Hypothesis Testing and Inferential Statistics

Question 1

How can we identify the difference between two or more data sets?

Answer 1

WE DON’T KNOW.

We CAN however calculate the probability of getting a large difference.

Question 2

What is a null hypothesis (Ho)?

Answer 2

The Ho is the probability that there is NO difference observed in your data (- nothing of interest is happening)

Question 3

What do inferential statistics enable us to do?

Answer 3

To CALCULATE THE PROBABILITY of differences in data through CHANCE.

Question 4

What is the difference between what we are looking for and how the statistics operate?

Answer 4

We are look for a DIFFERENCE in our data.
Statistical tests are testing that there is NO DIFFERENCE.
They are built on ACCEPTING and REJECTING the Null hypothesis (Ho).

Question 5

What do you do when you have formed a Ho?

Answer 5

Select a suitable statistical test to test the Ho and give you the probability of being true.

Question 6

What is the P-value?

Answer 6

The probability result from a statistical test which is used to accept or reject a Ho.

Question 7

What % difference is often used to define a P-value result?

Answer 7

0.05 or 5%

Question 8

When do we reject or accept the Ho with the P-value?

Answer 8

< 0.05 = REJECT the Ho (something interesting is happening) - the satistical alternative is accepted (Hi)
> 0.05 = ACCEPT the Ho (Nothing interesting is happening)

Question 9

List the 4 steps of Statistical Testing

Answer 9

1. Construct a Ho
2. Decide on the critical significance level (nearly always 0.05)
3. Calculate statistic + P-value
4. Reject or Accept Ho

Question 10

What are Type I and Type II errors?

Answer 10

Type I
You REJECT the Ho when it is actually TRUE. (always a chance it is right!)

Type II
You ACCEPT the Ho when it is actually WRONG to do so.

Question 11

What do Ho and Hi denote?

Answer 11

Ho - Null Hypothesis

Hi - Research/Information Hypothesis