Le scienze sociali vanno di moda e soprattutto quelle comportamentali.
Un gruppo di ricercatori (“mostly consisting of empirical psychologists who conduct research on basic, applied and meta-scientific processes”) ha pubblicato un breve lavoro in cui evidenziano i rischi di errori che le social and behavioural sciences possono produrre se applicate indiscriminatamente nelle policies pubbliche.
I loro dubbi riguardano i seguenti sei profili:
<< First, study participants, mainly students, are drawn from populations that are in Western (mostly US), educated, industrialized, rich and democratic (WEIRD) societies. Second, even with this narrow slice of population, the effects in published papers are not estimated with precision, sometimes barely ruling out trivially small effects under ostensibly controlled conditions. Third, many studies use a narrow range of stimuli and do not test for stimulus generalisability. Fourth, many studies examine effects on measures, such as self-report scales, that are infrequently validated or linked to behaviour, much less to policy-relevant outcomes. Fifth, independently replicated findings, even under ideal circumstances,are rare. Finally, our studies often fail to account for deeper cultural, historical, political and structural factors that play important moderating roles during the process of translation from basic findings to application. Together, these issues produce empirical insights that are more heterogeneous than might be apparent from a scan of the published literature>>
Propongono quindi un percorso a nove livelli per arrivare alla evidence readiness della loro scienza, sintetizzato nella figura 2 dell’articolo (sulla scia dei nove technology readiness levels della NASA, come si legge ivi).
Confortante che un gruppo di scienziato condivida in pubblico le preoccupazioni per errori nel metodo o nella prassi seguiti nel loro settore di attività.