Journal of Eco:lomic Behavior and Organization 3 (1982)65-e,1. North-Holland SEARCH FOR RULES FOR SEARCH John D. HEY* University of York, York YOI 5DD, :5 K Received June 1981, final version received July 1982 This l~aper presents the findings of some prelinfinary 'laboratory' investigati~, as into 'actual" search behavior. Speeitieally we looked at situations in which searchers' initial in~mretic,v -,bout the distributions was negligible. It seemed likely that simple 'rules of thumb" wo~tC '~; used in such ,'~tuations. Our findings confirmed this view: we ":dzn..t~ed five such rules "~,i~ic~ betw~n them "accounted' for a high proportion of observed behaviour. Moreover, these ~les, ,lthough not 'optimal') appear to be reasonably good and may be fairly robust; joint~, taey may consti :late a bettt;r explanation (and predictor) of actual search behaviour than do the carrently populur supposedly optimal rules. 1, Introduction and overview The purpose of this paper is to present the fiadings of some preliminary laboratory-type investigations into 'actual' search behaviour. Primarily motivated by a growing conviction that the increasingly sophisticated models of optimal search behaviour appearing in the economics literature are a decreasingly accurate reflection of reality, our investigations were designed to discover the actual rules that individuals use when carrying out search activity. Of particular interest to us were the rules used in situations whe,:e initially searchers had essentially no knowledge of the distribution over which they were searching. Casual empiricism and introspection suggested to us that simple 'rules of thumb' would be employed in such situations [sec Hey (1981)]. Our investigations were ~esigned to see whether, in fact, this was the ease, and is so, to determine the folm of the roles of thumb :Lctually employed. The findings reported [ere are clearly preliminary, in the sense tlaat our investigations suggested l.~ us a ,.vhole range of other possibilities which *1 am greatly indebted to Stephen Baker, Athan Balfc,ussias, Stella Balf~ussias. Peter Barker. Richard Barnett. Robert Beedl'am, Eduard Berenguer, Nell Blake, Da~,i~ G~wland, t~rla~: Hillier, John Hutton, Pa't Kemwn. Men, Korchid. i'eter Lambert, Guillermo Lopez, Anne Ludbrook, Edwaid Lynk, Alan Neilson, Don Poskitt, John Rae, Martin Reynold~,~, Robert Risohx~, Harold Seddighi, Peter Solar, Eiieen St:tcliffe, Andrew Tremayne, Stephen Trotter, Anthony Weekes, Jayne Wilae, Robert Woodlield and Simon Woodward for taking part in m? experime~lt. I am also greatly indebted to Professors Richard Day and Sidney Winter and two anonymous referees for helpful comments on an earlier draft ~f this pape:. ()16~' "~.& " ~/{*~ lf){~¢~ "¢dl'~ 75 ' I OQ~ N o r t h - } [ o l l a a d ~. g ,. :00/'0 .... , . .......... ~ ~o.