Stat Collecting Data

Census

a survey in which every item or subject in the population is selected

Population

the entire group you want information about

Clusters

distinct groups of data points in a scatterplot or a type of random sampling method

Bias

a systematic tendency to overestimate or underestimate a value

Undercoverage

certain individuals are less likely or unable to be chosen in a sample

Nonresponse

individuals chosen for the sample cannot be contacted or do not participate

Treatment

a condition imposed on individuals in an experiment

Experimental unit

the individual to which a treatment is applied

Replication

using enough subjects so differences aren’t due to chance

Completely randomized design

experimental units are assigned to treatments completely at random

Randomized block design

random assignment occurs within each block

Matched pairs design

a block design where units are paired by similarity or each subject receives both treatments in random order

Sampling

selecting a random group of people out of a whole population (that’s representative of the population)

Strata

group w similar characteristics assumed to be associated with the variables being measured

Prospective

tracks into future

Double Blind

neither subjects nor the people measuring know the treatment

Systematic random sample

randomly select starting point & select every #th individual after..make sure no patterns coincide with your systematic pattern

Simple random sampling

every possible group of the choses size also has an equal chance

Random Assignment (purpose)

Creates comparable groups in an experiment, allowing cause-and-effect conclusions. Causation

Cluster Sampling

Breaks population into groups, randomly picks some groups, and then surveys everyone in those groups

Blocking

controls for a variable that affects responses

Stratified sampling

divide by groups that matter and sample some from each group

Random sampling

representation/generalization