General
What is critiquing & what does it involve?
It's deciding whether we trust. It involves understanding how constructs are measured.
Types of claims?
Frequency claim: Descriptive only but doesn't explain why
Association claim: correlations & relationship but doesn't explain why
Casual claim: explain cause & effect (best but hardest to achieve)
A) Face Validity vs Content Validity
F: is when a test looks like it measures what it’s supposed to
C: is when a test covers the right things
Face Example
A sleep researcher makes a survey asking, "I wake up feeling rested." The questions clearly look related to sleep quality.
What is a construct & how to make it measurable?
It's a variable that cannot be measured. You turn the abstract concept into something measurable.
Reliability vs Validity
R: a measure yields the same result repeatedly (consistent)
V: a measure yields consistent and accurate results (if there's V there's R)
B) Criterion Validity?
is the degree to which a test performs well against an independent criterion
Content Example
A test on academic motivation includes questions about effort, persistence, and goal setting.
3 ways to Measure
Self Report, Observational, Physiological
3 Types of Reliability?
Inter-rater reliability: degree of agreement among independent observers
Test-retest reliability: same score on each measurement on something that should remain consistent
Internal reliability: items on scale are measuring the same construct
Two types of criterion validity?
Predictive: is the degree to which a test predicts a future outcome or behavior
Concurrent: is the degree to which a test correlates with another established measure or outcome at the same time
Criterion Predictive Example
A new driving test predicts who passes the official road exam.
Operation definition?
ways we measure abstract ideas
Construct validity path a & b?
A: two subjective ways to access validity
B: three empirical ways to access validity
B) Convergent vs Discriminant Validity
C: is when two different measures of similar constructs show a strong positive correlation (they produce similar results)
D: is when two measures of unrelated constructs show little to no correlation (they stay separate)
Criterion Concurrent Example
A psychologist gives patients a new depression questionnaire and compares their scores with a clinical interview for depression done that same week.
To critique claims:
1) Decide what type of claim it is 2) Examine 3 types of reliability 3) Examine types of validity
Construct Validity
How well measurement actually measures the construct it claims to measure
Internal vs External Validity
I: Carefully controlled to rule out all possible alternative explanations
E: Designed to apply in real-world situations