Hypothesis based on my teaching exp: A deep understanding of the concept "correlation≠causation" depends on one's ability to generate alternative causal explanations. Memorizing "correlation≠causation" doesn't transfer broadly, because students lack practice generating them.1/4
-
Show this thread
-
Which makes the problem harder, because content knowledge is important for generating alternative causal explanations! For example, consider the question: Are mean differences in achievement test score levels strong indicators of effective school district level education policy?2
1 reply 1 retweet 21 likesShow this thread -
Helpful background knowledge: 1) Far more variance in test scores within districts than between them; 2) selection into schools is highly nonrandom and depends on factors that plausibly influence achievement, e.g., SES. "Correlation≠causation" is a bad answer to this question.3
2 replies 1 retweet 17 likesShow this thread -
As usual, no single obvious solution, but something that I like to think helps my students is frequent practice generating these alternative hypotheses. Draw the DAG: district policies->test scores. Now tell me 3 potential confounds. Good, now onto design based solutions.4
3 replies 1 retweet 33 likesShow this thread
What does “design based solution” mean in this context? Tx!
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.