On Wed, 21 Jul 2010 06:32:01 -0700,, Scott O Lilienfeld wrote: >On a more conciliatory note (maybe I'm just in a charitable mood >this morning), Mike P. and I do agree on the merits of Don Campbell's >writing (and I agree with Mike that Shadish, Cook, and Campbell >is a treasure trove of a resource for thinking critically about >quasi-experimentation). Also, a quotation from Campbell from his >classic 1969 American Psychologist article (Reforms as Experiments), >seems apropos here: > >"The advocated strategy in quasi-experimentation is not to throw up >one's hands and refuse to use the evidence because of this lack of control, >but rather to generate by informed criticism appropriate to this specific >setting as many plausible rival hypotheses as possible, and then to do >the supplementary research..which would reflect on these rival hypotheses."
Another quote seems apropos here. Shadish (2010), in a Psychological Methods article that compares the causal models implied by Campbell's and Donald Rubin's (of missing value analysis and "potential outcomes" theory) approaches, makes the following statement: NOTE: CCM = Campbell's Causal Model |Perhaps the important difference is the greater emphasis on |human and scientific fallibility in CCM. CCM is paradoxically |skeptical about the possibility of doing the very things that CCM |sometimes requires to generate good causal inferences. Humans |are poor at making many kinds of causal judgments, prone to |confirmation biases, blind to apparent falsifications, and lazy |about both design and identifying alternative explanations. Yet, |in the cases where a stronger design cannot be used or fails, |CCM is remarkably dependent on those judgments for identifying |threats, deciding whether they have been rendered implausible |in any given case, and summing over them to reach a |causal conclusion, especially in weaker nonrandomized experiments. |CCM argues, therefore, that the responsibility for being |critical lies more with the community of scholars than with any |given researcher, especially with those whose interests would |most lead them to find fault (Cook, 1985). Of course, CCM |benefits from technical advances, like propensity score analysis, |that reduce reliance on fallible judgments. Hence, CCM has |turned attention to incorporating such developments (e.g., Luellen, |Shadish, & Clark, 2005; Shadish, Luellen, & Clark, |2006). (page 8). William R. Shadish (2010). Campbell and Rubin: A Primer and Comparison of Their Approaches to Causal Inference in Field Settings. Psychological Methods, 15(1) 3–17. -Mike Palij New York University [email protected] --- You are currently subscribed to tips as: [email protected]. To unsubscribe click here: http://fsulist.frostburg.edu/u?id=13090.68da6e6e5325aa33287ff385b70df5d5&n=T&l=tips&o=3708 or send a blank email to leave-3708-13090.68da6e6e5325aa33287ff385b70df...@fsulist.frostburg.edu
