AP Stats Unit 3 Study Guide

Simulation

Using random #’s to accurately mimic reality of real life events in order to estimate probabilities or behaviors of a system.

Steps to simulation

1. Identify what is being simulated. Explain how you will use random #’s to represent outcomes of events & outline the process.

2. Run several trails.

3. Summarize your findings.

Parameter

Numeric measurement from the population

Sample

Numeric measurement from sample

Bias

Sampling method that tends to under or over emphasize some characteristics of the population.

Voluntary response (Bias)

Respondent selects whether they’re part of the survey or not (Example: American Idol because viewers choose whether or not to vote.)

Response (Bias)

Anything in the survey design that influences how people answer (Example: Loaded question or order of questions)

Nonresponse (Bias)

Large group fail to respond

Undercoverage (Bias)

Part of population underepresented

Census

A sample that contains all elements of the population

Sampling frame

A list of individuals that are used to create a sample

Different types of sampling

Random, convenience (weakest of the 5), systematic, stratified,and cluster

Random (Types of Sampling)

Each person has same chance of being chosen (Example: put names into hat, select 100)

Convenience (Types of Sampling)

Weakest of the five; choose students that are readily audible (Example: choose from football and friends)

Systematic (Types of Sampling)

Select every Kth member of population (Example: every 8th student in alphabetical list)

Stratified (Types of Sampling)

Divide into similar groups, select some from each group (Example: divide by class, survey all seniors)

Multistage

Combining 2 or more sampling methods together

Simple Random Sample

Each group has the chnace of being chosen (only deals with the group not the individuals)

Factors

Must have at least one factor that can be manipulated

Levels

Specific values of each factor

Treatments

Combinations of specific levels from each factor

Blinding

(Only for drug studies); a research practice where information about which treatment or group a participant belongs to is hidden from people involved in the study to prevent bias

Single Blind

Either the participant or the person conducting the experiment don’t know who is getting the real drug

Double Blind

Both the participant or the person conducting the experiment don’t know who is getting the real drug.

Observational Study

Collect results but the researcher doesn’t change or manipulate the environment in any meaningful way

Retrospective

(Past); Use information that has already been collected

Prospective

(Future); gather information from now until the future

Experiment

Manipulates the factors and levels to create treatments; must manipulate enviornment in some way

Placebo

Patient improves but has fake drug

Experimental Units

Nonhuman; can be inanimate object or animals

Subject

Human participants

Confonding

When the levels of 1 factor are associated with levels of another factor in such a way that can’t be seperated

Response Variable

WHat is being measured in experiment or observational study

4 Principles of Experimental Design

Blocking, Control Variables, Randomization, & Replication

Blocking

Not required but improve almost every experiment or observational study - dividing into similar groups

Control Variables

Most important element of any experiment, control sources of variation of all factors’ that are being studied

Randomization

Randomly assign people to treatments; never randomly assign people to your experiment (never done!)

Replication

Have someone recreate your experiment, if they get the same results, then maybe you can become famous

Cluster (Types of Sampling)

Divide into similar groups, select all from 1 group (Divide by class, survey all seniors.