Research Concepts

Bias

Researchers take great pains to minimize error. In any human endeavor, error is inevitable. The simple act of data collection, measurement and recording can introduce small random errors. Even the best instruments in the physical sciences have tolerances below which they cannot measure and the same is true for social science instruments.

Error is unavoidable when samples are involved. During this course you will learn methods for controlling and describing errors that occur randomly -- errors that do not favor one outcome over another but simply occur by chance.

Other kinds of error can be systematic, tending to favor one outcome more than another. Statisticians call this kind of error bias. A famous example of bias involves a poll in the 1936 presidential campaign. This poll spectacularly predicted the wrong outcome in the 1936 election due to a biased sampling strategy. Another well-known example of bias in sampling is the Hite Report.

A well-designed study will include strategies to minimize bias.