Академический Документы
Профессиональный Документы
Культура Документы
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 support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms
Wiley, Midwest Sociological Society are collaborating with JSTOR to digitize, preserve and extend access
to The Sociological Quarterly
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms
THE EMERGENCE OF
BANDWAGON EFFECTS:
A THEORY
Richard L. Henshel
William Johnston*
University of Western Ontario
Since the inception of scientific polling over 40 years ago, serious questions have been
raised about the possible influences of poll forecasts on election outcomes. The concept
of a bandwagon effect from a pre-election poll, in which a prediction of victory for one
candidate or party leads to a higher vote for that candidate or party than would have
occurred otherwise, dates back at least to the 1940s, but this history is fraught with
difficulties. A principal difficulty is that studies of this phenomenon have produced
contradictory findings: some researchers detect a bandwagon effect and others do not.'
This undesirable situation has been recognized for some time (Dizney and Roskins
1962, p. 108). It is a basic thesis of this article that these difficulties have appeared in
part because of conceptual inadequacies, which are reflected in empirical research.
Therefore, we intend to retrace and reconceptualize in a fundamental way the linkages
involved in bandwagon phenomena. When the research is reexamined in the light of
improved understanding, the existence of bandwagon effects in some elections appears
unambiguous. Further, the strength of the effects due to polling may be increasing over
time, because of mechanisms that will be set forth below.
*Direct all correspondence to: William A. Johnston, Department of Sociology, University of Western
Ontario, London, Ontario N6A 5C2, Canada.
The Sociological Quarterly, Volume 28, No. 4, pages 493-511.
Copyright ? 1987 by JAI Press, Inc.
All rights of reproduction in any form reserved.
ISSN: 0038-0253
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms
494 THE SOCIOLOGICAL QUARTERLY Vol. 28/No. 4/1987
The objectives of this article are as follows: (1) to outline an alternative set of mech
nisms by which polls may create election bandwagons; (2) to resolve apparently con
dictory research findings; (3) to explicate feedback mechanisms through which polls
successive elections may become increasingly potent; and (4) to indicate by the fore
going the possibility of major political ramifications of polling. In addition, we will s
that some of the apparent extreme irrationality of election bandwagon behavior ha
rational, calculating basis.
A bandwagon effect in an election occurs when the predicted winner in an election poll
gains additional votes as a result of the publication of the poll. Conversely, an underdog
effect occurs when the predicted loser in an election poll gains additional votes as a
result of the publication of the poll.2 It is essential to the definitions that candidates gain
or lose votes not merely after a poll but because of the poll. This causal relationship can
be subdivided into effects caused by (1) voters switching their votes, the principal
concern with respect to pre-election polls, and (2) alterations in voter turnout, the
principal concern in election-day exit polls. In this article we will concentrate exclusively
on pre-election polls and their vote-switching potential.
In this article, bandwagon and underdog effects3 will be considered for (1) any con-
scious decision, (2) about specific, reasonably well-defined alternatives (to include per-
sons, parties, or concrete choices), (3) spelled out in advance of a specific date on which
the decision will be made, (4) by means of some majority or plurality principle, and (5)
about which some forecast of outcome has been communicated to voters. This delinea-
tion of scope deliberately includes not only secret ballot/mass elections for representa-
tives, but also conditions of open, direct voting on issues or voting in small face-to-face
bodies such as parliaments or unions. On the other hand, the conception retains sufficient
boundaries to insure that we continue to deal with a set of relatively well-delineated
phenomena.
The requirements that at least moderate specificity must exist in advance about the
possible choices, the decision procedure, and the timing of the decision ensures that the
voting under examination will be consciously recognized by those involved as specific
decision events about which predictions can be made. This criterion distinguishes the
present concern from the conception of "bandwagon" sometimes employed in collective
behavior models such as Granovetter's (1978, esp. p. 1424) or Schelling's (1978, esp. pp.
90-101). Requirement (5) demands further that an overall prognostication of the out-
come be disseminated in advance, thereby excluding exhortations or hegemonic influen-
ces of a powerful mass media (Noelle-Neumann 1984).
The most significant demarcation is expressed in components (3) and (5), which
require a concentration on the expected numbers of others rather than their actual
numbers, such as a poll of current opinion can provide. This future orientation rules out
those conformity studies (e.g., May and Martin 1975) in which information on the
distribution of present opinion is the key experimental variable. Our intent here is not to
imply that certain uses of the term bandwagon are improper, but to prevent the inap-
propriate extrapolation of research findings concerning conformity to public opinion into
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms
The Emergence of Bandwagon Effects: A Theory 495
(a)
(individual
voters)
(b)
alterations I
in financial +
contributions
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms
496 THE SOCIOLOGICAL QUARTERLY Vol. 28/No. 4/1987
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms
The Emergence of Bandwagon Effects: A Theory 497
CLASSIFICATION OF CAUSES
AND RESOLUTION OF CONTRADICTORY RESEARCH FINDINGS
We will now set forth the various possible causes of election forecast effects with the
intent of determining the relative predominance of bandwagon and underdog effects.
The causes of election forecast effects can be divided into two principal types, direct and
indirect; the distinguishing characteristic is whether the cause operates immediately
(directly) upon the individual voters, or through the decisions of certain key agents,
whose behaviors then influence the voters.
We have already discussed indirect causes. When we turn to direct causes, we note
that three types have been described in existing literature (four, if coalition theories are
counted); an additional type is set forth here for the first time. Because the first three
types have been addressed by others, their treatment here is brief. The first two of these
types are derived from the psychological literature on conformity, as adapted to band-
wagon phenomena by Fraser (1971).
Normative and informational causes are more or less terms for existing interpretations
of the bandwagon effect, and are approximated roughly as "conformity."
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms
498 THE SOCIOLOGICAL QUARTERLY Vol. 28/No. 4/1987
When we combine the three indirect causes with the four direct
clear that only a minority of the possible causes of forecast effect
explicitly. Some of the most interesting and potentially significan
disregarded thus far.
Now we can also offer a resolution for the conflicting findings r
wagon effect (see note 1). We begin by observing a perfect corresp
the studies surveyed, between finding and type of research: no
experiments finds a bandwagon effect (Cantril 1980; Ceci and K
Roshens 1962; Fleitas 1971; Gaskell 1974; Laponce 1966; Navazi
Kaplowitz 1977), while all of the field studies and theoretical treatm
theoretically indicate a bandwagon effect (Beniger 1976; Brams and
Gartner 1976; Grush 1980; Straffen 1977; Zech 1975).9 The crit
laboratory experiments seems to be the omission of indirect effects.
experiments on the bandwagon, which concentrate exclusively on co
simulate f, v, or e decisions among potential contributors. Thus
ignores the potentially powerful alterations on the flow of such cont
ble victory forecasts can create, alterations that indirectly affect
Meanwhile, field research on the bandwagon, which studies actual
inadvertently incorporates these indirect causes, albeit without rec
therefore does generally detect a bandwagon effect. The conflictin
note I arise in this fashion, and can now be understood and reconcile
that bandwagon effects do exist in actual elections.
Regarding the seven listed causes of election forecast effects-four direct and three
indirect-we must ask whether they produce forecast effects that are all in the same
direction (all bandwagon or all underdog), or whether they interfere with and nullify one
another. As Catherine Marsh (1985) pointed out, nullifying would lead us to underesti-
mate the intrinsic importance of election forecast effects.'0
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms
The Emergence of Bandwagon Effects: A Theory 499
o Ideological Personal
L Contribution Conution
o0 Contribution
0
o
t4,
0 50% 100%/
Predicted vote for contributor's
preferred candidate
We conclude that (1) informational and normative causes can produce only band-
wagon effects; (2) sympathy/empathy can produce only underdog effects; (3) calculation
can produce either effect but is most unusual in mass elections; and (4) financial assis-
tance, volunteerism, and endorsement (f, v, and e) intrinsically can produce either a
bandwagon or an underdog effect, but are very likely to have a net bandwagon effect.
The last of these conclusions requires explanation. We derived this conclusion by
noting that contribution of f, v, or e is activated in part by expectation of personal or
group gain in reciprocation for costly actions, and in part by ideologically based objec-
tives. The ideologically based acts probably peak in close races, as illustrated in Figure 2,
implying more or less equal likelihood for bandwagon or underdog effects. We reach a
very different conclusion, however, regarding f, v, or e offered for tangible benefits to the
contributor: it continues at a high level when one's candidate is far ahead and diminishes
when one's candidate is thought to be behind. (For the contrast in effects see the
illustrative curves in Figure 2.) These theoretical conclusions are supported by empirical
research." Thus, although f, v, or e have underdog components, the predominant (or
"net") effect should be bandwagon. The overall conclusion is that the principal direction
of votes is toward bandwagon effects, and the potential for bandwagon effects is far
higher than for underdog effects. This finding is particularly clear in mass elections,
where "calculation," as defined above, intrudes infrequently.
To this point we have analyzed and reconceptualized the bandwagon effect in a new and
(we hope) more satisfactory form, but pre-election polling seems to have further implica-
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms
500 THE SOCIOLOGICAL QUARTERLY Vol. 28/No. 4/1987
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms
The Emergence of Bandwagon Effects: A Theory 501
Shifts in f, v.
+ ond e in
+ direction
forecast
Poll
SShifts in voter
intentions in
direction
forecost
Shifts in election
outcomes in
direction forecst
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms
502 THE SOCIOLOGICAL QUARTERLY Vol. 28/No. 4/1987
100%
Realized Accuracy
t-
0
o Intrinsic Accurocy
threshold
I I I I I
Obviously, constraints must exist on the positive feedback; otherwise, all polls eventu-
ally would possess absolute credibility, and bandwagon effects would overwhelm all
other electoral influences. Positive feedback, however, does not necessarily imply
"explosive" deviation (Blalock 1969); convergent positive feedback with damped
incremental deviation, as illustrated in Figure 4, seems the likely possibility in the
present case.'7
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms
The Emergence of Bandwagon Effects: A Theory 503
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms
504 THE SOCIOLOGICAL QUARTERLY Vol. 28/No. 4/1987
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms
The Emergence of Bandwagon Effects: A Theory 505
Such simulations and existing trend data are supportive but hardl
focused testing scheme is required. In a direct empirical investigati
feedback loop, we encounter the well-known difficulties of studyin
tems. Because of these difficulties, a brief discussion of two possib
First, credibility feedback can be created by laboratory experiment
sive polls with rigged intrinsic accuracy, for a series of small body
important arguments in favor of an experimental approach to issue
stake here, see Brody and Brownstein [1975], especially pp. 214-
series of relevant phenomena, gradually developed by operationalizi
and predictive accuracy, would in principle offer an alternative, rigor
however, such a newly conceived time series could not provide a co
A more practical alternative is the retrospective construction of a tim
elections, using available proxies for the unmeasured data. For exam
real expenditures on private polls by political parties provides a rea
unmeasured poll credibility.
It may appear premature to spell out theoretical and political implications of the theory
in advance of direct empirical confirmation. Yet because considerable research is
required, it is highly desirable to list the issues whose resolution hinges upon the research
results. These issues may be divided imperfectly, but usefully, into the political and the
theoretical.
Political Issues
Ever since the inception of scientific polling, questions have been raised about the
potential for the polling institution to sway elections. Indeed, the first congressional
investigation of this issue took place only a few years after the first scientific sample poll.
The issue has been debated ever since, but apparently it has concentrated on only one
part of the set of possible mechanisms, with predictably equivocal results.
The present theory answers the influence question with a strong affirmative; it also
indicates a new path by which such influence can occur. The improved understanding
permits resolution of the contradictory research findings on poll effects. Further, confir-
mation of a bandwagon feedback loop would entail the conclusion not only that polls
influence elections but that this influence is growing over time.
Theoretical Issues
The theory provides an entirely new dynamic-the three indirect causes-to explain
how bandwagon effects may emerge. The theory subsumes not only formal polls but
also any poll-like predictors, such as the outcomes of early primaries, within a common
framework.25
The notion of indirect causes implies that bandwagon phenomena may be quite
rational-that is, generated by rational, calculating considerations. This idea is a funda-
mental departure from the traditional view of bandwagons as representing essentially
herdlike irrationality. If one argues that the electorate's voting response to shifts in f, v,
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms
506 THE SOCIOLOGICAL QUARTERLY Vol. 28/No. 4/1987
ACKNOWLEDGMENTS
Thanks are expressed to Hayward Alker, Fred Evers, Edward Grabb, John Kunkel, and
Benjamin Singer for helpful suggestions. We benefited a great deal from being able to
see a draft paper by Catherine Marsh. Earlier versions of our article were presented to
the XI World Congress of the International Sociological Association, New Delhi (1986),
and the Seventh International Congress of Cybernetics and Systems, London (1987).
NOTES
1. Beniger (1976), Grush (1980), and Straffen (1977) find some empirical evidence for an
election bandwagon effect. Brams and Garriga-Pico (1975), Gartner (1976), Straffen (1977),
and Zech (1975) provide theoretical support for the existence of such an effect. But Cantril
(1980), Dizney and Roshens (1962), Navazio (1977) and Tyson and Kaplowitz (1977) find no
effect. Ceci and Kain (1982), Fleitas (1971), Gaskell (1974), and Laponce (1966) find an
underdog effect. No attempt has been made to provide a definitive survey of findings.
2. For classical treatments see Merton (1948, and to a lesser degree, 1936), Simon (1954),
and Baumol (1957).
3. The term effect commonly used to describe bandwagon and underdog phenomena, might
imply some alteration in outcome, but in situations in which a bandwagon effect and an under-
dog effect exist simultaneously from the same poll, there may be no overall visible disturbance in
the proportional distribution of votes. In such instances, however, we might detect widespread
switching within different blocs of voters (something not visible at the aggregate level) or
increases in the total voting population (as bandwagon and underdog effects both work to draw
out additional voters). In light of the possibilities of countervailing forces, which produce little if
any overall change in proportions, it might be preferable to speak of bandwagon and underdog
"tendencies."
4. Occasionally, the belief that a candidate is a shoo-in could also have the effect of reducing
campaign support, thus leaving open the possibility of an underdog effect. However, this outcome
seems less likely because, as will be seen, there is a second reason people become contributors:
not only to produce victory but to produce gratitude for victory.
5. In this article, the meaning of the terms direct and indirect differs from that in two earlier
works. First, for Mendelsohn and Crespi (1970), "direct" effects refer to any influence of poll
forecasts on election outcomes. They conceive of these effects only in terms of conformity and
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms
The Emergence of Bandwagon Effects: A Theory 507
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms
508 THE SOCIOLOGICAL QUARTERLY Vol. 28/No. 4/1987
outside the technical polling community are not motivated to prefer percentage erro
for credibility, because people who read election polls are interested mainly in learnin
win the election. In short, the nontechnical audience should be expected to judge the
polling by its won-lost record, because it is concerned with that record. In operatio
what is relevant for credibility is that in the last election a poll called (for example)
races correctly, not that its MAPE was (for example) 1.4.
16. Baumol (1954) proposed another form of election poll feedback loop in which t
of multiple election polls in a single campaign were examined. In this case, the result
initial poll influence the public's preference distribution, which is reflected in turn
poll, which influences in turn.... We suspect that this sort of feedback occurs, although
not in the elegant manner foreseen, which ignores "random shocks" produced by the
themselves.
17. In the latter case, deviations of the loop variables from their initial values inc
progressively smaller amounts on successive iterations. When difference equations ar
the linearity of the loop relationships is assumed, the test of convergence (or "st
whether the product of the slopes empirically exceeds unity (Blalock 1969, pp. 81-82
18. The efficacy of political advertising can be challenged under certain condition
other situations it is clearly vital for campaigns, as are expenditures for other purpo
campaign (Grush 1980).
19. The significance of funding as an indirect cause might appear to be threatened
tendency of some interest groups to donate to both sides in order to maintain polit
regardless of who wins. Studies have shown, however, that the amounts given to
candidates are usually quite unequal. Thus the amount given to both sides can be rega
access base and can be ignored for present purposes. We must ask instead why poten
tributors decide to provide additional funding to their preferred candidate, how th
varies with the outcome predicted, and how much this additional amount helps the ca
20. Empirical confirmation of indirect causes does not exhaust important question
subject. As the predicted gap or margin widens between candidates or parties, is there
less monotonic change in the level off, v, and e (as selfish motivation would anticipate),
change curvilinear (as ideological motivation suggests)? Are f, v, and e affected in
ways? Do the degree and type of reaction vary across major sectors of the population
21. The initial model linked poll accuracy, poll credibility, and self-altering ten
three bivariate linear equations, one of which uses a first-order difference equation t
system. Modified systems were also explored, first, by assuming credibility linked to
racy by a distributed lag function, and, second, by assuming credibility linked to self
tendency by a power function. All four systems were studied through Monte Carlo simu
22. Although certain counterintuitive parametric values generated "explosive" curv
vertical asymptotes, results for all four systems overwhelmingly produced curves with t
configuration of "realized accuracy" in Figure 4. That is, they displayed damped inc
deviation and ultimately converged to horizontal asymptotes.
23. The brevity of this treatment is not intended to minimize the special difficulties p
empirical analysis of feedback systems. In the present case, however, a considerab
pendable time lag between the alteration in poll credibility after one election and th
bandwagon effects in the next election permits the use of discrete time procedures, s
difference equations. Provided tests for autocorrelation are negative, the bandwagon
system can be treated as quasi-recursive, with a lagged variable treated as "exogen
approach considerably simplifies the (still difficult) tasks of estimation and testin
analysis of the system's long-term stability.
24. Mendelsohn (1966, p. 219) provides an illustration of research on poll cre
Although his data pertain to election night projections, they illustrate the operation
suggested.
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms
The Emergence of Bandwagon Effects: A Theory 509
REFERENCES
Alexander, H.E. 1976. Financing the 1972 Election. Lexington, MA: Lexington Books.
. 1979. Financing the 1976 Election. Washington, DC: Congressional Quarterly.
1983. Financing the 1980 Election. Lexington, MA: Lexington Books.
Ashford, K. 1986. "The Role of Corporations in the 1980 U.S. Congressional Elections." Socio-
logical Inquiry 56: 409-432.
Baumol, W.J. 1957. "Interactions Between Successive Polling Results and Voting Intention."
Public Opinion Quarterly 21: 3 18-323.
Beninger, J.R. 1976. "Winning the Presidential Nomination." Public Opinion Quarterly 40: 22-38.
Blalock, H. 1969. Theory Construction. Englewood Cliffs, NJ: Prentice-Hall.
Brams, S., and J. Garriga-Pico. 1975. "Bandwagons in Coalition Formation." American Behav-
ioral Scientist 18: 472-496.
Brody, R.A., and C. N. Brownstein. 1975. "Experimentation and Simulation." Pp. 211-263 in
Handbook of Political Science, vol. 7, edited by F.I. Greenstein and N.W. Polsby. Reading,
MA: Addison-Wesley.
Buck, R.C. 1963. "Reflexive Predictions." Philosophy of Science 30: 359-374.
Cantril, A.H. 1980. Polling on the Issues Cabin John, MD: Seven Locks Press.
Ceci, SJ., and E.L. Kain. 1982. "Jumping on the Bandwagon with the Underdog." Public Opinion
Quarterly 46: 228-242.
Cronkhite, G., and J.R. Liska. 1980. "The Judgment of Communicant Credibility." Pp. 101 -139
in Persuasion New Directions in Theory and Research, edited by M. Roloff and G.R. Miller.
Beverly Hills, CA: Sage.
Dizney, H.F., and R.W. Roskens. 1962. "An Investigation of the Bandwagon Effect in a College
Straw Election." Journal of Educational Sociology 36: 108-114.
Field, M. 1983. "Political Opinion Polling in the United States of America." Pp. 198-222 in
Political Opinion Polling, edited by R. Worcester. London: Macmillan.
Fleitas, D.W. 1971. "Bandwagon and Underdog Effects in Minimal Information Elections."
American Political Science Review 65: 434-438.
Fraser, C. 1971. "Group Risk-Taking and Group Polarization." European Journal of Social
Psychology 1: 493-5 10.
Gartner, M. 1976. "Endogenous Bandwagon and Underdog Effects in a Rational Choice
Model." Public Choice 25: 83-89.
Gaskell, G. 1974. "Polls and the Voters." New Society 4(April 28) (600): 23-24.
Goeke, J.W. 1965. "Public Reaction to Pre-Election Polls and Computer Election Projections."
Public Opinion Quarterly 29: 444-445.
Granovetter, M. 1978. "Threshold Models of Collective Behavior." American Journal of Sociol-
ogy 63: 1420-1443.
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms
510 THE SOCIOLOGICAL QUARTERLY Vol. 28/No. 4/1987
Grush, J.E. 1980. "The Impact of Candidate Expenditures, Regionality, and Prio
the 1976 Democratic Presidential Primaries." Journal of Personality and Soc
38: 337-347.
Heard, A. 1960. The Costs of Democracy. Chapel Hill: University of North Carolina Press.
- 1968. "Political Financing." Pp. 235-241 in International Encyclopedia of the Social
Sciences, vol. 12. New York: Macmillan.
Henshel, R.L. 1978. "Disciplinary Prestige and the Accuracy of Social Predictions as a
Deviation-Amplifying Feedback." Pp. 143-158 in Sociocybernetics, vol. 2, edited by R.F.
Geyer and J. van der Zouwen. London: Martinus Nijhoff.
. 1982. "The Boundary of the Self-Fulfilling Prophecy and the Dilemma of Social
Prediction." British Journal of Sociology 33: 511-528.
. 1986. "Two Discrete-Time Models of Credibility Loops." Working Paper No. 1 (32
pp.) typed. Studies in Social Prediction, Department of Sociology, University of Western
Ontario.
Hovland, C.I., I.L. Janis, and H.H. Kelley. 1953. Communication and Persuasion. New Haven, CT:
Yale University Press.
Jacobson, G. 1980. Money in Congressional Elections New Haven, CT: Yale University Press.
Kayden, Xandra. 1973. "The Political Campaign as an Organization." Public Policy 31:
240-277.
Kelly, J. 1976. "Maryland: Money Talks." Pp. 14-38 in Campaign Money, edited by H.E.
Alexander. New York: Free Press.
King, R., and M. Schnitzer. 1968. "Use of Private Political Polling." Public Opinion Quarterly 32:
431-436.
Langbein, L. 1986. "Money and Access: Some Empirical Evidence." Journal of Politics 48
1052-1065.
Laponce, J.A. 1966. "An Experimental Method to Measure the Tendency to Equibalance in a
Political System." American Political Science Review 60: 982-993.
Marsh, C. 1985. "Back on the Bandwagon: The Effect of Opinion Polls on Public Opinion."
British Journal of Political Science 15: 51-74.
Maruyama, M. 1963. "The Second Cybernetics: Deviation-Amplifying Mutual Causal Pro-
cesses." American Scientist 51: 164-179.
Mason, W.M. 1973. "The Impact of Endorsements on Voting." Sociological Methods and
Research 1: 463-495.
May, R.M., and B. Martin. 1975. "Voting Models Incorporating Interactions Between Voters."
Public Choice 22: 37-53.
McCarthy, J.D., and M.N. Zald. 1976. "Resource Mobilization and Social Movements: A Partial
Theory." American Journal of Sociology 82: 1212-1239.
McClosky, H. 1968. "Political Participation." Pp. 252-265 in International Encyclopedia of the
Social Sciences, vol. 12. New York: Macmillan.
Mendelsohn, H. 1966. "Western Voting and Broadcasts of Results on Presidential Election Day."
Public Opinion Quarterly 30: 212-225.
Mendelsohn, H., and I. Crespi. 1970. Polls, Television and the New Politics. Scranton, PA:
Chandler.
Merton, Robert K. 1936. "The Unanticipated Consequences of Purposive Social Action." Ameri-
can Sociological Review 1: 894-904.
1948. "The Self-Fulfilling Prophecy." Antioch Review 8: 193-210.
Milbrath, Lester W. 1965. Political Participation How and Why Do People Get Involved In
Politics? Chicago: Rand McNally.
Napolitan, Joseph. 1972. The Election Game. Garden City, NY: Doubleday.
Navazio, R. 1977. "An Experimental Approach to Bandwagon Research." Public Opinion Quar-
terly 41: 217-225.
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms
The Emergence of Bandwagon Effects: A Theory 511
Zeitlin, I. 1981. The Social Condition of Humanity. Oxford: Oxford University Press.
Zetterberg, Hans. 1965. On Theory Construction in Sociology. 3rd ed. Totowa, NJ: Bedminster
Press.
This content downloaded from 202.92.130.58 on Sat, 10 Sep 2016 01:26:49 UTC
All use subject to http://about.jstor.org/terms