are able to recall than what does notncome readily to mind (the “availability”nheuristic), and they are far moreninfluenced by the features of thingsnthat resemble one another than thosenthat don’t (the “representativeness”nheuristic). Thus, students think thatninterracial violence is far more commonnthan it is because it is oftennreported in the media, it is vivid, and itnreadily sticks in one’s mind; whitenstudents usually think that most welfarenrecipients are black — when, innreality, most are white—because ofntheir profound ignorance of the differencenbetween proportions and absolutennumbers.nThe fact is, most students — and thenhypothetical man and woman on thenstreet — are neither good observers nornaccurate with respect to explanations.nThey often — and under certain circumstances,nusually — make erroneousnobservations. They walk around believingnthat certain entirely mistaken generalizationsnare true. Many do notnknow the difference between an assertionnand a fact. Or how to documentnan assertion or a generalization. Ornhow ideology may influence the observationsnthat they and others makenabout the social world. Many cannotnreach a reasonable conclusion evennafter they have been supplied withnevidence bearing on a given question.nThe problem is far more profoundnthan Professor Goldberg suggests: it isnmost people’s observations and explanationsnthat need training. We do notnhave to have the observational acuity ofnan Erving Goffrnan to know that, fornmost people, being able to observenevents accurately in the external worldnis a chancy business, one that would benimproved by making use of a systematicndiscipline such as sociology or socialnpsychology.n—Erich GoodenProfessor of SociologynState University of New YorknStony Brook, NYnMr. Goldberg Replies:nThe purpose of the “Gommon-SensenSociology Test” is to demonstrate thatnwhich cannot be demonstrated: thatnmany of the observations made bynstudents (and conclusions closely connectednto observations) are faulty andnthat these can be corrected by annacceptance of “opposite” observationsnmade by sociologists, “opposite” observationsnthat are usually concordantnwith an ideology that is shared by thenmajority of academic sociologists.nTo choose a typical example: then”correct” answer to the true/false question,n”The income gap between malenand female workers has narrowed innrecent years,” is: “[It] has widenednrather than narrowed; women hold fewnhigh-paying positions and the averagenworking white woman earns less thannthe average working black man.”nNow, the freshman student, who isnunfamiliar with the full meaning of thenterm “income gap,” clearly thinks thatnhe is being asked whether a womannperforming the same job as a man nownreceives a more-nearly equal incomenthan she formerly would have. Thenstudent responds that a woman does,nand this would have to be incorrect fornthe test to have its desired effect ofnmaking the student understand thatnone of his basic beliefs is incorrect. Thentest cannot do this, of course, becausenthe student is correct. That the samenforces that have made this true havenalso brought large numbers of womenninto the labor force at entry-level positionsn(thereby increasing the “incomengap”) is interesting, but not in the leastnsurprising to a student to whom isnexplained what the question and thisnterm literally mean.nA lack of specificity is often thenmethod that enables the test to seem tondemonstrate that the student is incorrect.nBut it is the tendentious and falsen”answer”—and not a mere and easilyncorrected lack of specificity—that isnthe deeper problem. That this is thencase is clear from the fact that neithernProfessor Goode nor anyone else can,nby increasing the specificity of thenquestions on the test, enable the questionsnto demonstrate successfully to thenstudent that his deeply held beliefs arenincorrect. Specificity serves to demonstratento the student that his beliefs arencorrect.nMy general point is that the observationsnpeople repeatedly make over timenare astonishingly correct and that thensociologist’s legitimate role is to explainnthat which is observed, not to pretendnthat the observation is incorrect.nProfessor Goode attempts to refutenthis in three ways. First, he points outnnnthat eyewitness accounts are notoriouslyninaccurate. This is certainly as correctnas it is irrelevant; I wrote not ofnobservations made midst gunfire andnmayhem, but ones made repeatedlynand over time. Second, in his discussionnof judgmental heuristics, ProfessornGoode gives only the examples ofninterracial violence and blacks on welfarento make his point. To be analogousnwith the basic beliefs considered by thentest we discuss, it would have to be thencase that students are incorrect in theirnbeliefs that more blacks attack whitesnthan vice versa and that disproportionatennumbers of blacks are on welfare.nThe students are not, of course, incorrect.nThat they are incorrect about thenabsolute figures relevant here hardlyncasts doubt on any basic, observationallynbased beliefs.nThird, Professor Goode offers alternativentest questions. The questionnabout social class and drinking strikesnme as unsurprising or surprising onlynin the illegitimate way in which thenquestions are presented on the Common-SensenSociology Test. Read literally,nthe question is sufficiently specific,nbut I doubt that many students —npicturing the cocktail party of the wellto-do—nare particularly surprised thatnas high a percentage of the highernclasses drink, if “drink” means “everntaking a drink.” I suspect that mostnstudents — who, after all, are freshmennon their first day of college — misreadnand interpreted the question as “membersnof the lower class drink more”n(i.e., have a higher per capita rate ofnconsumption and alcoholism). A colleaguenspecializing in the study ofnalcoholism tells me that this question isncomplicated methodologically, but thatna strong case can be rnade for a “false”nanswer.nProfessor Goode’s other two questionsnhave the virtue of specificity andnclarity, but they are not in the leastnsurprising — at least they were not tonhalf of my students nor to half ofnProfessor Goode’s (results that wenwould expect if the students have nonstrongly held beliefs about the issues).nIn contrast, the incorrect “answers” tonthe test we discussed were surprising tonnearly all of my students, as I’m surenthey would be to Professor Goode’s.nI would hope that we could make antest of questions that, on the first day ofnclass, only half the students answerednJULY 1991/5n