1
$\begingroup$

I have the following data of 2 diseases from 5 areas. I want to see if there is any relationship between the 2 diseases. Incidence Rates of 2 diseases (cases per million per year)

    Areas   Disease 1       Disease 2
    1       4.653           0.751
    2       6.910           1.121
    3       4.957           0.745
    4       2.870           0.848
    5       2.819           1.166

Actual number of cases are
Areas   Disease 1       Disease 2
1       1152            186
2       2601            422
3       1051            158
4        403            119
5        290            120

I am a beginner in Biostatistics. Kindly advise me in simple terms

Improve this question
$\endgroup$
2
  • $\begingroup$ What are those numbers? Are these average values, what is your sample size, etc. A little bit of context would be of great help. $\endgroup$
    – chl
    Commented Mar 9, 2011 at 10:40
  • $\begingroup$ @chl: The first paragraph shows the Incidence Rates of 2 diseases (cases per million per year) being compared The second paragraph gives the actual number of cases of 2 diseases being compared $\endgroup$
    – DrWho
    Commented Mar 23, 2011 at 7:32

1 Answer 1

Reset to default
3
$\begingroup$

There is no simple answer for how to "establish a relationship between two variables;" indeed, your question is one of the central issues in statistics and research is still going on on how to do this. But some basics: first you will want to plot your data, and then you will want to carry out a linear regression to test some specific type of relationship between variables in your data. You will need to obtain the "p-score" of the regression to get an idea of how well your purported relationship is supported by the data. Generally if you can get a very low p-score (e.g. p < 0.01), then it will be safe to say that there is a relationship between variables.

Improve this answer
$\endgroup$
5
  • 4
    $\begingroup$ A lower p-value does not mean that the observed effect is stronger. $\endgroup$
    – chl
    Commented Mar 9, 2011 at 11:10
  • 1
    $\begingroup$ Indeed, it is important for a statistician to understand the difference between statistical and practical significance. I interpreted the OP's question as asking about statistical significance. $\endgroup$
    – charles.y.zheng
    Commented Mar 9, 2011 at 11:25
  • $\begingroup$ Thank you for the prompt reply and the clarification. I provided initially Incidence Rates. Now I added number of cases also for clarity $\endgroup$
    – DrWho
    Commented Mar 9, 2011 at 12:38
  • $\begingroup$ @chl♦: Thank you for the prompt clarification. I provided initially Incidence Rates. Now I added number of cases also for clarity. $\endgroup$
    – DrWho
    Commented Mar 9, 2011 at 12:39
  • 1
    $\begingroup$ The point made by chl is important: any p-value below your alpha level of statistical significance is in the safe area to reject the null hypothesis, and further distance of the alpha level will not indicate further substantive significance. This applies to the formal model behind your scatterplot and its fitted regression line, just like it will apply if you detect a non-linear relationship, either visually or through formal diagnostics, and then switch to more advanced techniques in order to model it. $\endgroup$
    – Fr.
    Commented Mar 13, 2011 at 15:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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