L8 - hypothesis testing

Question 1

hypothesis testing steps

Answer 1

1. Formulate hypothesis
2. Collect data
3. Construct statistic (t,F,x2..)
4. Calculate value
5. Check distribution table
6. Make conclusion:
   Result is significant / there is a relationship / hypo is supported

Question 2

Pr(H0|data) != Pr(data|H0)

conditional probability

Answer 2

result does not tell us about the truth of the null hypothesis, given the data…

What it tells us is the probability of the data, given the truth of the null hypothesis

Question 3

Questionnaire question types

Answer 3

• Dichotomous (male/female)
• Nominal (action/adventure/thriller/comedy)
• Cumulative =guttman scale (chech the statement you agree with)
• Interval =5/7 point likert

Question 4

Reverse coding

Answer 4

Avoid response style bias = yea/nay saying

Question 5

Common method bias:

Problem + solution

Answer 5

Problem: fact that your X and Y are measured with the same method or from the same source may drive your results.
Especially when X + Y are perceptual instead of factual from same respondent.

Solution:

1. X + Y from different sources
2. If 1 is not possible, measure X + Y in different time/space (like separate study)
3. Control social desirarability, especially when studying perception /behavior

Question 6

Retrospective bias

Answer 6

Respondents reconstruct the past to make it consistent with subsequent performance results, conventional story lines, or current beliefs

Question 7

Self-serving bias

Answer 7

Individuals’ tendency to refer to internal attribution for their success and external attribution for their failure

(failure=bad luck, success=ability OR 85% rate other drivers bad, itself excellent)

Question 8

Non-response bias

Problem + solution

Answer 8

Problem: Systematic differences between respondents and non-respondents may drive your results

Solution:

1. Show there are no significant differences between respondents and non-respondents on a number of demographic variables
2. Do the same for early and late respondents (e.g. before and after you send a reminder)

Question 9

Back Translation

Answer 9

process of translating a document that has already been translated into a foreign language back to the original language, preferably by an independent translator

(translate english into dutch + dutch into english to compare)

Question 10

Sampling

Answer 10

Probability sampling = random
• Simple random sampling (n out of N)
• Stratified/proportional/quota random sampling (strata)
• Systematic random sampling (every kth)

Nonprobability sampling
• Accidental/Convenience sampling
Purposive sampling: snowball sampling / quota sampling

Question 11

Statistical power

Answer 11

Type 1 = a
Falsely concluding h1 = false positive

Type 2 = b
Falsely concluding h0 = false negative

Question 12

OLS

Answer 12

= Ordinary least squares = smallest quadrating method > Simple Linear Regression .

Question 13

H0

Answer 13

The null hypothesis is not always a zero coefficient