Sie sind auf Seite 1von 20

Asking Too Much, Expecting Too Little

Author(s): Stanley Lieberson


Source: Sociological Perspectives, Vol. 31, No. 4 (Oct., 1988), pp. 379-397
Published by: University of California Press
Stable URL: http://www.jstor.org/stable/1388967
Accessed: 14/08/2009 06:46

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=ucal.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.

University of California Press is collaborating with JSTOR to digitize, preserve and extend access to
Sociological Perspectives.

http://www.jstor.org
ASKINGTOO MUCH,
EXPECTINGTOOLITTLE

STANLEYLIEBERSON
University of California, Berkeley

We know there are myths in all professions,deceptionsthat


practitionersfoist on both the publicat largeandthemselves.We
knowthat actualpracticedeviatesfromthe overtrulesandpublic
posturethatis taken.Whatpeopleor organizationssay theydo is
often far different from what they actually do. Why should we
think that our field is any different?I want to considerwith you
the sharpdiscrepancybetweenwhatwe do andthe claimwe make
when we addresstopics that are controversialor in the public
limelight. I am convinced that our discipline follows certain
implicitrulesof thinkingthat aretotallyinappropriate,illogical,
and ultimately undermineour ability to advance knowledge
about society. Many of the examples will be from the field of race
and ethnicrelationsbecausethat is the areaI knowbest. But the
problemis widespread,coveringfar morethan that subject,so I
hope you will view these as illustrations rather than as a
delimitationof the issue. Thereis good reasonwhy the problem
exits-unfortunately, what I think of as strangling our discipline,
others may think is the sheer joy of sociology. Indeed, it is
possiblethat some people of good will and kind intent maywell
findthe implicationsof mycommentsratherrepulsive.Sociology
is a ratheramorphousand diversefield with many competing
paradigms,but I believe we all share the notion that we are a
truth-seekingdisciplineand what I have to say deals with that
task.

AUTHOR'SNOTE:Thisarticleis the 1988PresidentialAddressgiven at the


annualmeetingof thePacificSociologicalAssociationat Las Vegas,Nevada,on
April 7.
SOCIOLOGICAL PERSPECTIVES, Vol. 31 No. 4, October1988379-397
? 1988PacificSociologicalAssn.

379
380 SOCIOLOGICAL
PERSPECTIVES/ OCTOBER1988

THE AMBIGUOUSROLEOF
EMPIRICALSOCIALRESEARCH
Therearethosewho essentiallyrejectallempiricalinformation
(or at least statistical data) because either they claim it is an
inkblotthat you can interpretor massageany way you want, or
becausetheyholdthatempiricalanalysiscannotreallyget at what
is of importance.Butthatis the extremeviewand,I wouldguess,
is held by only a small segmentof our discipline.Most of us, I
believe,do not fall into this extreme;we wouldnot be troubled,
for example,to learnthat the courtshave approvedof measures
that comparethe probabilityof a qualifiedminoritybeinghired
with that for the majoritygroup (Gastwirthand Greenhouse,
1987, p. 39). (By the way, if you fall at the extreme view of
rejectingany and all empiricaldata, honestyshouldcompelyou
to rejector ignoresuchdatawhenit supportsyourtheoryas much
as you wouldwhen the data run counterto your ideas.)
On the otherhand,most of us areunlikelyto acceptempirical
data automaticallywithout carefulevaluation.We know there
can be many problemswith our research,and it is particularly
difficultin the usual case when we work with nonexperimental
data (see Lieberson, 1985;Achen, 1986, chap. 1). Any decent
undergraduatemethods text is filled with technicalissues that
must be consideredlest a researchprojectgo awryand generate
meaninglessoutput. But there are more than technicalissues
involved, ratherthere is also a set of exogenous pressuresthat
affectourstudyandthe interpretationof the results.It is not at all
unreasonableto viewthe resultsof socialresearchas a dependent
variablethat is affectedby the realityof whatit is meantto study
and other various technical mattersthat keep the study from
being ideal and also the social context of the work itself. By the
last I meana widevarietyof social pressureson the researcher(s)
such as the dominanttheories and paradigmsin the area;the
predispositionof the researcher;the dispositionsof foundations
andotherfundingsources;the dispositionsofjournaleditorsand
reviewers;seniorcolleagues;the prestigefulin theprofession;and
so forth.1
Lieberson/ ASKINGTOOMUCH 381

If the resultsof socialresearchareto be viewedas a dependent


variablethat is affectedby more than the realityit is meant to
ascertain,it does not follow that the influence of extraneous
forces is of equal magnitudefor all researchproblems. Some
problems are intrinsicallyvery hard to get at, for example,
heredity-environmentissues well known to us with respect to
black-whitedifferencesin IQ tests. Others are relativelymore
manageable.Consider,for example,the serendipitousdiscovery
by Stephen D. Stitt (a professor of law at the University of
Florida) and S. Kay Isaly (a paralegalwith a Jacksonvillelaw
firm).Whilepreparingthe defensefor a clientin a Floridamurder
case in 1977,they learnedthat 94%of the 114 men on Florida's
deathrowin 1977hadkilledwhitevictims,whereas4%hadkilled
blacks, with the remaining2%being equal opportunitykillers,
havingboth whiteand blackvictims(Zeisel, 1981,p. 458). There
werestronggroundsfor concludingthat raceof victimaffecteda
murderer'slikelihoodof receivingthedeathpenalty.Sucha result
is consonantwith a long-standingnotion in race relationsthat
white-dominatedsociety values white life more highly. Hence
violations against whites were punishedmore severely(Zeisel,
1981,pp. 460-461).
We also have several other very careful analysesof related
questions:Radelet (1981, p. 918) found in Florida that "those
accusedof murderingwhites are more likely to be sentencedto
deaththanthoseaccusedof murderingblacks... dueprimarilyto
the higherprobabilityfor thoseaccusedof murderingwhitesto be
indicted for first degree murder."Race of defendantdid not
strongly affect the indictment or the probability of a death
penaltyonceraceof victimis controlled.Onthe sametopic, a very
thorough logit and log-linearanalysisof the data reportedby
Radelet(1981)appearsin Agrestiand Finlay(1986,pp. 486-502)
on the influenceof racialcharacteristicson the odds of a death
penalty being imposed after an individual is convicted of
homicide. These resultscontrastwith earlierperiodsfor which
thereis evidencethatblackcriminalswereat substantiallygreater
risk of receivingthe death penalty (see Radelet, 1981, p. 918).
Now I doubtif thesestudiesarethe finalstatementson this topic,
382 SOCIOLOGICAL
PERSPECTIVES/ OCTOBER1988

butthe qualityof the analysesis highandit shouldnot be hardfor


criminologiststo evaluate the results and pursue the relevant
data, if unansweredissuesstill remain.
Most of us would have no difficultyappreciatingthis sort of
research:Race of offendereitherdoes or does not influencethe
sentencingof someonefoundguilty;likewiseraceof victimeither
does or does not havesome impacton the sentence;and, in turn,
severityof sentencingeitherdoes or doesnot reflectaninteraction
betweenthe victim'sraceandthatof the killer.In similarfashion,
we can appreciateand understandresearchdealing with com-
plicated equal employmentcases. There are special statistical
problemswhen the minorityfractionof qualifiedapplicantsis
unknownand/or whenthe numberof hiresis smallrelativeto the
numberof applicantsthatis, in turn,a smallpartof the largepool
of qualifiedpersons(Gastwirthand Greenhouse,1987).We can
see how suchstatisticalissuesmustbe studiedwhena courtdeals
with charges of discriminationby the Universityof Texas at
Dallas againstwomen,or whenit has to decideif hiringpolicies
by the governmentof Fairfax County discriminatesagainst
blacks.
Under any circumstance,it is quite appropriatethat the
existing researchresults be challengedif more than one inter-
pretationis plausible.Sucha give-and-takeis not alwayspleasant
for thosewho are-as it were-"taking"it, butit is productivefor
the discipline. For it leads to more and betterprocedures,the
examinationof newerand differentdimensionsto the problem,
theoreticalreformulations,and so on. In evaluatingthe workwe
shouldstartwiththe assumptionthat the researcheris givinghis
or her most honest effort to interpretthe results,but there are
these additional forces operatingthat may distort the inter-
pretation.In otherwords,no matterwhat one may think of the
methodsused, the qualityof the data, indeedthe intelligenceof
the author(s),the resultsareto be dealtwithdirectlyandin terms
of the standardsof scholarshipin the discipline.If you ignorethe
study and/or its conclusionscontradictone's own assertions,it
cannotbe on the groundsthat you don'tlike the results.
Above and beyond the social and structuralpressuresthat I
Lieberson/ ASKINGTOOMUCH 383

mentionedearlier-which we sharewithmanyotherdisciplinesin
both the hard and soft sciences-our discipline also has an
additionalpressurethat leadsus to ignoresome researchand all
too quicklyandunthinkinglyembraceotherstudiesof less merit.
In a nutshell,thereare a set of views about contemporarysocial
and political issues that at least appearto be quite common to
sociologists.In turn,resultsthatclashwiththeseviewsaretreated
differentlythan eitherthose that are in harmonywith them or
deal with topics that have minimalrelevancefor contemporary
policydebates.Ourdisciplinehas a set of illogicalproceduresand
inappropriateways of thinkingabout resultsthat are implicitly
definedas "unacceptable." What arethey?

AD HOMINEMTHINKING
Letus startwithwhatappearsat firstglanceto be an extremely
bad example of anything.As we all know, South Africa has
massiveracialgapsin income.Accordingto governmentstatistics,
for 1983the averagemonthlyearnings(in SouthAfricanrand)in
variousindustrieswas R1,200for whites;R571for Asians;R404
for colored;and R300 for blacks-the last figurebeing exactly
one-fourthof the whiteaverage(Departmentof ForeignAffairs,
1986,p. 470).2Whatcausesthisdifference?Discriminationof one
sort or another?Not accordingto a studyof wagedifferencesin
1970reportedin an Official Yearbook(Departmentof Foreign
Affairs,1985,pp. 481-482).Whiteshadearnings5.2 timesgreater
than blacksin 1970;of this hugegap, accordingto the statistical
analysispresented,between87%and96%of thetotalblack-white
gap is due to nondiscriminatoryfactors. Controllingfor racial
differences in age (an approximation of experience), hours
worked, education, and occupationalstatus, it was found that
about 70%of the black-whitedifferentialcould be explainedas
not dueto discriminationbutto "factorsrelatingto productivity"
(Departmentof ForeignAffairs,1985,p. 481).Butthatis not all,
they proceedto take into accountthe fact that qualityof black
educationis lowerbecauseof the standardsexistingfor different
384 SOCIOLOGICAL
PERSPECTIVES/ OCTOBER1988

teachers,differencesin the student-teacherratio,andgapsin the


percapitamoneyspenton students.By the timetheyget through
it turnsout that 91.5%of the gap (?4.5) is not discriminatory.
Whatdo you think of this result?Let us assumethat the data
are not fudged, that is, that the gaps are not greaterthan the
SouthAfricangovernmentreports,or educationaldifferencesare
not smaller.If we are to questionthe results,it cannotbe on the
groundsthat we know why the governmentis providingus with
suchdata, we cannotsay that we rejectthe resultsbecauseSouth
Africa is a racist countrythat is tryingto rationalizeor justify
intenseformsof discrimination.We may wonderaboutthat and
it may alert us to questionthe data very closely, but we cannot
rejectthe resultson these grounds.How can I make such a flat
declarativestatement about a field that is so wide open with
respectto its paradigms?The answeris verysimple,to rejectany
dataset or intellectualargumenton thegroundsthatthepresenter
has certainmotivesor attitudesis to takean ad hominemposition
that has nothing to do with the validity of the argument.If a
naturalscientisttells me that grapefruitcausescancer,I do not
think about what makes the scientist look at grapefruitso
negatively(eventhoughtheremay be suchreasons)and I do not
say that the scientistis antigrapefruitas if that is the explanation
for the scientist'sresults.It is perfectlypossiblethat the scientist
hates his or her father and mother and they own a highly
mortgagedgrapefruitfarm. Or the scientistmight have always
found the taste of grapefruitto be repulsive.But it is not of
consequencebecauseconsumptionof a given amountof grape-
fruiteitherdoes or does not affectthe riskof cancer.And that is
all that matters, pure and simple. Ad hominem arguments,
defined in my dictionary (Merriam-Webster,1985, p. 56) as
"markedby an attackon an opponent'scharacterratherthanby
an answerto his contentions"areillogicalandirrelevant.Now in
the SouthAfricancase,the analysiscollapsesbecauseit is so bad;
it takes black characteristicsthat are themselvesproducts of
discrimination,for example, the occupation held, or level of
education,and sneaksthese in as controlvariables.But keep in
mindthe point about the dangerof ad hominemarguments.
Lieberson/ ASKINGTOOMUCH 385

Let me give you a tougher example of this horrible misstep in


our thinking. In 1974, Robert William Fogel and Stanley L.
Engerman (1974a, 1974b) published a two-volume study in
economic history, Time on the Cross. The entire second volume
was devoted almost exclusively to evidence and methods, and the
first volume-although very readable-included six tables and 46
figures. So their work was hardly speculative. Their reinter-
pretation of slavery generated considerable controversy. Among
their conclusions are the following: Slavery was highly profitable,
the system was going strong at the time of Civil War, slave
agriculture was efficient compared to the northern system of
family farming, and the demand for slaves was increasing. Here
are two that I want to specially emphasize:

The material(not psychological)conditionsof the lives of slaves


comparedfavorablywiththose of freeindustrialworkers.Thisis
not to say that they were good by modem standards.It merely
emphasizesthe hard lot of all workers,free or slave, duringthe
firsthalf of the nineteenthcentury....
Slaves wereexploitedin the sensethat part of the incomewhich
they producedwas expropriatedby their owners.However,the
rate of expropriationwas much lower than has generallybeen
presumed.Overthe courseof his lifetime,the typicalslavefield
handreceivedabout90 percentof the incomehe produced[Fogel
and Engerman,1974a,pp. 5-6].

Fogel and Engermanmention, in the book's epilogue, that they


had presented papers in many places on their research dealing
with the economics of slavery: "Sometimes, after the end of a
discussion one of our colleagues would come up to us and, with a
nervous smile ask "What are you guys trying to do? Sell slavery?"
(1974a, p. 258). Elsewhere they mention a conference on these
matters in Philadelphia in 1967that eventually lead to accusations
of racism and questions of whether the authors were arguing that
blacks were better off under slavery than freedom, that the
abolition of slavery was a mistake (1974b, pp. 11-19).
Are Fogel and Engerman correct in their assertions? Consult
an important critique by Paul A. David et al. (1976) and the
386 SOCIOLOGICAL
PERSPECTIVES/ OCTOBER1988

literaturethat has builtup on theseissues.But whateveryou do,


the motives of the authorare not the final groundson whichto
decideif theirdataareanalyzedcorrectlyor if the conclusionsare
correct. As a matter of fact, do not be so confident that the
motivesof the social scientistis all that obviousor suspect,given
the conclusion reported. The scientist who found a linkage
betweengrapefruitandcancermaywellhavebeena greatloverof
grapefruitand graduallygot interestedin studying the fruit.
Likewise,thereis no reasonto thinkthatFogelandEngermanare
anythingbut concernedsocial scientists(see, for example, the
motivestheygivefor attackingthethentraditionalinterpretation
of slavery,1974a,pp. 258-259).Unlessyou believethat data are
totally the productof the researcher'swishes-and we discussed
this earlier-then you have to allow that the results a social
scientistreportsis indeedthe best sensehe or she can makeof the
data.
Onemayask:Does mypositionon ad hominemargumentsrun
counterto a basicthrustin the sociologyof knowledge,whichnot
only posits certainsocial conditionsunderlyingknowledge,but
also has as its missionthe studyof theseconditions?Theansweris
that thereis not reallyany directcontradiction.The key is not to
confuse the mission of the sociology of knowledge with the
substantiveproblembeing studied.It is perfectlyappropriateat
any and all times to ask about the social underpinningsof
knowledge,but that is not the samething as askingwhetherthe
knowledgeis valid-at leastvalidunderthe broadcriteriaof what
the societyis ableto defineas "true"in its currentstateof affairs.
If there are certain commonly shared criteriafor determining
whetherconsumptionof a givenquantityof grapefruitincreases
the risk of cancer,then that questionis not addressedby asking
why the scientists are doing these studies. They are separate
questions;and unlessempiricalworkis ruledout entirely-each
addressesdistinctiveand separatequestions.
I think there is a good reason for our extensive use of ad
hominemarguments-and if you keep your eyes and ears open
you will discover that they permeate our thinking and the
thinking of studentsin the classroom.Here are two examples
Lieberson/ ASKINGTOOMUCH 387

fromthe excitinglife of StanleyLieberson.Someyearsago, when


I was on the facultyat the Universityof Washingtonin Seattle,I
wrote a paper on the military-industrialcomplex (Lieberson,
1971). As far as I could tell with the data that I gatheredand
analyzed, the notion was greatly overblown:A minority of
industriesstood to gain by increasedmilitaryspending,but most
gainedmoreif the sameamountof federalspendingwas devoted
to nonmilitaryspending. I was able to explain how increased
military spending could still occur even if the majority of
Americanindustrydid not standto gain. Afterthe publicationof
thatpaper,on morethan one occasionI receivedjokinginquiries
as to whetherBoeing-a companywithlargemilitarycontracts-
hadpaidme to do the study.Second,a fewweeksago I attendeda
brown-bagtalk on values in primetime commercialtelevision.
The speaker, in the course of his discussion, observed that
televisionproducersandwriterswerelargelyJewishandtheyhad
significantinfluenceon thevaluesembeddedin televisionoutput.
Nobody in the audienceresponded.Judgingfrom the speaker's
name,I wouldguessthat he wasJewish.I couldnot helpwonder
if the commentwouldhavebeenas readilyacceptedif the speaker
had been eitherblack or a non-Jewishwhite.
As I discussedearlier,the data are often not that clean and
clear, there is a rarelya definitiveresult, and hencethereis the
propensityto go to the extremeof assumingthat the outcomeis
really a reflectionof the investigator'sdisposition. This is the
well-known"Fallacyof the Consequent."(By "well-known"I
mean I just found out about it two days ago.) This is whereyou
arguefrom a consequentto its cause, for example,if someone
dislikeshot and humidweather,he or she will not wantto take a
summervacation in Florida. But if you do not take a summer
vacationin Florida,it does not follow that you dislikeheat and
humidity.To put the argumentin a differentform-a statistical
one-if the dependentvariablehas severalcommoncauses,you
cannot infer with much confidencewhich cause is operatingby
simply observingthe presenceof the dependentvariable.I will
have an excellentexampleof the "Fallacyof the Consequent"in
the next section;see if you fall into the trap.
388 SOCIOLOGICAL
PERSPECTIVES/ OCTOBER1988

A final example of how ad hominem thinking puts pressure on


us can be drawn from William Julius Wilson (1987, p. 19), who
appropriately-in my opinion-describes the "liberal"perspec-
tive on the underclass as timid. Earlier in writing about the
deterioration of ghetto life through the years, Wilson observed:

The only studythat providedat least an abstractsenseof how the


problemhad changeddownthroughthe yearswas the Moynihan
report on the Negro family, which presenteddecennialcensus
statisticson changingfamilystructureby race.
However,the controversysurroundingthe Moynihanreporthad
the effect of curtailingseriousresearchon minorityproblemsin
the innercityfor overa decade,as liberalscholarsshiedawayfrom
researchingbehaviorconstruedas unflatteringor stigmatizingto
particularracialminorities[1974,p. 4, italicsare mine].

DOUBLESTANDARDS
Several years ago an article was published in Scientific
Americanproposingto show that reductionsin the fundingof
various programs "for children, mothers of young children and
pregnantwomen"duringthe administrationof PresidentReagan
had "contributed significantly to the change of trend in the infant
mortalityrate"(Miller, 1985, p. 31). The change in trend was
basedon the fact that the rateof declinehad sloweddownin one
year, 1984,a dropof 2.7%in thatyearas contrastedwith4.6%per
yearin precedingyears.
Now such a conclusion is totally unwarranted.In an out-
standing article, which I recommend to you if you have not read
it, Donald T. Campbell(1985, in Tanur et al.) examined the
problems of measuring the effects of innovations by means of
time series analyses. He was concerned primarily with the
influence of various government policies on automobile-related
incidentsof death,accidents,and so on. He observes,"Itis clear
that the more unstablethe line is before the policy change or
treatmentpoint,the biggerthe differencehas to be to impressus
as a real effect" (p. 163). In other words, if we see that the
dependentvariablebouncesaroundfromyearto yearpriorto the
Lieberson/ ASKINGTOOMUCH 389

introductionof a new program,it is difficultto know whethera


shift shortly after the innovationis a function of the policy or
merelyreflectschangesthat would have otherwiseoccurred.
The existence of a double standard,such that studies that
supportour predispositionsareexaminedless criticallythan are
studies that run counter to them, is by no means uniquely a
problemin sociology or in that partof our workthat deals with
issues of policy and publiccontroversy.3But it is exacerbatedin
these areas because it is very difficult to stand up to a double
standardwithout again having your motives questioned.If a
study provides very weak evidence to support the claim that
groupx is beingdiscriminatedagainstin some arenaof life, it is
hardto criticizethe studywithoutgivingthe impressionthat you
do not thinkgroupx is beingdiscriminatedagainstevenif all you
wantto sayis thatthe datafail to showit in a convincingway(not
thatthereis anythinginherentlyimproperwithsucha conclusion,
anyway).Thisagaininvolvesthe irrelevantad hominemnatureof
such a conclusion.Moreover,the inferenceis not a logical one
sinceit entailsan incorrectconclusionaboutmymotivedueto the
Fallacyof the Consequent.
The exampleof the Fallacythat I had in mindfor this section
stemsfrom my commentsaboutReagan.If I tryto arguethatthe
ScientificAmericanarticle providedextremelyshaky evidence
about the effect of Reagan'spolicies on infant mortalityin the
United States, does this mean you should infer that I am "for"
Reagan?Yes if you fell for the Fallacy of the Consequent.If
Stanley is pro-Reagan,then he will defend him. From such a
statementit does not follow: If Stanley defendshim, therefore
Stanley is pro-Reagan. Actually I am not: I am pro good
thinking;I am pro doing good workin sociology.

UNDOABLES

There is a second reason for my mentioningthis Scientific


Americanarticleabout Reagan and infant mortality.And that
pertainsto the fact that our eagernessto addresspublic policy
issuescanleadto totallyunjustifiedstudiesandwaysof thinking.
390 SOCIOLOGICAL
PERSPECTIVES/ OCTOBER1988

Namely,we seek to do studiesthat are simplyundoablefor one


reason or another. Sometimes the data just are not there to
answerthe question.This is not easy to grasp;Reagan'spolicy
either did or did not affect infant mortality.But that does not
mean that it is an answerablequestionwith the informationat
hand (in this case, the ratefor a singleyear).To use a gambling
example:Suppose I have a hunch that the roulettewheel will
come up red on the next spin and it does. You can ask do I have
some powerto forecastthe wheel?You can ask that questionall
rightand therehas to be an answer,but you arein no positionto
drawa conclusionafterthe one spin-or at leastmakeit with any
confidence.Thisis not a cop-out;indeedit is a veryhelpfulwayof
thinkingaboutprogramevaluationsif questionsareraisedwhen
thereis not enoughtime and informationto reacha meaningful
conclusion.
I have discussedundoableselsewhere(Lieberson,1985, pp.
6-11)so I do not wantto belaborthe point hereanyfurtherother
thanto emphasizethatwe shoulduse ourknowledgeandskillsto
addresspublicpolicy issues,but we do harmto ourselvesandto
society when we pretend to have reasonablegrounds for a
meaningfulempiricalconclusionwhen the data are inadequate.
(If, on theoreticalgrounds,we havesomethingto say,thenthatis
appropriateas long as it so labeled.)A speciallyhelpfulcontri-
butioncan also be madeto societywhenwe areableto showhow
it is not possible to answer the empiricalquestion under the
existingconditionsor why it is the wrongempiricalquestion.4

IF RESULTSHAVE POTENTIAL
CONSEQUENCESTHAT ARE VIEWED
BY SOMEONEAS UNDESIRABLE,THEN
IT DOES NOT FOLLOWTHATTHE
RESULTSARE WRONG

Suppose we attemptto ascertainthe truth of the following


statement:If P, thenQ. ThestatementaboutP leadingto Q is not
in errorif we do not like Q; nor can we concludethat P does not
exist simply becausewe object to the consequencesof P being
Lieberson/ ASKINGTOOMUCH 391

true. This is a very obvious point, but I believe such faulty


reasoningis commonwhenthe issuehas consequencesfor public
policy. For example,WilliamJulius Wilson's(1978)book, The
Declining Significance of Race, was criticized in part because it
would be used asjustificationby the dominantwhitesocietyfor
changingsomepoliciestowardblacksor mightbe misunderstood
in such a light. It is perfectlyappropriateto criticizea scholarly
productbecauseof errorsin its results,procedures,theories,or
logic. But it is inappropriateto criticizea work becauseof what
mightbe done with the results.Doing so will producea perverse
form of logic: I do not like Q, thereforeI rejectthe truth of P
and/or the causal linkage between P and Q. (This is quite
differentfrom acceptingthe truth of the P - Q linkage, but
separately opposing Q on openly stated value reasons. But
distortingthe P - Q result simply in order to avoid Q runs
counter to the ideals of any truth-seekingdiscipline.)5Moore
(1970, p. 8) has stated the issue very well: "It is importantto
rememberthat in pointing out the political implicationsof an
argumentone has said nothingwhatsoeveraboutthe validityof
the argument.It is a commonpolemicaldeviceto discredita line
of argumentby demonstratingthat it leadsto unwantedpolitical
conclusions"(citedin Wilson, 1980,p. 21).
The second issue pertainsto one of the great functions of
sociological work, and that is the ability to present other
perspectivesto peoplewho otherwisehaveno contactwiththem.
Workingin race and ethnic relations,one becomessensitiveto
this very quickly since we often do not fully understandthe
perspectivesthatmembersof anothergrouphavetowarda social
issue. Sociologycannotsay whichperspectiveis "correct"but we
certainlycan play a valuable role in helping our studentsand
societyin generalrealizethat thereare differentperspectiveson
the same topic that we would not otherwiseunderstand.One
wouldthink this positionis hardlycontrovertible,but it is when
the investigator helps us understand a perspectivethat the
discipline appearsto disapproveof the point of view of anti-
abortionists,bigots, and so forth. Althoughperfectlyhappy to
understandtheviewpointsof personswe aresympathetictoward,
we often act as if someonepresentingus with the viewpointsof
392 PERSPECTIVES/ OCTOBER1988
SOCIOLOGICAL

others is necessarily someone who is sympathetic or inclined


in that direction-as opposed to providing us with basic
information.
This issue was suggested to me while reading an interview with
a Nobel Prize winning writer, Isaac Bashevis Singer, who was
asked whether he endorsed the views of the hero in one of his
novels. After Singer denies such an endorsement, the interviewer
presses him further, and Singer responds:

The tendencyto identifya protagonistwith a writeris a verysilly


one. Some peoplecameto the conclusionthat Dostoevskyreally
was a murderer.Give him freedomand he would kill a Russian
womanand hernieceand so on. It'sridiculous.The fact that you
know the way of thinkingof a murdererdoes not meanthat you
area murderer.It meansthatyou canprojectyourselfinto his way
of thinking.This is the very essenceof talent. But this suspicion
that the writeris alwaysthe heroof his book cando greatdamage
to literature[p. 151].

In the same way as Dostoevsky is not a murderer because he


wrote a novel about a murderer, likewise sociologists should not
be taken as endorsing a position simply because they help us
understand those who hold such positions. It is as important for
us to learn about perspectives that run counter to the dominant
ethos of the discipline as it to learn about perspectives that we
tend to support. It should not be the author's duty to assault the
subjects or counter their perspectives in ethnographic research.
Moreover, even if the scholar were to be personally supportive of,
say, right-to-life legislation, it should not be relevant to evaluating
the product as scholarly information.

POLICY RESEARCH HAS


CERTAIN DISTINCTIVE QUALITIES

Finally, if we are serious about doing researchthat has a public


policy orientation, it is necessary to consider some of the special
problems that occur in this type of research.This is not the time or
place for an entire essay on this matter, but I will briefly discuss
Lieberson/ ASKINGTOOMUCH 393

two of theseissuesin orderto openup at leastconsiderationof the


fact that policy-orientedresearchis not the same as "academic"
work.
(1) You andI wouldbe outragedif a drugcompanyintroduced
a new medicinewithout first consideringpossible side effects
causedby the drug.Not only do we wantto ascertainwhetherthe
drugis usefulin treatinga givendisorder,we also wantto knowif
the side effects are worse than the cure, that is, if the treatment
does moreharmthangood. At the veryleast,a patientshouldbe
forewarnedin orderto determineif the drug presentsa greater
riskthanhe or she is willingto take. Knowledgeaboutsideeffects
also helps us decide if the net effect of the new drug is an
improvementoveralternativetherapies.Thisis so importantthat
we have a governmentagency to ensure that drug companies
make the appropriate tests and report all of the relevant
information.
In policy-orientedresearch-for the very same reasons-it is
also necessaryto consider"sideeffects."But this requiressocial
researchof a sort that is radicallydifferentfrom the customary
focus on how one or more independentvariablesaltersa given
dependentvariable.In policy researchwe have to be concerned
not only with how certain causal variables might affect the
dependentvariableof policy interest,but also with the influence
changesin theindependentvariables(andthedependentvariable)
may have on otherattributesin the societyas well. Unlessthis is
done in policy research,we are just as irresponsibleas a drug
companythat neglectsto study a new medication'sside effects.
Yet one rarelyencountersdiscussionsof the possibleside effects
of givenpolicyproposals.If we knowenoughfromthe analysisof
a data set to determinethe social consequencesof a givenpolicy,
should we not also expect some understandingof the policy's
influence on other societal phenomena?Likewise,if we have
enoughtheoreticalknowledgeabouthow to altera givensocietal
condition,it is reasonableto expectinformationon sideeffectsas
well. Should the standardsfor social researchbe lower than the
standardswe hold for drugcompanies?
(2) In MakingIt Count(chap.4), I suggestedthatsocialevents
have lives of their own; so that if the originalcause of a given
394 SOCIOLOGICAL
PERSPECTIVES/ OCTOBER1988

dependentvariablewereto returnto an earlierstate,it wouldnot


necessarilylead to a similarshift in the dependentvariable.For
example,if the variable(s)that causedthe Women'sMovement
wereto returnto an earlierstate, it is unlikelythat the Women's
Movementwoulddisappear.If this provesto be the case-and I
suggested ways of testing that proposition-then the usual
researchapproachwill not be a satisfactoryway of addressing
policyissues.Normally,we askwhatindependentvariablesaffect
the observeddependentvariable;but this is of relevanceonly
whenthose linkagesarewhat I havecalled"reversible" (1985,p.
65 ff.). Butin situationswherethe associationwithX is totally or
largely irreversible,then policies based on how X caused the
currentlevel of Y will hold little help for understandinghow to
changeY sinceshiftsin X backto anearlierstatewillhavelittleor
no impacton Y. In the area of race and ethnicrelations-and I
assumein otherfields of sociologyas well-I am impressedwith
how criticalthis factoris. All too often it is assumed,in the most
cavalierfashion, that to understandthe causes of some of our
most seriousandpressingproblemsmeanssimultaneouslylearn-
ing how to removetheseproblems.This may or maynot be true,
but will be truenot merelyif the analysisof the causesof current
conditionsis correct,but also if such causal links are totally or
largelyreversible-a very separatequestion.

CONCLUSIONS
I likesociology.Indeed,if I coulddo it overagain,I would.Our
disciplinecan makeveryimportantcontributionsto socialissues
and social policy. But at one and the sametime I also think our
disciplineis asking too much, yet expectingtoo little. We are
asking too much of ourselvesbecausewe insist that our social
researchgeneratecertain results that are somehow defined as
appropriateandpoliticallycorrect.We have a set of illogicalbut
punishingconclusionsabout those whose results do not meet
certainpredeterminednotions. We also are a bit too prone to
think that any question asked can be answeredwith the infor-
Lieberson / ASKING TOO MUCH 395

mation that we are presentlycapableof obtaining.And we are


also too willingto allow sociallyand politicallyrelevantsubjects
to be studiedin the formin whichthe societystatesthe questions,
ratherthan by the way our knowledgetells us to approachthe
problem.Onthe otherhand,becauseof this overlyambitiousand
overlypoliticizedagendawhenwe dealwithsubjectspertainingto
national controversy and concern-whether that be gender,
abortion, poverty, power, education, the family, crime and
delinquency,raceand ethnicrelations,politicalinstitutions,and
the like-we do not expectenoughfromourscholarsin thewayof
intellectualintegritysuchthattheyapproachthedataas bestthey
can. We tend to accept or not accept a result because of its
possiblepoliticaland/or socialimplications.Butthatis irrelevant
to thejob at hand.Orif it is to be definedas relevantto thejob at
hand, then what does that mean our disciplineis? If we are a
lawyeror an advertisingagency,thenwe aretryingto compileall
of the evidencewe canto supportone positionor another.Thatis
understoodandwe wouldcertainlyfirean attorneywhodefended
us by indicatingthere are twelve reasons for thinking we are
innocentbut nine reasonsfor thinkingwe areguilty. We cannot
have our cake and eat it; asking ourselvesto present "good"
resultsbut suppress"bad"results,andact as if thosereportingthe
latterhavecertainsuspiciousmotivations.If we aretruth-seekers,
then the resultsarewhatthey are,pureand simple,like it or not.

NOTES

1. Here and elsewhere in the article we will not deal with intentional cheating with the
data in the sense of fudging about the actual numbers-as opposed to ignoring
contradictory results or otherwise analyzing the data with the intent of obtaining a certain
result.
2. I assume the gap would be even greater if agriculture was included, as well as data
for so-called "independent"African states.
3. It is my recollection that it was Peter Rossi who I first heard use the term double
standard in this way.
4. For an example of the latter see the discussion of the prematurequestion that occurs
when we do not yet have the underlying knowledge necessary to reach an answer to the
question that is posed (Lieberson, 1985, p. 9).
396 SOCIOLOGICAL
PERSPECTIVES/ OCTOBER1988

5. For a morecomplexstatementabouttheissuesinvolved,atleastas it is manifested


in criminology,see the reviewby Hirschi(1981).

REFERENCES

Achen,ChristopherH. 1986. TheStatisticalAnalysisof Quasi-Experiments. Berkeley:


Universityof CaliforniaPress.
Agresti,Alan and BarbaraFinlay. 1986. StatisticalMethodsfor the Social Sciences.
Seconded. San Francisco:Dellen.
Campbell,DonaldT. 1985."Measuringthe Effectsof Social Innovationsby Meansof
TimeSeries."In Statistics:A Guideto the Unknown.Seconded., editedbyJudithM.
Tanur,FrederickMosteller,WilliamH. Kruskal,RichardF. Link,RichardS. Pieters,
and GeraldR. Rising.Monterey,CA:Wadsworthand Brooks/Cole.
David, Paul A., HerbertG. Gutman,RichardSutch,PeterTemin,and GavinWright.
1976.ReckoningwithSlavery.New York:OxfordUniversityPress.
Departmentof Foreign Affairs. 1985. South Africa 1985 Official Yearbookof the
Republicof SouthAfrica.Eleventhed. CapeTown,South Africa:CTP.
1986. South Africa 1986, Official Yearbookof the Republicof South Africa.
Twelfthed. CapeTown,South Africa:CTP.
Fogel, Robert Williamand Stanley L. Engerman.1974a. Time on the Cross: The
Economicsof AmericanNegroSlavery.Boston:Little,Brown.
----1974b. Timeon the Cross:Evidenceand Methods.Boston:Little,Brown.
Gastwirth,JosephL. andSamuelW. Greenhouse.1987."Estimating a CommonRelative
Risk: Applicationin Equal Employment."Journal of the AmericanStatistical
Association82:3845.
Hirschi,Travis.1981."Reviewof Taboosin Criminology,editedby EdwardSagarin."
Crime& Delinquency27:560-563.
Lieberson, Stanley. 1971. "An EmpiricalStudy of Military-Industrial Linkages."
AmericanJournalof Sociology76:562-584.
1985. Making It Count: The Improvementof Social Researchand Theory.
Berkeley:Universityof CaliforniaPress.
Merriam-Webster. 1985.Webster'sNinthNew CollegiateDictionary.Springfield,MA:
Merriam-Webster.
Miller,C. Arden.1985."InfantMortalityin the U.S."ScientificAmerican253:31-37.
Moore, Barrington,Jr. 1970. Reflectionson the Causesof HumanMisery.Boston:
Beacon.
Radelet, MichaelL. 1981. "RacialCharacteristicsand the Impositionof the Death
Penalty."AmericanSociologicalReview46:918-927.
Singer,Isaac Bashevisand RichardBurgin.1986. Conversationswith Isaac Bashevis
Singer.New York:Farrar,Straus& Giroux.
Wilson,WilliamJulius.1978.TheDecliningSignificanceof Race:Blacksand Changing
AmericanInstitutions.Chicago:Universityof ChicagoPress.
---1980. "A Responseto MarrettandPettigrew."Contemporary Sociology9:21-24.
1987.TheTrulyDisadvantaged: TheInnerCity,the Underclass,
andPublicPolicy.
Chicago:Universityof ChicagoPress.
Lieberson/ ASKINGTOOMUCH 397

Zeisel,Hans. 1981."RaceBiasin the Administrationof the Death Penalty:the Florida


Experience."HarvardLaw Review95:456-468.

StanleyLiebersonis presentlyProfessorof Sociologyat HarvardUniversity,on


leave during the 1988-1989academicyearfrom the Universityof California,
Berkeley.He is currentlyengagedin a comparativeand historicalstudy of the
social conditionsaffectingthe namesgiven to children.His most recentbook is
From Many Strands:Ethnic and Racial Groups in ContemporaryAmerica
(coauthorMaryC. Waters).

Das könnte Ihnen auch gefallen