1. The d.c. component of an a.c. signal.
2. The d.c. voltage used to switch on or off a bipolar transistor or diode
(see forward bias, reverse bias), or the d.c. gate-source voltage used to
control the d.c. drain-source current in a field-effect transistor. The word is
also used as a verb: to switch.
3. In statistical usage, a source of error that cannot be reduced by increasing
sample size. It is systematic as opposed to random error.
Sources of bias include (a) bias in sampling, when members of the sample are
not fully representative of the population being studied; (b) nonresponse bias
in sample surveys, when an appreciable proportion of those questioned fail to
reply; (c) question bias, a tendency for the wording of the question to invite
an incorrect reply; (d) interviewer bias, a problem of personal interviewing
when respondents try to reply in the way the interviewer is thought to expect.
A narrower definition of bias in statistical analysis (see statistical methods)
is the difference between the mean of an estimating formula and the true value
of the quantity being estimated. The estimate
for the variance of a population is biased, but is unbiased when n is
replaced by (n - 1)