Biostats Week 2/1

Simple Random Sample

The total possible size of people is selected so that all possible combinations of individuals in the population are equally likely to be in sample

undercoverage

some groups are underrepresented on the population listing

volunteer bias

Self-selected participants of a survey tend to be uncommon in population

nonresponse bias

large percentage of individuals refuse to participate in a survey/cannot be contacted

Probability Sampling

Each member of the population has a known probability of being selected into the sample (ex:srs)

Stratified Random Sample

Draws independent SRS from within homogeneous groups- specific to that group

Cluster Samples

Randomly selecting clusters of people consisting of smaller subunits

Multistage Sampling

Large-scale units are selected at random

Experimental studies

One group gets exposed while other stays nonexposed

Observational studies

Investigators merely classify individuals as exposed or nonexposed- doesn’t interfere

Placebo

Even if a change happens ( and doesn’t affect anything ) we expect change to be felt

Explanatory variable

Treatment/exposure that explains or predicts changes

Response variable

Outcome/response being investigated

Lurking Variables

Not included in a statistical analysis but still affects the relationship between variables being studied.

Randomized controlled trial

assignment of the treatment is based on chance

Double blinding study

Investigators and study subjects making measurements don’t know

Triple-Blinded

Study subjects, investigators, and statisticians analyzing the data are kept in the dark

Equipoise

Balanced doubt