02_Understanding Hypothesis Testing

Question 1

Based on inferential statistics: how to decide if results are reliable/reproducible and not due to chance (error) ?

Answer 1

All inferential statistics tests start by assuming the NULL hypothesis—any differences observed are due to chance and not reliable
H0 = There is no reliable difference between outcomes H0 = There is no reliable difference from 0.0 (corr.)

Question 2

Goal: Reject the NULL hypothesis and accept the ALTERNATIVE hypothesis. 
The alternative (or research hypothesis) is our working hypothesis—what we expect to happen

Answer 2

Goal: Reject the NULL hypothesis and accept the ALTERNATIVE hypothesis. 
The alternative (or research hypothesis) is our working hypothesis—what we expect to happen 

H1 = There is a reliable difference between outcomes H1 = There is a reliable difference from 0.0 (corr.)
We can have multiple alternative hypotheses: H1, H2, H3, H4, etc.

Question 3

Logic of Hypothesis Testing?

Answer 3

RQ: Do physicians spend less time with obese patients than non-obese patients?

H0 = There is no statistically significant (i.e., reliable) difference between the mean amount of time doctors spend with obese patients and the mean amount of time doctors spend with non-obese patients.

H1 = There is a statistically significant (i.e., reliable) difference between the mean amount of time doctors spend with obese patients and the mean amount of time doctors spend with non-obese patients.

Mean time obese = 24.7 min.
Mean time non-obese = 31.4 min.
Is the observed difference (6.7 min.) due to chance?
p = 0.0057
Low probability that the observed difference is due to chance; reject the null

Question 4

What is a Causal Hypotheses?

Answer 4

a. The researcher has a specific hypothesis about a causal relationship, and 

b. It is possible, feasible, and ethical to manipulate the causal variable and randomly assign participants to conditions or order of conditions 


Question 5

What is a non Causal Hypotheses?

Answer 5

a. The researcher does NOT have a specific hypothesis about a causal relationship, and 

b. It is NOT possible, feasible, or ethical to manipulate the causal variable and randomly assign participants to conditions or order of conditions