skip to main content

Enough Already about "Black Box" Experiments: Studying Mediation Is More Difficult than Most Scholars Suppose

DONALD P. GREEN, SHANG E. HA, ; JOHN G. BULLOCK

Annals, Vol.628 pp.200-200, 2010 [Periódico revisado por pares]

Texto completo disponível

Citações Citado por
  • Título:
    Enough Already about "Black Box" Experiments: Studying Mediation Is More Difficult than Most Scholars Suppose
  • Autor: DONALD P. GREEN, SHANG E. HA, ; JOHN G. BULLOCK
  • Assuntos: Experimentation; Experimenters; Relationship; Dramatically; Introduction; Intervention; Experimental; Explanatory; Continually; Counteracts; Civil Procedure; Criminal Law & Procedure; Evidence
  • É parte de: Annals, Vol.628 pp.200-200, 2010
  • Descrição: ... Keywords : mediation; causal mechanisms; causal inference; experiments An academic phenomenon that occurs with astonishing regularity may be observed whenever experimental researchers present evidence of a causal effect. ... For example, the introduction of limes into the diet of seafarers in the eighteenth century dramatically reduced the incidence of scurvy, and eventually twentieth-century scientists figured out that the key mediating ingredient was vitamin C. ... Conventional regression approaches to the study of mediation rely on strong and often implausible assumptions, even when applied to data in which a causal factor has been manipulated experimentally. 2. ... In the spirit of making this example mimic an ideal experiment, suppose these two unobserved factors were independent of one another and drawn independently for each observation. ... The bottom line is that when subjects are governed by different causal laws, analyses that presuppose that the same parameters apply to all observations may yield biased results. ... One can learn a great deal of theoretical and practical value simply by manipulating variables and gauging their effects on outcomes, regardless of the causal pathways by which these effects are transmitted. ... For any researcher working in the early phases of an experimental research program, devoting resources to the manipulation of mediators (and investigation of subgroup differences in causal effects) is a gamble, as there is no guarantee that the experimental intervention will produce a substantively interesting average effect on the outcome.
  • Idioma: Inglês

Buscando em bases de dados remotas. Favor aguardar.