Chapter 6 - Effect Sizes and Confidence Intervals

Point estimate or single-number estimate

This number alone doesn’t tell the whole story of the data.

Effect size or practical significance

measures the size of the difference between two populations or the relationship between two variables 

WHAT n-size WHAT standard error 

INCREASING n-size DECREASING standard error 

P-hacking

the inappropriate manipulation of data analysis to enable a favoured result to be presented as statistically significant 

What are the two types of error:

1. WHAT

2. WHAT

What are the two types of error:

1. Type I error

2. Type II error 

Type I error or false positive

declaring significance when there is none

Type II error (false negative)

When you miss significance that is there 

The probability of committing a type I error is called WHAT

The probability of committing a type I error is called alpha (α)

The probability of committing a type II error is called WHAT

The probability of committing a type II error is called beta (β)

Type WHAT is less common 

Type I is less common 

Power

The likelihood of correctly rejecting the null hypothesis

Power depends on two things:

1. WHAT

2. WHAT

Power depends on two things:

1. How much sampling error there is

2. Sample size 

Effective size is measured using WHAT

Effective size is measured using Cohen’s d 

Cohen’s d ranges from WHAT-WHAT

Cohen’s d ranges from 0 - maybe 1 or 2 

Cohen’s d, eta squared (η2), and beta (β) are the WHAT generated by different types of statistical tests.

Cohen’s d, eta squared (η2), and beta (β) are the EFFECT SIZES generated by different types of statistical tests.