1 of 22

The best three point shooters in the nba Those in top 50 in 3p% or top 50 3PM

  • What is going on here?
  • If steph is such a great shooter, why is his 3P% mid?
  • Why is there a negative relationship?

Think about

  • sampling error
  • Sampling Bias
  • DAGs

2 of 22

Class 01

3 of 22

Welcome back / Q&A

4 of 22

5 of 22

Foundation of stats

Inferring causation

6 of 22

Types of studies

In a Observational Study we measure things in the world.

In an Experimental Study we measure things in the world after poking them one way or another.

7 of 22

DAGs

Are tools to help us consider causes & confounds in observational studies

8 of 22

Another evaluation

9 of 22

Foundation of stats

Parameters vs estimates

10 of 22

Think / Pair Share

What is a parameter?

What is an estimate?

Is there anything you would like to change about this answer?

11 of 22

Populations Vs Samples

Parameters Vs Estimates

A population consists of all observation that could ever exists. � We often think about this as e.g. all individuals�Summaries of a population are parameters. Parameters are the truth

We never have information for the full population

Rather we have samples from a subset of the population�An estimate is out guess at a population parameters from our sample

12 of 22

Populations Vs Samples

Parameters Vs Estimates

A population consists of all observation that could ever exists. � We often think about this as e.g. all individuals�Summaries of a population are parameters. Parameters are the truth

We never have information for the full population

Rather we have samples from a subset of the population�An estimate is out guess at a population parameters from our sample

13 of 22

Populations Vs Samples

Parameters Vs Estimates

A population consists of all observation that could ever exists. � We often think about this as e.g. all individuals�Summaries of a population are parameters. Parameters are the truth

We never have information for the full population

Rather we have samples from a subset of the population�An estimate is out guess at a population parameters from our sample

14 of 22

Foundation of stats

Sampling, what could go wrong?

15 of 22

Think / Pair Share�

What is a sampling error?

What is an biased sampling ?

Describe each and provide an example

16 of 22

Sampling error, Sampling bias, Non-independence

Sampling error is the chance difference between a sample estimate and the true parameter. It is directionless and is what most stats try to deal with.

Sampling bias is when something about the sampling process makes it so that sample estimates will systematically differ from true population parameters.

Non-independent sampling is when something about your sampling procedure means that the capturing one observation changes your chance of getting another

E.g. sampling repeatedly from a dorm or house, or family etc..

17 of 22

Sampling error, Sampling bias, Non-independence

Sampling error is the chance difference between a sample estimate and the true parameter. It is directionless and is what most stats try to deal with.

Sampling bias is when something about the sampling process makes it so that sample estimates will systematically differ from true population parameters.

Non-independent sampling is when your sampling procedure makes it so that capturing one observation changes your chance of getting another

E.g. sampling repeatedly from a dorm or house, or family etc..

18 of 22

Sampling error, Sampling bias, Non-independence

Sampling error is the chance difference between a sample estimate and the true parameter. It is directionless and is what most stats try to deal with.

Sampling bias is when something about the sampling process makes it so that sample estimates will systematically differ from true population parameters.

Non-independent sampling is when your sampling procedure makes it so that capturing one observation changes your chance of getting another

E.g. sampling repeatedly from a dorm or house, or family etc..

19 of 22

Case Study 1: “Astrology test” link2setup link2quiz

Case Study 2: “The Stanford Marshmallow Experiment” link

20 of 22

Chime in grades

I need to fix them!

21 of 22

Class discussion about Case Study [5 min]

22 of 22

Comments/ Questions