Dear Colleagues,
Some may be interested in this seminar.
P
What’s wrong with ‘What Works for Whom, Where’?
Speaker: Nancy Cartwright, University of Durham and University of California, San Diego
Venue: UCL Institute of Education, Room G16, 9-11 Endsleigh Gardens, London WC1H 0ED
Date: Wednesday
22 November
Time: 12.30 – 13.30
As we know, Random Controlled Trials, if well blinded and monitored, can be taken to give an unbiased estimate of the mean treatment effect in the population enrolled in the study. The first thing to note is that without additional assumptions, the estimate
need not be very precise. This means that the result from one run of the experiment may give results far from the true mean on that population. The second thing is to ask: If this particular population with these specific individuals in it is not the one of
interest, what use can be made of the estimate of its mean treatment effect?
It is now widely acknowledged that few social interventions work the same way. Hence the move to ‘What works for whom, where?’. This conceals a heavy metaphysical assumption well beyond the empirical evidence: that the intervention has what philosophers
call a ‘goal-directed power’ towards the outcome.
This talk will argue that it is unlikely that many of interventions we consider in social policy have any such ‘goal-directed power’. When they work they generally do so because local arrangements call a number of different causal principles into play
together that would not be at work together in other arrangements.
Biography
Nancy Cartwright is Professor of Philosophy at the Department of Philosophy, University of Durham and at the University of California, San Diego (UCSD). She is past President of the Philosophy of Science Association and was President of the American Philosophical
Association (Pacific Division) in 2008.
Her research interests include philosophy and history of science (especially physics and economics), causal inference, causal powers, scientific emergence and objectivity and evidence, especially for evidence-based policy.