2
$\begingroup$

I intend to use the following data of patients with liver/wo disease in my research: https://archive.ics.uci.edu/ml/datasets/ILPD+(Indian+Liver+Patient+Dataset)#

But there wasn't any information on if the patients were a random sample. I can only guess that the situation was of people going to the clinic and then finding out that they had liver disease or not.

So I propose to obtain a random sample from the data of patients by:

1.Giving each patient a number from 1 to 583 (that is the total number of patients in the data set)

  1. Based on some random number generator I choose only 250 patients (this is just an arbitrary number haven't learnt power analysis yet). I intend to use a two way anova with response as Alamine Aminotransferase with gender and liver disease or not as categories

Will I then be able to use this new data set in my research and say that the sample is random? Cause I can't really do it if the data is not of a random sample

$\endgroup$

1 Answer 1

1
$\begingroup$

If the original sample is biased then drawing a random sample from it will not help I am afraid. The bias will still be there and you will have lost precision. But consider why you wanted a random sample. What is your target population, the general public, people with a particular symptom, people who worry about their liver, people with alcohol use disorder, people who visit A&E/ER, people who visit hepatology clinics? Perhaps your current sample could be considered representative of one of those populations. When you publish your research readers will make up their own minds whether it applies to their clinical situation.

$\endgroup$

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.