Ragin, Charles (2008) Fuzzy-Set Social Science

Embed Size (px)

Citation preview

  • 8/12/2019 Ragin, Charles (2008) Fuzzy-Set Social Science

    1/3

    Southern Political Science ssociation

    Fuzzy-Set Social Science by Charles C. RaginReview by: Bear BraumoellerThe Journal of Politics, Vol. 70, No. 1 (Jan., 2008), pp. 291-292Published by: Cambridge University Presson behalf of the Southern Political Science AssociationStable URL: http://www.jstor.org/stable/30218885.

    Accessed: 13/07/2014 13:14

    Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at.http://www.jstor.org/page/info/about/policies/terms.jsp

    .JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

    content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

    of scholarship. For more information about JSTOR, please contact [email protected].

    .

    Cambridge University Pressand Southern Political Science Associationare collaborating with JSTOR to

    digitize, preserve and extend access to The Journal of Politics.

    http://www.jstor.org

    http://www.jstor.org/action/showPublisher?publisherCode=cuphttp://www.jstor.org/action/showPublisher?publisherCode=spsahttp://www.jstor.org/stable/30218885?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/30218885?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=spsahttp://www.jstor.org/action/showPublisher?publisherCode=cup
  • 8/12/2019 Ragin, Charles (2008) Fuzzy-Set Social Science

    2/3

    BOOK REVIEWS 291Fuzzy-Setocial Science. yCharlesC. Ragin. Uni-versityfChicagoPress, 000.)doi: 0.1017/S0022381607080309Charles aginwouldhavebeenhardpressedocomeupwith book that s more ikely,t first lance, odeflect he nterest f large-N oliticalmethodolo-gists ndpractitioners.thas fuzzy nthe itle,ndto researchersnterestedn precision,hat can'tbegood. t takes eriouslyhepropositionhatnecessaryandsufficientonditionsxist. tsfoil,hroughouthefirsthapterspecially,s conventionaluantitativeresearch, hosepractitionersre blindto so manyaspects f he ocialworld hat ven generouseaderwouldhave oconclude hat hey re rather im.Settinghebookasidefor hese easonswouldbea veryubstantial istake,owever. agin's ocus asmuch moreto do with nalyzingausalcomplexitythan withfuzzy etsper se; indeed, he latter reutilized npursuit f the former. ndanyquantita-tiveresearcher ho has contemplatedsing nterac-tion ermsn a simple egressionquation asalreadyadmitted hat causal complexity-in hiscase, thecontingencyf theeffectf onevariable n the evelofanother-may appen.Themainquestions not,therefore,hethertexists,utrather owwestudyt.Ragin's nswer o that uestions anextensionfthe brilliantlynnovative ualitativeComparativeAnalysisQCA) technique hat he laid out in his1987 ook,TheComparativeethod. CA'semphasisonBoolean lgebras a remarkablyfficienteans funderstandingausal complexity,oupledwith tsutilityn even verysmall-Nsituations,made itimmenselyttractiveoqualitativeesearchers.uan-titativeesearchersere ess enamored f its deter-ministic ormulationsnd substantialensitivityomismeasurement.Fuzzy-SetSocial Science reflectsts author'sconsiderable eflectionn these ssues n the nter-vening13 years.On offers a new and improvedQCA-fuzzy-setQCA, or fsQCA-thatreplaces hestrict ither-orormulationsfQCA with hemorenuanced oncept fdegree fmembershipn a set.Here, gain, ifferencesith uantitativeesearchreeasyto exaggerate:nyonewho has formulatedsurveyuestionskingwhetherespondentsre veryreligious, somewhateligious, notvery eligious,or not religious s already hinkingn fuzzy-set-theoreticerms. he maindifference,tseems,ies nhow he nswersthen uantified:mindless -3-2-1coding, agin rgues, oes notreflectheresearcher'ssubstantivenowledgef hemeaningf he nswer.bettermeasure freligiosity ight e 1-0.8-0.2-0,r

    somethingimilar,ependingn the xactwordingfthe uestion. y he ame oken, measuref theismwould notnecessarilyimply e the nverse f themeasure freligiosity:ven a little eligiosityoulddisqualifyhe espondentrom he etof theists,o abetterodingmight e 0-0-0.2-1. ne canargue heexact umbers,rhow heyhould ederived,ut helarger oint s thatwe're ctuallypeakinghe amelanguage.Thefirsttep ngettingromodingmembershipin fuzzy ets to understandingausal complexityrelies n thevery mportantnsighthat bsences fdatacan be as importants, if not more mportantthan, heirpresence.The absenceof warsamongdemocratictates,or xample, as pawned massivescholarlyiteraturend has evenbeen noticed nthepolicyrealm.)These absences ell us quite a bit,according o Ragin: dearth f observationsn theupper-leftalf f n X-Y catterplotuggestrelation-shipof causalnecessity,hereas dearth fobserva-tions n the ower-rightalf uggest relationshipfcausal sufficiency.gain,the languagemakesthedifferenceeem larger han it is, as the dyadicdemocracys sufficientorpeace formulationllus-trates icely.Theremainderf themethod orunderstandingcausal complexityomes down to addressingheproblem fcombiningausalconditions.orexam-ple, one might rguethat so manyyoungpeopleenter aw schooleachyear hat he markets nearlysaturated,nd onlyyoung ttorneys ho arenatu-rally iftednd who workveryhardwillhaveverysuccessfulareers. he argumentmplies hat hereshould e a triangularelationshipetweenmember-ship in the set of giftednd studiouspeople andmembershipnthe etofsuccessfulawyers, ith heformerstablishinghe imit o the atter: ew fanylawyershould emore uccessfulhan heyregiftedand studious. ut even fwe canmeasure egree fgiftednessusingnformationromollegeGPA, ankofcollege, ATs, LSATs,etc.) and amountof timestudyingndconverthosemeaningfullyntodegreesof membershipn the sets of gifted eople andstudiouspeople,we are still eftwiththeproblemof aggregation:f Jones'membershipn the set ofgifted eople s 0.8 and hermembershipn the etofstudious eople s0.5,what shermembershipn thesetofgiftednd studious eople?Ragin's esponsesthat heminimum efines he onjuncturef he wosets:Jones'membershipn the set of gifted ndstudious eople s 0.5. Similarly,ones'membershipinthe etofgifted R studious eoplewouldbe themaximumf thetwo, r0.8.)

    This content downloaded from 152.74.16.35 on Sun, 13 Jul 2014 13:14:12 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 8/12/2019 Ragin, Charles (2008) Fuzzy-Set Social Science

    3/3

    292 BOOK REVIEWSAt this point, even sympatheticuantitativeresearchers aywellrebel, ecauseboththe cutoffvalue fornecessitynd sufficiencyX = Y) and therules or ombining ultipleauses min(xl,x2) ndmax(xl,x2))are peggedby assumption atherhanallowed ovary s thedata demand.Ragin ecognizesthispoint n thefirstnstance,ntroducing fuzzyadjustment how he could resist fuzzyfudgefactor, don'tknow)to movethediagonalup ordown bit ocaptureases hat renear he ine.Buttheuse ofminimand maximas a dauntinglytrongassumption.t rules utthepossibilityhat ne of hetwo factorslays bigger ole nproducingheout-come.ThoughJoness moregiftedhan tudious,orexample,tudiousness ay ontribute ore o successthangiftednessn general: erhapsmin(giftedness,2*studiousness) ould better apture he quantityrelevanto success.Notallowing nyweightingf hevariables ill onstituteveryubstantialandicap opeoplewhobelievenestimatingoefficientsasedonthedata.Opinionwill robablye somewhat ore ividedon the ole statisticalomponentf theprocedure,binomial est o determine hether r not thepro-portionf successes casesbelow heX = Y linefornecessity,r above forsufficiency)s significantlygreaterhan given utoff-0.65 or usually eces-sary and 0.80 for almostalwaysnecessary or

    example. sidefromhefact hat, ven o thosewhogrant probabilisticonception f necessity,neexceptionn fivemaywell stretchhe conceptofnecessityeyond cceptableimits,heres a purelystatisticaloint hat houldbe mentioned.he80%benchmarks chosen ecause trepresentssurpris-ingly igh ercentageelativeowhatwewould xpectbychance.But themaximizationnd minimizationoperationsn the data have an impact n whatwe

    would xpect y hance: akinghemaximumf hreevariableswhose observationsre drawnpurely trandomfrom uniform istributionn the unitinterval illproduce newvariable hat s heavilyskewed owardhehigh ndof he cale. fweplot hisnewvariable gainstnother ariable, ,with bser-vationsagain drawnat randomfrom uniformdistributionn theunit nterval,t would not at allbe surprisingofind hat 0%ofthem all elow heX = Y line. When triedtmyself,ithn = 20,theaveragewas75% across100trials.)One couldsurelytailormore easonableenchmarksasedonexpectedoutcomesforrandom data giventhe numberofobservationsnd the operations hatproducedX,and theprocedure ouldbenefitromt.Finally, ivenRagin'sfocus on sets it seemssurprisinghat he has nothing o say aboutwhatresearchersnthe etofunconventionaluantitativescholarshave to sayaboutcomplexity. iven thatconventionaluantitativecholarshipeems o con-sist of linearregressionnd correlations,his s asubstantialmission. nly nteractionerms remen-tioned splausibleuantitativenalogs ofsQCA,ndtheirshortcomings-theeed for a largeN, thedangersfmulticollinearity-arexaggerated.In all, the book constitutes very nnovativeapproach o aninterestingroblem,nd the lementsof hat pproachremore amiliarhan quantitativeaudiencemightuspect. hat s nottosay hat twillcommand niversalgreement,ut it should causeeventhe most conventionaluantitativecholar oquestion he conventional isdom n manyways.Qualitativecholars,nturn,ouldbenefitrom heirreactions.nshort,t s a book hat erymuch eservesto be talkedbout, ndthatshigh raisendeed.

    BearBraumoeller,he OhioStateUniversity

    This content downloaded from 152.74.16.35 on Sun, 13 Jul 2014 13:14:12 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp