Terms Test Statistics Ch 10-12

What is randomization?

A deliberate method of "haphazard arrangement" of observations as to simulate chance.

Why to Randomize

To reduce bias by spreading out the uncontrolled variables equally among groups.

What is Bias.

With respect to sampling systematically tend to select one group over another.

Bias = inaccuracy resulting in the over/underrepresentation of a certain group!

Models of Randomization

-Random number table
-Random number generator (calculator/ computer)
-Coin toss
-Dice rolls
-Card Draw.

Sampling method

-Sample random sample
-Stratified sample
-Systematic sample
-Cluster
-Multistage.

Experimental designs

-Completely randomized design
-Block design
-matched pairs.

To describe randomization and how it is used

-Clearly identify model used
-Describe labeling process
-For sampling, label population
-For experimental design, label all experimental units/ subjects
-indicate a process for handling "repeats" in random digits
-Use appropriate number of digits when assigning labels
Clearly indicate number of and size of groups with clear, "stopping points".

Population

The entire collection of individuals or things about which information is desired.

Sampling Frame

The list of members of the population that can be chosen for the sample.

Sample

A hopefully representative smaller part of the population selected for the study.

Variable

A characteristic whose value may change within one, and/ or among individuals.

Data

Information, in context, gathered from making observations on one or more variables.

Univariate Data

Data on a single variable.

Bivariate/Multivariate Data

Data on two or more than one variable.

Categorical Data

Nonnumerical data.

Quantitative Data

Data that can be measured with a number and units.

Bias

The tendency of a sample to differ from the corresponding population in some systematic way.

Undercoverage

Introduced when one portion of the population is systematically excluded.

Response

Introduced when the method of observation or the equipment used produces values systematically different from the population.

Nonresponse

Introduced when data is not obtained from all individuals selected from the sample.

Voluntary Response

Introduced when data is obtained only from those who willingly choose to respond.

Good Sampling Technique

Sample Random Sample (SRS)

A sampling method in which every sample of a given size is equally likely o be chosen.

Stratified

Dividing the population into subgroups and then taking a random sample from each.

Multistage

Dividing the population into subgroups and choosing one or more groups which then may be subdivided again until a random sample is chosen from one or more of the resulting subgroups.

Cluster

Dividing the population into subgroups and then using one of those groups as the sample.

Systematic

Choosing a sample according to some "rule" (every nth time).

Observational Study

A study of existing data without attempting to influence the response.

Experiment

A procedure in which the effect of a treatment is investigated.

Factor(explanatory variable)

Variable thought to affect the response.

Levels of

A factor can occur in different values (levels).

Treatment

What is imposed on the units, is a combination of factors and levels.

Experimental Unit

What receives the treatment in the experiment.

Lurking or Confounding Variables

A variable that may affect the response but which is not of interest.

Response Variables

Variable being measured and compared at the end of the experiment.

Control

Holding lurking or confounding variables consistent.

Randomization

Using chance ro assign units to treatments.

Replication

Applying the treatment on more than one unit.

Control Group

A group that receives no treatment or a placebo treatment.

Blind

When the subject is unaware of which treatment they are receiving.

Double Blind

When neither the subject nor the administrator of the treatment know which treatment is being used.

Blocking

Creating experiment groups that are similar by an extraneous factor.

Matched Pairs

A form of blocking where the two groups are identical.

Placebo

A treatment that looks like other treatments but has no level or value.

Relationship VS Cause and Effect

Only after an experiment can a cause and effect relationship be stated.