new_n_lost wrote:
A. Confounding error << What about this one. Selection bias is mostly if people selected are different from the people who have the disease. here i think both the populations placebo and experimental have the disease. Am i missing something ??
Correct answer is A.  Confounding error.
Explanation provided: A confounding error is committed when a variable other than the one being studied is influencing the results. In this study, the treatment group's lower blood pressure may be secondary to their younger mean age rather than to the antihypertensive mendication. / Answer D is incorrect because selection bias results in an increase in the likelihood of one outcome occuring.
I'm baffled.  Like some of you I also chose D and the explanation hasn't really convinced me. Can anyone explain how to rule out selection bias. Becos based on the explanation it still seems plausible - age of the people "increased the likelihood of one outcome". What am I missing?
|
| new_n_lost
Politically InCorrect
Topics: 643
Posts: 5,888
|
According to my thinking or what i have understood is that Selection bias always results in the manipulation of the data and therefore the result. Now if they had taken hypertensive ppl and given them the placebo and the drug to the non hypertensive people they would be doing selection bias.
And confunding error would be to assume that drug caused the major drop in the blood pressure rather accounting the fact that the difference in age will relatively miscontrue the result cos the age attributes to the hardening of arteries which is relative to the result.
The ideal study would have been people with teh same age and disease risk factors split up and then the results were collected between the placebo and drug group.
Moi 2 cents lets expand on this one.
___________________
FORUM RULES-- Those who believe in telekinesis, raise my hand. I get enough exercise just by pushing my luck --P4U World.." The pure and simple truth is rarely pure and never simple."
|
| Tiff
Forum Guru
Topics: 46
Posts: 450
|
Thanks nnl. I couldn't really see it, but your example helps it make sense to me. 
|
| drgho
Forum Junior
Topics: 1
Posts: 114
|
May be to do with deliberately introducing the error or not.
For example, one researcher conducts the research deliberately selecting the people with younger age for treatment group and older age for control group --> will lead to selection bias.
But, another researcher just did random assignment in a double blind study and later found out the antihypertensive group has a younger age group --> probably confounding.
but this question, not really mention about study design....so..quite tricky.
in order to exclude selection bias they should mention about how they select the sample and assign the groups.
|
| macintosh
Forum Elite
Topics: 38
Posts: 178
|
a selection bias occurs when the researchers are biased in their selection protocal. An example would be to gie the drug to seeingly more sick patients.
A confounding error is not a bias, but an error that arises in the data when all other factors are matched.
|
|
| |
| | | | | | 
