Ch 5 & Ch 6

Question 1

Reliability

Answer 1

Consistency of measure and Established empirically

Question 2

Reliability measures

Answer 2

• Test-retest: stable constructs
• Inter-rater: observations
• Internal: multi-item self-reports

Question 3

Test-retest vs replication

Answer 3

• Test-retest reliability: same sample measured twice on same operationalization
• Replication: new sample to test if there is same relationship between two variables

Question 4

Validity

Answer 4

• An evaluation of how well that operationalization is measuring what it is supposed to measure
• Measurement reliability and measurement validity are separate steps in establishing construct validity

Question 5

Reliability vs Validity

Answer 5

• Not the same or opposites
• Reliability is necessary but not sufficient for validity

Question 6

Criterion Validity

Answer 6

Correlating the measure with some form of behavioural evidence or concrete outcome (depression measure with a diagnosis)

Question 7

Convergent & discriminant validity

Answer 7

Correlating the measures of the a) same thing (con.) or b) distinct constructs (dis) (depression measure with a) other measures of depression, and b) measures of physical health)

Question 8

Construct Validity in Surveys - Wording of Questions

Answer 8

• Leading questions (phrasing a question with an indication of what the answer is (+/-)
• (Double) Negative wording (Highly cognitively tasking)
• Double-barrelled questions (Two questions in 1, making it hard to know what the answer is for)

Question 9

Construct Validity in Surveys - Response Sets

Answer 9

Patterns of responding, like yay saying (only yes/(+) answers), nay saying (only no/(-) answers), fence-sitting (middle/neutral response especially present for controversial or confusing questions) and faking good/bad (making oneself look better/worse

Question 10

Construct Validity in Observations - Observer bias

Answer 10

When observers’ expectations influence their interpretations of behaviour or the outcome of the study; observers cannot have any expectations of results so they are not lead to see their expected patterns

Question 11

Construct Validity in Observations - Reactivity

Answer 11

A change in behaviour when a participant (or subject) knows someone is watching, which can lead to measuring affected behaviour rather than natural behaviour; even present when human-like avatar observed people (sense of feeling watched)