hypothesis testing

null hypothesis

• States that the observation is due to chance or random factors 

  • IV does not influence the DV 

  • No difference between groups/conditions 

  • No real relationship between 2 variables 

  • The manipulation does not work the way you think it does 

• More conservative, baseline assumption (null hypothesis is the ‘boring’ hypothesis→ says nothing cool is happening) 

• Alternate hypothesis (research hypothesis) 

• States that the observation is not due to chance or random factors 

  • IV does influence DV 

  • Difference between groups/conditions 

  • Relationship exists between 2 variables 

• Researchers usually want to find evidence that supports Ha

The 2 hypotheses should be formulated so that they are

• Mutually exclusive: cannot have both hypotheses be true at the same time 

• Exhaustive: cover all possible outcomes 

• 1st: fine 

• 2nd: fine 

• 3rd: not exhaustive 

• 4th: not mutually exclusive (they can both be true at the same time)

• Non-directional v directional

• non-directional: When you dont pick a particular direction (can go either way)

  • (Ha doesnt pick a direction, it didnt pick higher and it didnt pick lower, just different)

• directional: picked direction —> stress improves memory/ Ha: stress impairs memory

•  True state of affairs 

The truth about the null hypothesis

Type 1 error

• Rejecting the null hypothesis when it is actually true 

• ‘False alarm’/ ’false positive’

 Type 2 error

• Retaining the null hypothesis when it is actually false 

• ‘Miss’, false negative

Can you make a type 1 error and type 2 error at the same time?

No. they are mutually exclusive errors

always begin with assuming the null hypothesis is true 

E.g. there are no unicorns in the world → easier to disprove bc youd just have to find 1 unicorn to disprove it.

(idrk)

do you retain or reject Ho if probability of obtaining the data you got (or more extreme data) if the null hypothesis is true is high? and same if the probability is low