How is this biased?

[yangxu]

The National AIDS Bahavioral Surveys found that 170 individuals in its random sample of 2673 adult heterosexuals said they had multiple sexual partners in the past year, that's 6.36% of the sample... but my prof said this is biased, my guess is that the survey concerns a sample consists of only adults who are heterosexuals and based this random sample's result to cover all types of sexual preferences for adults in the community... is this a reasonable explanation? Please enlighten me, thanks a lot in advance!

[Bald Wonder]

In a study the two major classes of bias are:

 

1. Selection bias

2. Observation/information (misclassification) bias

 

If the purpose of this study was to determine causes for AIDS in a random population (including homosexuals, blood transfusion recipients, IV Drug abusers, etc.) then this is definitely a case of selection bias.

 

He could also be referring to recall bias, which is the most common type of observation bias in this type of survey.