Chapter 5 class version b(1)

Ch. 5 continued:Measuring Variables and Sampling

9.25.2012

Road Map

• Discuss Exam 1

• Quick review of 9/20 class

• Finish Chapter 5 material (sampling)

Exam 1

• Will share class statistics next class– (Brian will enter grades in BbLearn right now)– Mean on multiple choice = 83%

• Discuss short answer questions

• Any questions about multiple choice?

Quick Review

• Scales of measurement• Reliability• Validity• Appropriate use/interpretation of reliability

and validity information

SAMPLING

Sampling• Very important part of research methodology

Let’s establish some vocabulary:

• Population- full set of elements that exist

• Sample- a set of elements taken from the population

• Element- the basic unit of sampling

More Lingo

• Sampling method

• Representative sample

• Equal probability of selection method – (EPSEM)

Statistics vs. Parameters• Statistic• Parameter

• Sampling Error: Difference between the sample values and the “true” population value

• illustration

• Inferential Statistics: goal is to draw conclusions (inferences) about population based on sample statistics

Even More Lingo

• Census

• Response rate: % of people selected to be in the sample who actually participate in the study

• What if our response rate is really low?

Will the Lingo Never End?

• Biased Sample: A non-representative sample

Hopefully you have:• Proximal similarity: generalization to people,

places, settings, and contexts similar to those described in the study

Sampling Methods

Simple Random Sampling

• popular and most basic type of random sampling

• Think slips of paper in a hat

• With replacement• Without replacement – preferred

Stratified Random Sampling

• Divide the population into mutually exclusive groups (strata)

• Then select a random sample from each group

• Stratification variable

• Proportional vs. Disproportional

Cluster Random Sampling

• Cluster- collective type of unit that includes multiple elements (people)

• Examples?– Schools, classes, families– clusters are randomly selected

Systematic Sampling

1. Determine the sampling interval (k)2. Randomly select an element between 1 and k3. Select every kth element.

• Sampling interval- The population size divided by the desired sample size– symbolized by the letter k

example

• Population N=100 • Desired sample n =10• k = Population N/sample n = 100/10 = 10• Step 1 select element between 1 and 10– we randomly select 7

• Now select every 10th (k) element• Sample= 7,17,27,37,47,57,67,77,87,97

Warning for Systematic Sampling

• Periodicity - problem if there is a cyclical pattern in the population from which you’re sampling

Example:• If I have lists of classes organized by student

grade (highest to lowest) and the length of each class list is = to k.

• Might always be selecting the A or F students.

Nonrandom Sampling

• Weaker method (less representative of population)

• But sometimes necessary for practical reasons

• Four types– Convenience sampling– Quota sampling– Purposive sampling– Snowball sampling

Convenience Sampling

• Use of people who are readily available, volunteer, or are easily recruited for inclusion in a sample

Two most common research participants

• White Rat• College student

Quota Sampling

• Researcher sets quotas– numbers of the kind of people wanted in the

sample

• Then locates (via convenience sample) the numbers of people to meet the quotas

Purposive Sampling

• Sampling of individuals who meet specific criteria/ characteristics

Snowball Sampling

• When research question requires individuals who are hard to find

• Example: researching HIV+ white females

• Start with small group who meet criteria• They spread the word

– “snowball effect”

Random Selection vs. Random Assignment

• Random Selection: select participants for study– Purpose: create a representative sample

• Random assignment: place participants in experimental conditions – Purpose: create equivalent groups for use in an

experiment

How do we determine sample size?

• If population <100 then get them all

• In general, get as big a sample as possible

• Sample size calculator: G*Power