Unit 3 (AP Classroom 1B): Data Collection

Population

The entire group of individuals we want data or information about.

Sample

The subset of the population we collect data from.

Census

Contacts every individual of the population.

Bias

Using a method that favors some outcomes over others. Systematically favoring certain outcomes (sometimes due to confounding or lack of control).

Types of sampling methods

Includes: Voluntary response, Convenience sample, Simple Random Sample (SRS), Stratified Random Sample, Cluster Sampling, and systematic sample.

Voluntary Response Sampling

Very Biased; Participants choose if they want to response. The researcher makes no attempt to ensure the survey is completed.

Convenience Sampling

Very biased; Individuals chosen because they are easiest to reach (the most convenient).

Simple Random Sample

SRS; Every group of size n has an equal chance of being chosen.

Cluster Sampling

When the population is divided into heterogenous groups and some of those groups are randomly selected.

Systematic Sampling

Sampling in a pattern. Every "nth" person. From a list, or as people walk by a location.

Observational Study

We observe individuals without trying to influence the responses.

Experiment

We deliberately impose a treatment in order to see whether or not it causes a different response.

Confounding

When the explanatory variable is linked to a third variable (the confounder) and the confounding variable also affects the response variable.

Undercoverage

(Type of bias) Groups of the population are admitted or not included in the sample selection and those groups are different than the sample in an important way.

Non Response

(Type of Bias) Members selected can't be contacted or choose not to answer/participate in the study and those individuals are different than the people who do respond.

Stratified Random Sample

When we have homogeneous groups and randomly sample from each of those groups (to ensure representation from each of the groups).

Experimental Units

The individuals on which an experiment is performed (usually this term is used for non-human subjects). The smallest unit to which treatments are randomly assigned.

Subjects

The experimental units (when human).

Treatment

Specific condition applied to the subjects.

Factors

The variables in experiment that the researcher is manipulating.

Level

The specific values of the factor the researcher uses.

Placebo

An inert treatment used to compare results/fake treatment.

Control Group

The group receiving the placebo or the standard treatment. Used as a baseline of comparison.

Random Assignment

Using a random process to assign experimental units to treatments.

Replication

Use enough experimental units in each group so that the any differences in the effects of the treatments can be distinguished from chance.

Double Blind

Neither the subjects nor the evaluators knows which treatment is being applied to the groups. (a research assistant makes a secret code for the end of the experiment)

Block Design

Separating subjects into similarities before applying random assignment to each.

Matched Pairs Design

A type of block design. Either the subjects are sorted into blocks of size two. Or, each subject is compared to itself--each subject receives both treatments but in a random order.

Statistically significant

Differences in results from treatment groups and control are large enough that it is unlikely they would occur by random chance alone.

Statistics

values we calculate from our sample

parameter

numerical summary of a population