Definite stats newbie, and I'm looking for a little bit of direction on which statistical test to use. Here's a quick summary: enter image description here

I'm trying to determine if the cases are on more than one narcotic at a time with a .05 significance. Is this an easy as doing an OR (1304*3404)/(6095*133)?

  • $\begingroup$ what's the difference between the case and the control? it looks like the control is also on narcotics. $\endgroup$ – markovchain Mar 19 '14 at 14:20
  • $\begingroup$ @markovchain The cases are people who have sustained a particular type of injury, and the cases have not sustained that injury $\endgroup$ – wootscootinboogie Mar 19 '14 at 14:25

Since you're looking for a statistical test, I'm assuming what you want to know is if there is a significant difference between the case group and the control group, and if there is a significant difference between those who use > 1 narcotics and those who use <= 1 narcotics. Am I right?

In that case, you want to use a 2-way ANOVA because it's easy and simple. You can do this test on software or some online site will probably do it for you for free. Here is a link to the Wiki for the test. Here's a link to a site that will do the test for you.

What you have to keep in mind though, is that ANOVA assumes a normal distribution. Plus, you can't perform 2-way ANOVA unless you have at least more than 1 replicate in one of your tests. The power of the test also decreases if you don't have the same number of replicates in all cases.

The way to evaluate the test is, if the p-value is less than 0.05, you say that there is a difference between the groups.

For your data, you can't perform 2-way ANOVA just yet because you lack a replicate (a repetition). I recommend you get another replicate first. Otherwise, you will have to do two one-way ANOVAs since that allows single replicate analysis.

  • $\begingroup$ Your statement in the your first paragraph is correct, I wish to see if there's a difference in the amount ( > 1 narcotic vs <= 1 narcotic for cases vs. controls). These data are historical (five years old or so) and consist of the entire population for my cases, and a randomly selected control group. I don't see how I could do a repetition of the data. $\endgroup$ – wootscootinboogie Mar 19 '14 at 14:42
  • $\begingroup$ So your numbers, 1304 133 6095 3404, are counts of the population inside your experiments? If you assume there hasn't been a significant change in your data over the years, you could segregate the population by year and you'll have 5 replicates per combination. $\endgroup$ – markovchain Mar 19 '14 at 14:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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