1.
anecdotal evidence: an informal story or observation
2.
bias: designs resulting in samples that do not represent the population
3.
blocking: subjects are separated into groups (blocks) with similar characteristics to eliminate condounding variables; treatments are then assigned at random within each block
4.
census: effort to record data about entire populatioin
5.
cluster sample: a sampling design in which entire groups are chosen at random
6.
confounding variable: a variable other than the treament that has a large influence on the response variable
7.
control group: a group with no active treament designed to measure influences on response other than treatment
8.
convenience sample: a sample of an easy-to-find group (subjects are not likely to reflect the attitudes of the population)
9.
double-blind: subjects and evaluators do not know which subjects are being assigned which treatments
10.
experiment: researchers assign treatments at random to subjects to see if response variable differences are caused by treatment differences
11.
matched pairs design: a block of size two in which one treament is assigned at random to each subject, both subjects are similar in many respects so differences in response are primarily due to differences in treatments
12.
nonresponse bias: subjects refuse to participate (usually in voluntary response samples)
13.
observational study: treatments are not assigned to subjects; response variable differences cannot be proven to be caused by treatments
14.
placebo: a treament has no active influence; subjects unaware
15.
population: all subjects to be studied
16.
response bias: wording of questions or interviewer attitudes can affect responses
17.
sample: subgroup of population about which data is actually collected
18.
simple random sample: each sample has the same chance of selection to represent the population
19.
statistical significance: the effects of other treatments are so large that they would be likely to recur in repeated trials
20.
stratified random sample: divide the population into subgroups with common characteristics, then select a simple random sample inside each subgroup
21.
treatment: different settings of the explanatory variable
22.
undercoverage bias: some groups within a population have no chance of being selected in a sample
23.
voluntary response sample: subjects volunteer to be included in sample; usually only strong opinions reply
