AP Statistics Unit 3

Confounding variable

An unforeseen factor that is related to the explanatory and response variable and generates an apparent relationship between two variables

Observational study

A study without imposing treatments; you cannot infer a cause and effect relationship

Retrospective observational study

Examine current or past data for a set of individuals

Prospective observational study

Follow a sample of individuals into the future collecting data

Experiment

A study where different treatments are imposed on subjects

Census

Collects data from all individuals in a population

Simple random sample

A sample in which every group of a given size has an equal chance of being choen

Cluster random sample

Population is divided into clusters, SRS of clusters is taken, and all individuals between chosen clusters are sampled

Stratified random sample

Population is divided into strata based on a similar characteristic, SRS within each stratum is taken, selected individuals are combined into a larger sample

Systematic random sample

Randomly choose a point, then sample at a fixed periodic interval

Bias

A systematic tendency to favor certain responses over others

Variation

A measure of precision

Undercoverage bias

When part of the population has a reduced chance of being included in a sample

Non-response bias

When individuals chosen for a sample don’t respond

Voluntary response bias

When an invitation is sent out to all individuals in a population to participate and those choose to participate differ from those who don’t

Question-wording bias

When survey questions are confusing or leading

Self-reported response bias

When individuals inaccurately report their own traits

Principles of a well-designed experiment

• Compares at least two treatment groups

• Random assignment of treatments

• Replication

• Control of potential confounding variables where appropriate

Placebo

“Fake” treatment that is similar to the treatments being tested

Placebo-controlled

The use of a placebo to determine if an effect is truly due to the treatment and not simply because of the placebo effect

Single-blinding

When subjects are not aware of the treatment being administered

Double-blinding

When both the subjects and the researchers are not aware of the treatment being administered

Matched pair design

A form of a randomized bloc design that uses blocks of two similar experimental units, one receiving each treatment, or both receiving both treatments in random order

Statistically significant difference

Difference larger than would be expected and is evidence that the treatment caused the effect

Convenience sample

A sample that involves selecting readily available members of the population rather than using any underlying probability-based selection