Biased sampling

Explanations > Theories > Biased sampling

Description | Research | Example | So What? | See also | References 

Description

When we need to make a decision and when we have very little real information about the subject in question, we will blithely base our decision on that very limited and hence biased sample.

For example, if we go to a restaurant for one meal and then someone asks us what the restaurant is like, we will give an authoritative pronouncement, based on that single experience.

Research

Hamill, Wilson and Nisbett (1980) gave people a hardship story about a woman on welfare. Some they told she was typical of people on welfare, others that she was atypical and to others (control group) they said nothing. Both those who were told she was typical and atypical rated her as less likely than the control group.

Example

An early political poll in the 1930s was to select random names from car owners and telephone directories. But only richer people had these, so the results were biased (and hugely incorrect). 

So what?

Using it

To help someone else decide in your favor, constrain and control the information available to them.

Defending

Beware of deciding on small samples, especially if they come from a single and possibly suspect source.

See also

Anchoring and Adjustment Heuristic

References

Hamill, Wilson and Nisbett (1980)

|awa|