External Validity
Define external validity
Are these findings generalizable? (i.e., can they be generalized to another, or external, population, or are they just specific to our population?)
Define internal validity
Is there a causal relationship between our independent (x) variable and dependent (y) variable?
Define construct validity
Are we measuring what we say we’re measuring?
Define hypothesis validity
•Hypothesis validity: Is our hypothesis ambiguous? Is it falsifiable?
•Statistical conclusion validity: Are our statistical methods of analyses and conclusions sound?
•How large was sample size?
•Low sample size = Low statistical power
•Statistical Power: The chance of finding something IF it exists
taking continuous data and making it categorical (i.e., if you have a numerical scale of 0-100, don’t turn it into “high” vs. “low”)
Provide an example of a threat to external validity
Provide an example of a threat to internal validity
Provide an example of a threat to construct validity
Using a scale of social connectedness to measure extraversion, using a scale of social anxiety to measure introversion
Provide an example of a hypothesis with weak hypothesis validity
"The researcher thinks happiness and exercise are related"
When evaluating statistical conclusion validity, what sample size do we look for?
n > 100 per group, or 100 overall if no groups
Provide an example of ecological validity
A study measuring participants' brain activity in a game of online chess (cannot generalize findings to chess players in general- in-person chess may be different)
Difference between independent and dependent variable
IV = what you're manipulating (predictor variable, x)
DV = what you're measuring (outcome variable, y)
Evaluate the hypothesis validity of the below hypothesis:
"Based on previous research findings that individuals living with chronic pain have higher average depression scores than people without chronic pain, Laura hypothesizes that pain and depression will be positively correlated."
Pretty strong (direction of relationship stated, based in theory/previous research)
Weaknesses: chronic pain same as "pain"?
What is an example of bias in research
journal bias for positive findings (file drawer effect)
researcher bias to find positive findings to succeed in academia, publish their study, etc.
You are a researcher looking to study how sleep impacts undergraduate students' GPAs. You hypothesize that sleep and grades are positively correlated (more sleep --> higher grades).
How would you make sure you have strong external validity? What would you want to ensure?
- have sample representative of the population you wish to generalize (undergraduate students)
What is a confounding variable?
A "third variable" that explains the observed relationship between the IV (x) and DV (y)- a variable that provides an "alternative explanation" to the relationship proposed in a study
Why is dichotomizing our data problematic?
We lose a TON of information!
ex: if I’m administering a new medication for doses between 0-100mg and find that the “high group” (51-100mg) is more efficacious than the low group (0-50mg), which dosage do I go with now? We don't know!