Week 3 - Three claims, four validities Flashcards
Define and differ measured and manipulated variables?
Measured are observed and recorded
Manipulated are controlled by experimenter
What is a frequency claim?
Described rate or degree of a single variable
Impt: only one measured variable
What is an association claim?
What are the three types?
One level of a variable is likely to be associated with a particular level of another variable
Impt: two variables are measured
Type 1: posi, move in the same direction
Type 2: nega, move in separate directions
Type 3: zero, no correlation
What are the three claims?
Frequency - rate of one degree
Association - one is related to one
Causal - change in one causes change in one other
List the four validity?
Construct
External
Statistical
Internal
Define construct validity?
Define external validity?
- How well a variable is operationalized, does the measured variable represent the construct?
- How do we know the results can be generalized, do the results say something beyond the measured sample?
Define statistical validity? - what are the two types of errors?
Define internal validity?
- Are the conclusions consistent with the statistical analysis
a) Type I - concluding there is an effect where there isn’t one
b) Missing an effect when there is one (false neg) - In a relationship between one variable and another, he extent to which a rando variable could have an effect (third variable problem)
What needs to be satisfied for causation to be plausible?
How can you test causation?
Co-variance - there is an association between one variable and another
Temporal precedence - one came before the other in time
Internal Validity - there are no alternative explanations (confounds) for the associations
Experiments, friend, experiments
How do you interrogate a frequency claim?
C - how well were variables measured
E - generalizability, how pick and are they representative
S - recognizing that there is a margin of error included within the study
How do you interrogate a association claim?
C - assess the construct validity of each variable, how well each was measured
E - will it generalize
S - the extent to which the statistical conclusions are accurate and reasonable.
Type I error (false positive): mistake an association where there is none
Type II error (miss): mistake no association where there is one
How do you interrogate a causal claim?
Evaluate three criteria for causation (T, C, I)
Perform experiments - random assignments increase internal validity
C - measure the validity of both variables
E - do the results generalize
S - how strong, stat wise, are the correlations
Provide an example of a manipulated and measured variables
IV: Amount of salt added to water
DV: Speed of the water boiling
