[DrSeddik]

What's the difference between confoundment, sampling and selection bias?

I know that failure of follow up is a selection bias. Sampling bias is "selecting" candidates for a study not representative of the population, and confounders are other factors in the selected population that come in association with the studied variant.

For example selecting mostly young population for an atherosclerosis related study is not representative of the population, and it confounds with other factors as smoking.

 

[khanar]

You answered your question yourself! You explained them all

 

[DrSeddik]

so if you do a research on atherosclerosis and you get candidates with young ages, what kind of bias would that be?

 

[khanar]

so if u choose do a research on atherosclerosis and you get candidates with youg ages, what kind of bias would that be?

Definitely selection bias

 

[DrSeddik]

so why not sampling or confoundment

 

[khanar]

This is my understanding of it;

You select young people to study atherosclerosis in the population. (Selection bias).
You withdraw blood from them (to measure vasodilators levels) right after exercise. (Sampling bias).
You did not stratify them by smoking status. (Confoundment bias).

 

[DrSeddik]

I get the difference between sampling and selection, but don't you consider the age as a confounder?

 

[khanar]

I get the difference between sampling and selection, but don't you consider the age as a confounder?

No, in our example Age is not a confounding factor because all the population of your study are young. So it does not confound with your analysis. But if you forget to check their smoking, obesity, family history, etc status, that's confounding because maybe some of them are smoking, obese, ...etc
Confounding factors arise when a portion (not all) of your population likely to have it.