Producing Data Designing Samples Experiments Deliberately change a variable in order to observe response Here you are actually INFLUENCING the response.

Producing DataDesigning Samples

Experiments

• Deliberately change a variable in order to observe response

Here you are actually INFLUENCING the response

Only plausible way to show cause & effect

You CAN control lurking variables to help minimize their

effect

SRS (Simple Random Sample)

• A random sample where each individual has an equal opportunity to be chosen and each sample has an equal opportunity of being producedRandom # Table Calculator

Assign each item a #

Given a starting pt, look for (__) digit #’s within the

appropriate interval

randInt(lowest#,highest#, # desired)

Pick 3 Random Citizens

52 3 41

6 7 8 9 10

11 12

13500 51807 42130 23171

randInt(1,12,3)

Alternative Sampling Methods

• Stratified Sampling• Divide the population into categories

(strata) and select a random sample from each category

• Cluster Sampling• Divide the population into categories

(clusters), randomly select one or more categories, and select EVERYONE within the chosen categories.

Pick 3 Random Citizens

52 3 41

6 7 8 9 10

11 12

We’ll group them by age

Here are the groupings!!

1

1 2 2

3

3 4

4

randInt(1,4,1)

randInt(1,4,1)

randInt(1,4,1)

Pick 3 Random Citizens

We’ll group them by location.

Here are the groupings!!

randInt(1,4,1)

But this is a little different

1

2

3

4

Stratified Sampling

• The students have been asked by Governor Martin O’Malley to survey the state of Maryland to determine the level of participation by teenagers in the state’s public school system. Describe 2 ways to design a stratified sample for this study.

1)Split the state into Counties (Strata). Assign a number to each individual in each county and take a random sample of n people from each county. PROBLEMS???

2) Split the state into Zip Codes (Strata). Assign a number to each individual in each zip and take a random sample of n people from each one. PROBLEMS???

How would you change these to be Cluster Sampling?

Try This One Too!!

Describe how I would collect a random sample of 15 students

from this class of 50. SRS?, Stratified? Cluster?

Talking about the “People”

• Be careful when making inferences from a sample to the population

Biggest Problem/Misuse

In StatsOVERGENERALIZATION

Collecting Data from “relatively” small sample sizes and making inferences about a population

The Bigger Your Sample – The More Accurate

Your Inferences.

Get a BIG Sample!!At least 15; try for 40 or more if possible.