Stats 4.1

Bias

A study flaw that leads to unrepresentative and/or inaccurate estimates

Under coverage bias

When part of the population has a reduced change of being included in the sample

Digital survey —> lack of access to technology

Nonresponse bias

When individuals chosen to be part of a sample don’t participate —> tend to only get people who feel strongly about the topic to participate

Voluntary response bias

When individuals in a study are volunteers —> tend to only get people who feel strongly about the topic to participate

Something released to many people —> released on the radio or social media

Question wording bias

When questions are confusing or leading the participant a certain way

“What are your feelings on homework holidays, knowing that you will have a lot of assignments on Tuesday?”

Self reported bias

When individuals in the study report their own traits —> responses can be dishonest which will affect collected data

Survivor bias

When only subjects that withheld certain circumstances are sampled —> ignoring the subjects that did not survive

What can you do to avoid bias?

Sample randomly

Simple Random Sample

A sampling method where every possible group of individuals in the population has an equal chance of being selected

“every possible group of individuals” —> every possible combination of individuals has an equal chance of being selected

If you have A, B, and C then all of the following combinations have an equal chance of being selected as the sample size:

A and B

A and C

B and C

Two key methods of SRS

1) Drawing chits

2) Random number generator (used more)

3 steps to describe the sampling process

1) Assigning numbers to the population: “assign every individual in NYC a different/unique integer (can add number: from 1 to n);” “from 1 to n where n is the population”

2) Random number generator: “use a random number generator to select 1,000 integers, skipping any repeated values”

3) Distributing the survey: “administer the survey to the 1,000 students whose numbers were selected (in the previous step)”

ARD —> assign values, random number generator, distribution of survey

Acronym for describing how to do a SRS

ARD —> assign values, random number generator, distribution of survey

Stratified random sampling

When you divide the population into groups called strata based on a characteristic they share, and then randomly select individuals from each group.

Ex: 100 freshman, 100 sophomores, 100 juniors, 100 seniors —> grade levels are strata (the groups), and 10 subjects from each strata (grade level) are chosen for the sample

Benefit: most representative as people from every group are included

How is strata determined?

By variables that have an impact on the response

Ex: How much homework do you receive a night? —> grade level would be strata here because it impacts responses; freshman tend to get less homework

Clustered random sampling

When population is divided into groups called clusters, and a few out of all of the clusters are selected. Every individual in the selected clusters is surveyed.

Ex: Divide the students in a school into clusters based on advisory (advisories are the clusters). Then, randomly select 4 advisories, and survey every person in each of the four selected advisories.

Benefit: easiest as it saves time, money, and is convenient

-Clusters “mirror population,” because they feel like a mixdd group

Systematic random sampling

When you select individuals at a regular interval after randomly choosing a starting point.

“Begin marking every 8th seat until you get a sample of 10 seats”