Chapter 3 Flashcards
Conceptual and Operational Variables
a. Variable: An attribute that varies within a project, has at least 2 levels.
i. Can be measured or manipulated
b. Constant: Something that could vary but has only one level.
c. Conceptual variable/Construct: Abstract concept defined by researcher (intelligence)
d. Operationalize: to turn it into a measured or manipulated variable
Claims
a. Frequency: Describes/Indicates the rate or level of some variable
i. 1 variable, measured
b. Association (correlations/covariances): argue that one level of a variable is likely to be associated with a particular level of another variable
i. 2 variables, both measured
ii. Types: Positive, negative, zero, curvilinear
iii. An association claim expressed with great certainty is not a causal claim—certainty is unrelated to the type of claim.
iv. Is linked to, prefers, is more likely to, may predict, is tied to, etc.
c. Causal claims: Argue that an association is casually directed; changes in one variable cause a change in the value of the other variable
i. 2 associated variables, one manipulated and one measured
ii. Causes, affects, promotes, reduces, etc.
Types of validity for Frequency Claim
a. Frequency: 39% of teens text while driving…
i. Construct: How do we operationalize and measure texting while driving?
1. Phone record, survey, etc.
ii. External validity: To what populations, settings, and times can we generalize our estimate?
iii. Statistical: Is our estimate accurate and reasonable> How large is margin of error and CI?
Types of validity for Association Claim
b. Association: Coffee consumption linked to lower depression.
i. Construct: how do we operationalize and measure coffee consumption and depression?
ii. External: What populations, settings, times can we generalize to?
iii. Statistical: Is the data linear, how strong is the association, is it likely an error occurred?
Types of validity for Causal Claim
c. Causal: Self distancing leads to greater persistence
i. Internal: are there are plausible explanations for the relationship?
1. Was an experiment conducted, is there temporal precedence, was there random assignment?
ii. Construct: Were the variables manipulated and measured appropriately?
iii. External: What populations can we generalize the results to?
1. Researchers often prioritize internal validity at the expense of external validity
iv. Statistical: How strong is the relationship self-distancing and persistence?
Validities
a. Frequency and Association rely on construct, external, and statistical validity; Causal claims rely on construct, external, statistical, and internal.
b. Construct: How well the variables are operationalized, measured, and manipulated
c. External: How well the claimed effect generalizes beyond the people, stimuli, and circumstances of the specific study
d. Statistical: How strong is the effect, is it statistically significant, how well does the study minimize type 1 errors?
e. Internal: How well a claimed casual statement can be supported against alternate explanations
Type I and II Errors
a. Type I: Incorrectly rejecting a null hypothesis (false positive)
b. Type II: Incorrectly failing to reject a null hypothesis (miss)
Criteria for establishing causation
a. Covariance/Correlation
b. Temporal Precedence (for causal variable): the cause must occur before the outcome
c. Internal Validity: All other potential causes should be ruled out as explanations for the observed outcome
Independent and Dependent Variables
a. Independent: Variable that is manipulated
b. Dependent: Variable that is measured
Random Assignment
a. Randomizing the group each subject will be assigned to
b. Effective at minimizing likely differences between groups before manipulating
Importance of Validities
a. Most important one depends on goals. For a causal claim, construct and internal might be most important. Sometimes prioritizing one validity means sacrificing another.
