unit 2 test

What does SRS stand for

Simple random Sample

What are the three steps of a SRS

1. Label individuals

2. Randomize (random number generators)

3. Select

What are individuals

A person or item in a population

What do we need to do with individuals for sampling

The individuals need to get assigned (write down names) ( give them a number)

What are the three popular samplings methods

1. simple random sample

2. Convenience sample

3. Voluntary sample

What is a convenience sample

People that are easy to reach are sampled

What is a voluntary sample response

People choose to respond to the question

What sampling methods are biased

1. Convenience sample

2. Voluntary response sample

Why is a voluntary response sample biased

People choose to respond, so that means they care more about the topic. This leads to extreme reviews, which are typically negative

Why is a convenience response biased

it is likely not representative to the whole population

What is census

A census is when you collect data for the whole population

What is challenging about a census

It is difficult to get everyone to respond, and it can be expensive

What are stratified random samples

Stratified random samples splits populations or samples into groups. A simple random sample is then taken

What is a strata

A strata is a group of individuals with similar characteristics

What does with or without replacement mean

With replacement means individuals can be sampled twice. Without replacement means an individual cannot be sampled twice

Why is a stratified random sample better than a simple random sample

A stratified random samples ensures all strata are represented, so it’s more representative

What does it mean to have a representative sample

Having a similar or same percentage of the total population in the sample

What does low bias mean

Low bias means it’s close to the true mean (answer) because there’s little systematic error.

What does low variability mean?

It means that the data is close together and not spread out

Does every possible sample combination have an equal chance of being selected in a stratified random samples?

No, some combinations are mathematically impossible to occur due to the various strata groups

What do stratified random samples affect in term of variability

It can decrease variability because it reduces the risk of sampling an unbalanced sample

What does homogenous mean in a sample?

Homogenous means that the sample is consistent, regular, uniform, and equal

What does heterogeneous mean in a sample

Heterogenous means a sample or group is assorted, diverse, mixed, and not equal

what does a Stratified random sample from

It samples some from all homogenous strata

What does a cluster random sample sample from

It samples all from some groups

Is a cluster sample homogenous or heterogeneous

It is heterogeneous because it samples a group even if they aren’t alike

Is a stratified random sample homogeneous or heterogeneous

It is homogenous because the strata are alike

Can you have multiple strata?

Yes you can have sub-strata

What is a benefit of a cluster sample

It is easy and includes a full group

What’s the benefit of a stratified random sample?

Each strata is represented so no group is left out. This allows for better data accuracy

What’s the benefit of a systematic random sample?

It is fast and easy while being more precise than a cluster sample

How is a cluster random sample and a stratified random sample similar?

They both break individuals up into groups

How is a cluster and stratified random sample different

A cluster takes all from some groups while a stratified random sample takes some from all strata

What is the under coverage bias?

It is a bias in which you sample a group while deliberately ignoring a certain group to where that group is not represented

What is a non-response bias

It is a bias in which an individual is not reached or refuses to respond

How is a non-response bias and a under coverage bias different

While they both leave out people, the way they leave out people is different. For example, under coverage would be not attempting to knock on doors in a neighborhood, while non-response would be knocking on doors in the neighborhood but the person isn’t home

What is a response bias?

A response bias is a problem in how you are gathering the data. The question could be worded in a way that is biased or the scenario could be presented in a bias way (a firefighter wearing his uniform to ask a question about the fire department)

What are the two poor sampling methods that can lead to bias

Convenience sampling and voluntary response can lead to bias

If you have a non-response bias, can you reduce the bias by sampling more people

No, sampling more people will not reduce the bias because there will be more people that won’t respond still

When designing a sample survey, what is the best way to reduce non-response bias

Make sure that there are multiple ways to contact and follow up with the participants

What is a population

The total group (football players in a county)

What is a sample

A portion of the population (30 randomly selected football players in a county)

Does increasing sample size decrease variability and bias?

Increasing sample size only decreases variability

What is variability

Answers varying by chance

What is bias

Systematic errors that pull results in a direction

How do you reduce bias

Randomization and blinding reduces bias

How do you decrease variability

Blocking for an experiment and stratifying for an observation decreases variability

What are the steps in a study

1. Get experimental units

2. Randomly assign groups

3. Give treatment to group

4. Compare findings

What is an explanatory variable

The explanatory variable is the independent variable. In other words it’s the variable that causes (explains) the result

What is a response variable

It is the dependent variable. It is the variable that happens because of the explanatory variable.

What is a confounding variable

A confounding variable is a variable that can skew both the explanatory and response variable. It typically skews the response variable. It is something that may influence or confuse a variable.

How can you help mitigate confounding factors?

Use randomization

What is the difference between a observational study and an experiment

In an experiment there is a treatment. In an observations study there is not a treatment. Causation can occur from an experiment.

What is an experimental unit

What/who the treatment is imposed on

What are treatments

What is done or. It done to experimental units

Can a control group be a treatment

Yes it is a treatment

What four things need to be present for a well-designed experiment

1. There needs to be two or more things to compare

2. There has to be random assignment

3. The study needs to be able to be replicated ( have multiple participants or do it multiple times)

4. There needs to be a control (baseline) group

How do you do random assignment

1. Label the individuals

2. Randomize the groups into treatments

What is the placebo effect

The placebo effect is when a fake treatment works due to human bias

How do you prevent the placebo effect?

1. Single blinding

2. Double blinding

What is single blinding

Single blinding is when just the participants don’t know whether they have the treatment or control treatment

What is a double blind

A double blind is when neither the researcher or the participant knows the group

What is blinding used for

To prevent the placebo effect?

What is the point of double-blinding

It helps remove bias, even if unintentional, from the researcher. He or she can’t act differently around the groups because he or she doesn’t know who’s who

What are blocks?

It is a group of similar experimental units that have similar traits

What is a randomized block design

1. Experimental units are split up into two groups (typically) (for example A and B

2. The experimental units are then grouped into blocks

3. The blocks are then randomly assigned the treatment (for example 1a 2a, 1b 2b

4. You compare the sub groups (for example 1a vs 2a and 1b vs 2b)

5. You then compare the results from the sub groups

What is completely randomized design

The experimental units are randomly assigned into treatment groups. (Basically simple random sample but instead of getting a sample it’s assigning people into groups)

What is a match paid design

Subjects are paired together in groups of 2 and then randomly assigned treatment. For example the top two students in the class would be pair and so on. One person in a group would get one treatment while the other person will get a different one. Who gets what treatment is random

What does a random sample allow us to do?

A random sample allows us to generalize our conclusions to the population from which it is sampled. This means we can make inferences for the population.

What does random assignment allow us to do

Random assignment allows it to conclude a treatment causes changes in the response variable. This means we can make an inference about causation.

What wording can be used if a random sample is used

Association can be used

What word can be used if random assignment is used

Causation can be used

what type of study allows for association

An observational study

What type of study allows for causation

An experimental study

If there is random assignment and random sample what inferences can be made

There can be inferences about a population and cause

What inference can be made if there is a random assignment but no random sample

There can be an inference of cause, but there can not be an inference for population.

What inference can be made if there is a random sample but no random assignment

There can be an inference about the population, but there can not be an inference about the cause

What inference can be made if there’s no random assignment or random sample

No inferences can be made