4a: 3.1-3.4 Sampling

Question 1

In Stats, what does random mean?

Answer 1

Equally likely to be chosen

Question 2

What is the main goal of taking a sample?

Answer 2

Getting a group that represents the whole population well

Question 3

What does it mean for a sample to be biased?

Answer 3

Any data gathered from the sample will not accurately represent the whole population. If you repeated the method many times, you would consistently either over or underestimate the real proportion/mean.

Question 4

What is a simple random sample (SRS)?

Answer 4

Every possible group of n individuals is equally likely to be chosen.

Question 5

Both split the population into groups first and then use some kind of random process, so what is the difference between stratified and cluster samples?

Answer 5

Stratified is when you take some from each group. Cluster is when you randomly pick some groups and then take everyone from those groups. Ideally, stratified is used when the groups are different from each other (you want to make sure to get some old people and some young people in your sample).

Question 6

What is voluntary response and why/how does it produce bias?

Answer 6

Rather than selecting people using a random process, you let people choose to respond by calling or going to a website or something. Generally, only people who are especially passionate (and usually in the same direction) will take the time to respond.

Question 7

What do you need to include in your answer when describing bias?

Answer 7

1. Give a possible reason for how the sample or experiment does not represent the population. 2. State whether that statistic obtained will be an over or underestimate of the true value

Question 8

How would you take a stratified sample in the cafeteria?

Answer 8

Randomly select 2 (or however many you need) students from each table

Question 9

How would you take a cluster sample in the cafeteria?

Answer 9

Randomly select 2 (or however many you need) tables and sample all the students at those tables

Question 10

Bias: how do you tell if a sample will be an under or overestimate?

Answer 10

Think: what “number” will I get from this sample? Will it be higher than the real number or lower? Example 1: I think that the percentage of my sample of students in detention will agree that the tardy policy is fair will be smaller than the real percentage of all students at the school. So this sample will underestimate the real percentage. Example 2: If I sample heights during break, the average I will get will clearly be a bigger number than the real average of people in the building. So, this sample would be an overestimate of the real height.