AP Statistics Unit 5

What is needed for conclusions about a population to be valid?

A sound sampling design.

What is sampling?

Studying part of the population to learn about the whole.

What is a census?

Contacting every individual in the population.

What is a voluntary response sample?

People choose themselves by responding to a general appeal.

Why are voluntary response samples biased?

People with strong (often negative) opinions are most likely to respond.

What is convenience sampling?

Choosing individuals who are easiest to reach.

What is bias in sampling?

A method that systematically favors certain outcomes.

What is an SRS?

A sample where every set of n individuals has an equal chance of being chosen.

What does an SRS guarantee?

Each individual and each possible sample has an equal chance.

What is a table of random digits?

A long string of digits 0-9 with each digit equally likely and independent of others.

What are the steps to choose an SRS?

Label individuals; Use Table B; Stopping rule; Identify sample.

What is a probability sample?

A sample chosen by chance where all possible samples and their probabilities are known.

What is a stratified random sample?

A sampling method where population is divided into similar groups (strata), a SRS is taken from each, and combined.

What is cluster sampling?

Divide population into clusters, randomly select clusters, include everyone in chosen clusters.

What is undercoverage?

Some groups are left out of the sampling process.

What is nonresponse?

Chosen individuals cannot be contacted or do not cooperate.

How does wording of questions affect surveys?

Leading or confusing wording can cause strong bias.

What is response bias?

Bias caused by respondents’ or interviewers’ behavior (e.g., lying, interviewer influence).

What are experimental units?

Individuals the experiment is performed on.

What are subjects?

Human experimental units.

What is a treatment?

A specific experimental condition applied to units.

What is an explanatory variable?

A variable that explains or influences the response.

What is a response variable?

What is measured as the outcome.

What is a factor?

An explanatory variable in an experiment.

What is a level?

A specific value of a factor.

What is control in an experiment?

Effort to minimize variability in how units are obtained and treated.

Why is control needed?

To protect from lurking variables that affect the response.

What is a control group?

Group receiving no treatment or a placebo for comparison.

Why is a control group important?

It isolates treatment effect by providing a baseline.

Difference between control and control group?

Control = limiting lurking variables ; control group = comparison group that doesn’t experience the treatment.

What is the placebo effect?

Response caused by expectation, not the treatment.

What is lack of realism?

When subjects, treatments, or environment do not match real life.

What is a matched pairs design?

Two treatments compared using pairs of similar subjects.

How is randomness used in matched pairs?

Randomly assign which subject in each pair gets which treatment.

What is another version of matched pairs?

Each subject receives both treatments, serving as their own control.

What is a block?

A group of subjects known beforehand to be similar on variables affecting response.

What is a block design?

design where experimental units are arranged into blocks to account for unwanted variation. By grouping similar subjects or conditions together, researchers can reduce variability within each block.

What is a completely randomized design?

All units are randomly assigned to treatments.

What can completely randomized designs compare?

Any number of treatments, from one or multiple factors.

What is treatment?

A specific condition applied to units.

What is a group in experiments?

Subjects receiving the same treatment.

What is statistical significance?

An effect so large it is unlikely to happen by chance.

What are the 3 principles of good experimental design?

Control lurking variables; Replicate treatments on many units; Randomize assignment.

What is replication?

Applying each treatment to many units to reduce chance variation.

What is randomization?

Using chance to assign units to treatments.

Why compare multiple treatments?

Helps control effects of lurking variables.