.

How should I compare these groups? Should I do multiple chi square tests taking 2 groups at a time? My number in each group is above 30 (but the number is different in each), and I am using SPSS 20.

share|improve this question
Related question:… – nico Feb 22 '13 at 17:32

As I understand it, you have 4 treatment groups - and for each child, you have several binary outcomes? You could build a score for each patient by adding the incidence of each potentially adverse outcome. Then compare the 4 groups based on the score.

a) if your score is the result of adding up a lot of binary outcomes (say 8 or more), then do a one-way ANOVA, assuming "normality" or your score.

b) If there are fewer scores (say, 4 items), then you could do a generalized linear model with binomial response and where "treatment group" is the explanatory variable.

If you go with Peter Flom's suggestion of doing a logistic regression on each outcome, be sure to adjust the p-value. Using 5% will boost the probability of a type 1 error somewhere in your results - so take something smaller. You want $1-(1-p)^r \leq 0.05$, where $r$ is the number of outcomes and $p$ will be the significance level for each test. For 4 outcomes, you want to go with a significance level of around 1% or 2%.

share|improve this answer

You should probably use logistic regression with each of your incidence variables as dependent variables and group as independent variable. You could do one logistic regression for each outcome.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.