American Psychological Conditions Does Theory Research Progress? Burke - PDF document

American Psychological Conditions Does Theory Research Progress? Burke Marketing they persevere procedures until obtaining theory-predicted result. This that are and this hinders successful Prefatory Note These preliminary unintended interpretations usefulness of theories. some confine the term used here, includes also much more that express rules ofcorrespondence, associate a multiple operations. Research of op- but commonly in terms of theories. For example, "presentation reward contingent a response statement in of reward, contrast, "presentation of a of its of the corresponding operations in a include concepts. preceding statements-to reduce surplus meaning (implied eralization). A latter one, the apparatus, would have still less theoretical. In present terms, then, it generalizes beyond risk of two main of this are the by its title, is necessary process) of inconsistent. It will be potentially informative observations. is less researcher's goal of increasing 943-1 978)' A sleeper takes some apparent, in other words, a delayed-action a delayed persuasive the communication untrustworthy), has particular research is often referred effect was first on the basis of their research Army during World War (Reviews of of delayed persuasive studies can be found of a of findings special role quirk. An interesting attracts the student's attention. or writer in terms general principles, the student duly impressed go beyond very closely after it a place it did of several during the empirically troublesome and that basis was al. (1949 (1949 done in confidence of in the ability of British allies. Hovland et after it was shown, some opinion items showed significantly more improvement in regard for at an was given 5 days after the finding a Explanation and acceptance. of a Hovland group's laboratory attention was given experiments (Hovland four interpretations earlier been proposed two assumptions: with time, association of topic weakens association of In their influential Hovland, Janis, by Festinger this support interpretation prepared the which it social psychology texts published between 1955 was engaged test a cognitive response theoretical analysis of persuasion (eventually published of research on the In that Cook introduced a distinction a communication's after a effect (significantly slower of the impact of a more thorough review of GREENWALD, PRATKANIS, cue than cue or been any satisfactory evidence studies had relative sleeper 1973, for a similar assessment of early sleeper encountering Cook's evidence, Gillig Greenwald (1974) con- the (absolute) Greenwald findings, based on clear evidence ative sleeper none for Perhaps Gillig and Greenwald's experiments had not found a sleeper sleeper-effect experiments-were at technical development. Ronis, Baumgardner, Leippe, Cacioppo, also Baumgardner, Leippe, Ronis, Greenwald, 1983) reported initial series a laboratory construct a it possible persuasive messages' impacts within a session. Initial applications conducted in In four effect was absolute sleeper initial research is more detail as includes studies conducted at Ohio a 7-year period sleeper effect research being had also undertaken a search Their search on technical improvements and tual analysis described in Cook's 1971 paper. Gruder et (a) a a have] substantial initial impact on attitudes; (b) this change is totally inhibited discounting cue; cue and quickly enough so itself still study obtained a sleeper portantly, replicated it. implemented, one obtains a sleeper remained disturbing some of have satisfied conditions for a test it was initial tests (1985) did not a sleeper present authors' from most previous sleeper Gruder et distrust the a critical their procedure. studies with varying whether discounting cue infor- presented before or after Empirical resolution: discounting cue is critical. procedure, conducted between 1977 discounting-cue treatments. (These studies are reported marized in Figure variation made discounting cue information came multimessage, one-session, Figure 1 message, two-session with which produced a sleeper effect. Accordingly, in producing the sleeper added condition discounting cue conducted such a replication, adding to Gruder et the critical condition Gratifyingly, Pratkanis et other topics, and assessment is typically The plotted slopes, weighted the number with their sleeper effect instead preceded dependence of discounting cue such that (a) initial strong this countereffect the impact a decaying of persuasive an increasing impact. though, there one another and different rates. initial interpretation is assumed strongly affected cue, such (a) initial unit in analysis of Figure of separate does not Sleeper Effect effect research. textbook writers prema- turely adopted a single explanation (the to the alternatives, such effect was relative sleeper dissociation interpretation's overlook null results. Fourth, dependence of Gruder et (1978) replicated sleeper unusual procedure on presenting persuasive message) done at Ohio mation biased of being a single theoretical hypothesis, course of improve methods persuasion research, their inadequacy made clear (1978) discovery of reliable sleeper In retrospect, in the methods question itself had been asked. In particular, both the group been asking "Does question with dissociation theory. search has made it apparent have been "Under sleeper effect This latter question authors' research on Had this question that the been achieved the latest in the methodological developments sleeper effect (Cook, an analysis research questions. Research questions can theory-centered fashion and the research strategy discipline is testing theories. this ad- mirable strategy is easily misused produce nearly search conclusions. appreciate this, consider tactical choices smoothly enough theoretically predicted results not confirmed, proceeding (a) the theory being disconfirming results), the theory is who pre- in the conducting additional more data, revising procedures and then more data. interpretation of (The dis- does the GREENWALD, PRATKANIS, the disconfirmation dilemma. (Question disconfirmation dilemma; the resolution leading toward indicate its infrequent researcher repeatedly the disconfirmation dilemma retesting the than reporting the theory-centered rather than theory-testing.) is possible for in the researcher's faith in the theory be shaken data, it testing. Rather, strategy is theory-confirming researcher cedures until prediction-supporting several false resulting confir- in response initial disconfirmations. empirical comparison attempted, the is unlikely firmation's dependence conclusions from such research tried-and-abandoned procedures, those con- in the more guiding theory. Such conclusions constitute (Avoidably?) Overgeneralized Conclusions rarely provide the information whether a was engaged Greenwald's (1975) did report approach, such as produces a disconfirming (b) being reluctant findings for which it of hindsight) of research of psy- a monopoly on the chology; see Brush, 1974, pp. 340 examples from 1959, when Miller used a similar principle analyze persistence the impact of a sequence of two opposing persuasive suggested previously, in reaching collective fixation dissonance theory that one action more a small seemingly insufficient) certainly adequate) the report very active research controversy limiting conditions justification effect Insko, 1973; years between discovery establishment of limiting conditions concentrated research any other social psychological In retrospect, in establishing a of incentives insufficient justification problem as (dissonance, learning impression management); results; (c) because they did not easily fit these theoretical positions; social psychology was delayed limiting conditions of published 1933) was theory-the law which they contingent. For decades the effect was widely This situation against the null hypothesis." prepare reports journal editors' prejudices against publishing it seems that authors those prejudices in the intensely persevering research efforts of a learning results once credited a variety Buchwald, 1969; 1953; Post- 1963). After falling nevertheless persisted the domain so. Eventually, by pigeons, were have been organism can generate. more accurate of rewards they work laboratory studies demonstrated servation of aggression is likely These consistent behavior with frequency Public Health (1972; National Institute Mental Health, television is a significant aggressive behavior. Singer, 197 experiments did not findings because the laboratory studies typically included counterparts in ordinary laboratory studies accepted principle typically causes imitative conclusion from relevant research. 1984, for a clusion of, limited imitative of witnessing even very limited imitation of observed can be There are problem areas the status well-established conclusions. such areas, look especially (a) topics for which a strongly expected nonreplications makes the phenomenon Avoidable overgeneralization can occur among literature deal with as much as among their own for example, that in the authors at no take note empirical phenomena inconsistent with evidence inconsistent with it?) mission of have overgeneralized also likely overstated their in confirmation-biased fashion. Whenever a conclusion based stated, there be alterable conditions on which conclusion depends. as a conclusion, necessarily overgeneralized. inevitability of overgeneralization of describing, in a all possibly relevant variables inevitability of generalized conclusions. shown in case of of these (often arbitrarily) fixed levels such as randomization only a sense of being deliberately varied between experimental units (dependent complete description results of values of all focal a description of course, possible. Typically, the description to an observed configuration conclusion statement implicitly generalizes across all levels across contextual there must be conditions, among the infinity of unexamined contextual variations, aspect of scientific progress consists of limit known In the case of al. (1949) of limiting the communication with discounting mation (Hovland Weiss, 195 dition of a strong enough have residual time the discounting information suasive message al., 1985). Prospectively, it is certain can be further limited. (And, for reasons in the this accounting separately rec- with focal as misidentifications or construct variables (cf. GREENWALD, PRATKANIS, LEIPPE, AND an observed observed and dependent] variables has no dependence on or uncontrolled contextual variables is inevitably in of Research researchers' biases toward confirming their preferred theories. After all, social systems; their participants can be complementary and, therefore, mutually compensating for a discussion of theoretical competition counted on empirical resolution. of a there is empirical resolution, of psychic are advocates6 then, the great majority phenomena report that are established domains-are researched almost grinding theoretical axes lie detection, nontraditional procedures that are proposed tech- control, anxiety reduction, for which technique's small group of domains of research, greater career advancement, a new one who familiar idea may similarly allocate journal space the more left by This pattern researcher would theoretical predictions However, if a theory-confirming its conclusion has been Fate of Suggested in the process of researchers' strategies, published research have suggested remedies, which (a) using a alpha criterion than the customary "alpha inflation" of the same 1959); (b) standardly reporting measures because null-hypothesis rejection enough observations on reports just methods bias against publishing null results developing statistical drawing conclusions form other than using interval techniques (Bakan, 1966; Edwards, 1962; Greenwald, find occasional remedial techniques not much (even by remedial measures practices, faulty though present analysis identifies confirming function. previous critics remain unused they uniformly reduce prospects for retical predictions. (Never theory confirmations the alternative procedures should lectively avoid!) is a verdict. To use of that are that, obstruct scientific progress. Appeal of bias delays finding for beliefs presently held. long ago, answers. Although these answers they have virtue of a useful heuristic. is, even though they often work well (see The extent present social under the Theoretical opponents have well search. Instead, operated as detectives, purported demonstrations indeed, problematic only when an alternative might be be superior the re- superior alternative is available, confirmation bias for the available hypothesis. (Against this, course, it can better alternative confirmation bias predictions confirmed should increase a researcher's and support the of an confirm their predictions ones who sustain or expand their confirmation-biased researchers are thus more research activity unbiased ones, then it confirmation bias trait that selected (in the disciplinary environment research occurs. conjirmation bias hypotheses, confirmation bias a conservative tains a theoretical veteran researchers pass on specific theoretical confirmation-biased strategies their students, the theories the senior practitioners next generation (cf. remark: "Once a [theoretical] system . . as a sacred grating behind which each in order through its interstices" interstices" p. 3941). To all those in the a theory, there tangible rewards status and perpetuating the theory These interpretations (confirmation-biased) the- ory-confirming research strategies great survival to compete strategies, alternative similar self-sustaining potential; they must permit researchers succeed sufficiently sustain their partic- ipation in research in the researcher's perseverance testing theoretical predictions that produced disconfirming (1970) considers that ordinary scientific activity thrives on confirmation within the paradigm. Although notes the steady progress comes from this "normal such tradition-directed does not Anomalous results accumulate until major theoretical reorganization revolution) can commodate them. researchers' theory-confirming Kuhn's conception (as characterized Lakatos, 1970). scribed does achieve steady resources are consumed in the production avoidably overgeneralized Popper's falsification-seeking in principle that the much lip service who would falsification seeking subtly different theory confirming. Result-Centered Research Strategies Opposed although Kuhn maintain that theory testing is the central pole scientific activity revolves. In contrast, our theory should not test a theory become ego-involved theoretical prediction, lead even- avoidably overgeneralized this section research strategies (but not far from the center process. In as a radical should not routinely needlessly overgeneralized findings of an ex- one can many conditions confounded together 1959) contrast dogmatic (con- critical (disconfirmation firmation-seeking methods not science, example, the that researchers who set cation) road testing theoretical hypotheses remind the reader do not research. Rather, these two The necessity is discussed at the becomes a dogmatic one somewhere concluding section. GREENWALD, PRATKANIS, "Under what conditions does communication accompanied cue produce condition-seeking researcher de- liberately seeks conclusion (such persuasive message presented after effective message, use of generates a progression of research increasingly precise Under what conditions does Under what conditions does Under Condition when a research form, the answer must conditions under which it conclusions corresponds between statistical method can contrast with of condition-seeking, which dis- finding depends, that can produce unobtainable result. gineer a As with the method of seeking, research on the approach can be phrased what conditions . . what conditions adolescents resist of the significant reversals of familiar theory-based example, "Under conditions does viewing televised vi- subsequent tendencies "Under what conditions does child's liking for directly with of incentive effect, well obtain the sought result, design strategy design strategies already obtainable which a generality of increases (from a finding. Condition-Seeking Strategy that the strategy, even if label, is This comment occurred in two forms: (a) is a long-standing researchers whereby product of petition in a research not the goal of individual researchers agree with these observations, that contemporary therefore free of limited informal tradition favoring interaction-effect re- as indicated in social psychological the same discussion of the importance nonexistent in most detailed 1969, p. Perhaps the valuable kinds statistical interaction independent variables dependent variable. phenomenon varies as certain conditions usually helps to narrow the phenomenon. of condition-seeking discussion of of avoidable alizations, earlier this article, (1973, 1983) criticisms of conditions on (perhaps independently researchers) of a condition-seeking strategy. However, this informal widespread advocacy the method. collective condition seeking. many examples seeking emerging as the competitive theory-centered individual efforts. As effect was resolved by progressive, collec- tive discovery of qualifying Another problem competing groups of social psychologists was risky shift effect, individuals (Stoner, 1961; scientific progress in the of this risky shift controversy was eventually resolved by that produced the a con- servative shift (Brown, 1965; Myers from Elliot Aronson (May 17, Judson Mills set-size effect, finding of increase in of a defined target set, in the target set set-size effect is still subsided greatly upon establishment did not Schneider, 1977). between-groups versus condition seeking. insufficient justification, risky shift, set-size effects of the many cases in which active empirical disputes resolution upon of limiting for previously overgeneralized conclusions. make clear indeed, condition occur as the of a of researchers pursuing theory-centered of researchers suing theory-centered often confirmation-biased) taken as a great deal wasted effort in the inefficiency is largely that did which a effect fails generally available analysis might Existing Uses Design Strategy design strategy previously been its value. Consider, for Sikes, Blaney, desegregated schools undo the edu- the United versely affect oversight was, eralization of conclusions from Stephan (1978) his colleagues identified designed a classroom" (based White, Hood, use of equal-status inter- heterogeneous (majority plus minority) groups. (1953) creation remembered better which a influences a Orne's con- struction of exhibited phe- were previously believed exclusively associated of sensory only have but more need for method produces a progression of may suggest of increasingly precise with previously discovered. (Consider a hall of third order untested interactions can be be p. 1 191.) However, the suggestion that progress via condition seeking occurs in the form of increasingly trivial results is based one that entific progress. need for guide empirical along with is provided by new findings, method from being just a of cultivating In order increasing precision of findings theories must grow in their new findings. statement that can progressive discovery of limiting conditions history of research insufficient justification problem. Statement is worded analysis originally provided theory; it bulges with appended limiting conditions findings of of research. in terms discrepant statement, large) incentive offered for the statement person freely chooses to statement (Brehm Cohen, 1962) and to the the statement Insko, 1973) toward agreement statement occurs the extent statement implies the subject's of progress the form of with ever-shrinking is powerfully subsuming the insufficient justification integrates dissonance reduction with a GREENWALD, PRATKANIS, change (cf. Greenwald, 1980; Schlenker, 1982; is, for insufficient justification effect: It should not occur among includes traits that statements. (Nevertheless, dition seeking should than Statement McGuire's Analysis the Relation the relationship theory has (1973, 1983) writings. influence can most readily through some from McGuire's (1983) description the methodological he calls from McGuire's section entitled "Corrupting taken for some set stage management skills, re- sources, and later to find or in which researcher ingenious enough and out "right" and it is this san- the research program revealed the manuscript submitted to information obtainable emerges from the pattern in which none at theory (like is an distorted revresentation . . is each and each is dangerously false . . . . regards] empirical confrontation, to determine the hypothesis the hypothesis's in which condition-seeking and strategies overlap makes (1983, pp. ommendations primarily displaced from its as a central goal of Conclusion: The Role Theory in Research the Middle Ground Between philosophical analyses scientific method de- mand a skeptical regard toward theory: And, just it is impossible prove a theory true, so also is joint impossibility proving or dis- theories might appear such a skepticism is difficult maintain. The from experience is greatly verting experience as a basis for rests on an Fortunately, there an immense middle ground unjustifiable positions regarding theories as disprovable. And it evaluate theories within middle ground. directed although the essential containers scientific knowledge necessary vehicles containers and criteria such as storage capacity and that are appraisable without disproof, or most appropriate evaluating theories. work of rather than as and disproving them. the greatest researchers is until a perseverance is the context a theory-centered method, activity of theory testing defined activity results pre- theory. In produce a theory-predicted pattern, stress on procedure. consequence is researcher finds a procedure produce the desired relationship. when, as often the research product its absence, there solid gain the conditions on the empirical pattern the researcher describe, in a published the procedures reported prediction-confirming Such a finding is and the predicted it is difficult patterns, these strategies put stress on theory, along with increase its and thereby become increasingly observations just made can be recast in the func- of scientific research within evolving system of knowl- Alternative research strategies can understood in manner in which they variations and the for such selection pres- operate directly on theories. article of (and perhaps a chief defining scientific activity) theory is evaluated primarily indirect means empirical test. In empirical test, the conceptual elements a theory operational form, and WHAT CONDITIONS? this procedural proxy the survival test sus disconfirmation. principle, theory-centered methods operate the empirical crucial experimental (cf. Chamberlin, the researcher becomes an theory advocate, this a theory-confirming strategy (see Figure the philosophical impos- a theory theory come what theory-confirming researcher the rule-of-correspondence link and procedure that becomes exposed disconfirmation test, rather the theory confirming context, survival is governed litical selection criteria than not surprising, then, that replace a well-established theory. result-centered methods (such as condition research procedures provide the variations on which selection in advance. a procedural variant has been a theory does not produce the desired pattern, it it is a theory a function procedures. Importantly, centered researcher works with designs pinpoint critical procedural variations producing interaction-effect find- The supporting role that theory in result-centered meth- no means a one. Result-centered research such as producing based on stroying malignant not nonmalignant or creating computerized simulations that the selection pressures more rapid theory-centered methods. Obstruct Research identifying mark recommended condition-seeking a research question that begins with the question this article's conditions does theory Theory obstructs research progress testing theory the central of research, (as often happens) the faith in the correctness the theory than in the suit- the procedures test it. In other words, research progress when the theory may be in the face of prediction-disconfirming results. Re- many scientific disciplines encourage researchers theory advocates, conditions under occur commonly. since Francis Bacon, along with odologists of more recent eras, have been aware of the potential been urged toward greater objectivity Perhaps research eventually become years, methodologists hand. In the meantime, recommendations for research strategy should take account the reality when ego-involved-in centered methods recommend permit the researcher producing a desired pattern results rather confirming a particular whether this alternative a more one-can be as the heuristic that Anderson, C. (1983). Abstract 93- 108. (1982). Research on journals: Implications editors and authors. (1968). Dissonance theory: Progress and problems. In R. al. (Eds.), Theories ofcognitive Chicago: Rand Stephan, C., (1946). Forming impressions of Journal of Ab- Self-efficacy: Toward a unifying theory Social learning H., Leippe, reliable persuasion and Social R. C. (1975). Learning, motivation, cognition. In of learning NJ: Erlbaum. York: Wiley. (1964). Interference in G. (1974). Should history of science be of "right" and on subsequent behavior: A Psychological Review, (1973). Attitude and attitude of incentive, choice, Personality and Campbell, D. (1986). Science's social of validity-enhancing collective belief and the social sciences. In Pluralisms and subjectivities in social sciences. in preparation. M., Ellsworth, in social psychology research on the American Psychologist, C. (1965). method of multiple Science, 148, published 1890) Journal of discounting cue hypothesis and Unpublished manuscript, Northwestern R. (1979). pitfalls in accepting 1954 Social Science Statement segregation: A (1955). Construct validity in Fazio, R. (1980). Expectancy confirmation interaction sequence. American Psychologist, (1962). Dissonance Journal of Personality, for psychological Psychological Review, R. D. and aggression. (1955). Social processes. In 6, pp. 16). Stanford, A theory ofcognitive University Press. Festinger, L., Cognitive consequences and Social prophecy fails. of Minnesota Feynman, R. Surely you're D. (1982). American in the television violence Bulletin, 96, (1983). School desegregation: science role. it time the sleeper and Social D. A. null hypothesis and investigating theoretical models. Psychological Review, (1975). Consequences prejudice against (1976). Within-subjects of personal dissonance: Case study the conceptual Personality and Social Gruder, C. L., Cook, (1978). Empirical the absolute sleeper dicted from the discounting cue hypothesis. (3rd ed.). Hovland, C. I., Janis, I. H. H. (1953). Princeton, NJ: Princeton University Press. source credibility I. L., incentive conditions in modifying attitudes. and Social "Reinstatement" of the com- in delayed opinion change. and Social in hypothesis-testing Chicago: Uni- Center for ed.). Chi- cago: University Criticism and the (pp. 9 196). London: University Press. University Press. (1942). Mechanization effect of (1957). Primacy-recency in impression formation. C. I. et al. (Eds.), in persuasion (1925). The psychology of (1977). Publication prejudices: experimental study of confirmatory in the Therapy and The nature attitudes and change. In Aronson (Eds.), MA: Addison-Wesley. progress in social (1983). A contextualist theory of Its impli- innovation and reform in psychological Academic Press. (1985). Attitudes and attitude change. In social psychology New York: Random House. (1967). Theory-testing in A meth- Philosophy of Science, K. (1948). self-fulfilling prophecy. Antioch Review, Television and social behavior: New York: (1959). Recency primacy in as a timing of Journal of and Social E. B., M. (1976). Consulting and Clinical 92- 103. Norms and counter-norms in moon scientists: the ambivalence of scientists. American Sociological Review, Social influence and social change. New York: Ac- Doherty, M. simulated research environment: experimental study of (1975). The of group National Institute Mental Health. (1982). and implicationsfor the eighties. U.S. Government Printing inference: Strategies NJ: Prentice-Hall. London-New York-Leuven: Press-Leuven University Press. R. (1953). lications Universitaires. Demand characteristics and Academic Press. experiment: Antisocial and hypnosis. K. E. factors in the sensory deprivation "panic button." and Social R. (1964). Strong inference. published 1934) (1963). Rewards and punishments human learning. Postman (Ed.), R. (1981). A hunt effect in published master's thesis, heuristic and sche- Unpublished doctoral dissertation, reliable persuasion dead. Long Manuscript submitted Ronis, D. L. 980). Repetition agreement with opposing arguments: L., Baumgardner, H., Leippe, R., Cacioppo, search of reliable persuasion effects: I. computer-controlled procedure for studying persuasion. Journal of and Social On eliminating apprehension from Journal of Personality and Rosenthal, R. (1966). behavioral research. Rosenthal, R., L. (1968). Ross, L., Lepper, M. Perseverance in self- perception and processes in debriefing paradigm. and Social (1981). The personal history. Journal of Personality and Social R. (1982). Translating actions analytic approach to the L. Berkowitz 15, pp. Academic Press. (1966). Data-dredging procedures in survey J., Hood, and cooperation: Robbers Cave (1977). Controlled automatic human Perceptual learning, automatic attending, Snyder, M., (1978). Reconstructing the past: made in Publication decisions effects on inferences drawn from tests Journal of American Statistical Association, character recognition: Some idence from reaction time measurements. comparison of and group decisions involving risk. Unpublished master's thesis, Massachusetts Institute Self-verification processes: sustain our (1983). Self-definition and Greenwald (Eds.), 2, pp. theory of action of the connection upon it. Purposive behuvior in animals Public Health Service. Television and Government Printing policy in social sciences. American Statistician, the failure hypotheses in Quarterly Journul opinion change. and the fudge factor. cognitive dis- Hillsdale, NJ: Erlbaum. (1951). On stereotype or Revision received