STATS Sampling and Surveys

Population

the entire group of individuals we want information about

Census

collecting data from every individual in the population

sample

a subset of individuals from the population from which we actually collect data

sample survey

a study that uses an organized plan to choose a sample that represents a specific population

bias

when a study design consistently undestimates or ovestimates the value you want to know

convenience sample (type of bias)

What is it: Choosing individuals who are easy to reach

Problem: Usually biased because easy-to-reach people may be different from the whole population.

Example: Asking the first 30 students you see in the library about homework time

voluntary response sample (type of bias)

What is it: People choose themselves by responding to an invitation

Problem: Usually biased because people with strong opinions are more likely to respond

Example: Online polls where people click to vote

Simple Random Sample (SRS); (good Sampling method)

What is it: Every group of n individuals in the population has an equal chance to be selected

How to do it:

• Label each individual with a number

• use technology or a random number table to select individuals

Why is it good: Avoids bias by using chance to select the sample

Stratified Random Sample

What is it: Divide the population into groups (strata) that are similar within, then take the SRS from each group.

Example: Divide students by grade level, then randomly select students from each grade

Advantage: More precise estimates when groups are different from each other.

Cluster sample

What is it: Divide the population into groups (clusters), randomly select some clusters, include all individuals in chosen clusters.

Example: Randomly select 5 classrooms and survey all students in those 5 rooms.

Advantage: Saves time and money when individuals are naturally grouped.

Systematic Random Sample

What is it : Select every kth individual from a list after randomly choosing a starting point.

How to do it: Divide population size by desired sample size to get k, randomly select a starting point between 1 and l, then select every kth individual after that.

Example: To sample 50 students from 500, select a random start between 1-10, then select every 10th sample.

Advantage: Easy to carry out and spreads the sample evenly across the population

The 10% Condition

the sampling size should never be more than 10% of the population size

Undercoverage (type of bias outcomes)

Some memeber of the population cannot be chosen in the sample

Example: Phone surveys miss people with no phones

Nonresponse (types of bias outcomes)

An individual chosen for the sample can’t be contacted or refuses to participate

problem: People who don’t respond may be different from those who do

Response bias (type of bis outcomes)

The behavior of the interviewer or respondent causes bias

Example: People pay not answer honestly about sensitive topics

Quetion Wording

How a question is asked can influence the response

Example: leading questions push people towards a certain answer