AND DEBATES

COMMENTARY

Scapegoating in Baseball

Oscar Grusky's "Managerial Succession suggest each has different implications

and Organizational Effectiveness" (Amer- and can be tested with little difficulty. A
ican Journal of Sociology, July, 1963) small amount of data toward such a test is
makes use of an ingenious source of data offered below. As a prior step, we present
for comparative organizational analysis. the three explanations of the succession-
By picking major league baseball clubs for effectiveness relationship and the different
his focus, he has available to him twenty predictions which they make.
organizations which are identical in goals
and in a wide variety of other structural THREE EXPLANATIONS

characteristics. There is considerable po- 1. The common-sense one-way causality

tentiality here, but unfortunately Grusky's theory.-This explanation fully accepts the
analysis offersus no more than the fact that fact that the field manager of a baseball
those clubs which have done the worst over team is a major influence on a team's per-
the years have changed their field manag- formance. When the team is doing poorly,
ers most frequently. he is rightfully held responsible. Conse-
For this rather obvious correlation Grus- quently, he is fired and replaced with an
ky suggests two possible explanations. He alternative manager who, it is hoped, will
quickly disposes of the common-sense ex- do better. A new manager typically will
planation of one-way causality: the man- raise the performance of a team, since he
ager is fired because the team performs can benefit by avoiding the errors that his
poorly. Despite the fact that this explana- predecessor made.
tion is consistent with the data he presents, There is no reason, by this theory, to
he charges it with, among other things, the expect any diminution in a team's perform-
failure "to stimulate careful empirical test." ance after a new manager has been hired.
Such a charge is doubly puzzling. First Instead, deteriorating performanceleads to
of all, it is not at all clear why this par- managerial succession which in turn leads
ticular explanation offers any less of a clear to increasedeffectiveness of performance.A
empirical test than the alternative offered. team which is fortunate enough to have
Second, one looks in vain for the careful acquired a good manager will perform ef-
empirical tests that Grusky's more compli- fectively over a prolonged period and will
cated explanation has produced. At the have few managerial changes. A team that
very least, one might have expected the is not so fortunate in its choice of manag-
specification of how one might get data ers will have poorer performance and a
that would allow a choice. Faced with two high frequency of managerial succession.
explanations that apparently handle the 2. The Grusky two-way causality the-
available data equally well, we prefer the ory.-Grusky also assumes that the field
simpler one, Grusky's invocation of Ernest manager is a major influence on a team's
Nagel notwithstanding. performance.However, the relationship be-
However, we wish to suggest that the tween effectiveness and succession, he ar-
common-sense explanation, Grusky's ex- gues, is reciprocal rather than one way. It
planation, and a third one which we will is certainly true that a team performing
69

This content downloaded from 138.025.078.025 on August 03, 2016 03:48:15 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
70 THEAMERICAN
JOURNALOF SOCIOLOGY
badly will frequently cause the manager to baseball talent is the most important de-
be fired. However, such managerialchanges terminant of performance,that is, the play-
tend to have a number of interrelated and ers themselves. The manipulation of this
undesirable consequences. To quote Grus- talent by the field manager will make very
ky: little difference. The New York Yankees,
A managerialchange inevitably upsets old while Casey Stengel was the manager, won
patternsof behavior.New organizational poli- ten pennants in twelve years. Is this the
cies . . . producechangesof great magnitude measure of Stengel's skill or do the New
in the internalstructureof the team.Members York Mets, who are winning no pennants,
are forcedto adaptnot only to the successor's more accurately reflect it? We suggest that
new way of doing things but also to the new the Yankees would have done as well and
informalcoalitionsthat inevitablydevelop. the Mets would have (or more accurately,
Frequentmanagerialchangecanproduceim- could have) done no worse.
portantdysfunctionalconsequenceswithin the Two qualifications are worth making be-
team by affectingstyle of supervisionand dis- fore one dismisses the field-manager role
turbingthe informalnetworkof interpersonal
relationships.. . . The resultinglow primary out of hand. First of all, a truly talented
groupstabilityproduceslow morale and may manager such as Gene Mauch of the Phil-
therebycontributeto team ineffectiveness.De- lies may make a long-term contribution
clining clientele support may encourage a through his skilful development of the po-
greater decline in team moraleand perform- tentialities of young players. This may not
ance. The consequentcontinueddrop in prof- be reflected in immediate change in team
itabilityinducespressuresfor furthermanage- performance.
rial changes. Such changes in turn produce A second qualification acknowledges
additionaldisruptiveeffects on the organiza- some over-all effect by field managers but
tion and the vicious circle continues. assumes that the variance in skill between
Clearly, a managerial change by the those who become field managersis so small
Grusky theory should produce a further that managerial skill may be considered
deteriorationin performanceby an already a constant. It is worth noting, in this re-
faltering team. spect, that few "new" managers make their
3. The ritual scapegoating no-way cau- appearance when managerial shifts are
sality theory.-Unlike the above theories, made. Instead, there is a pool of former
this one assumes that the effect of the field managers, frequently employed as coaches
manager on team performance is relatively by various teams, who are usually called
unimportant.1In the long run, the policies upon when changes are to be made. Most
of the general manager and other front- of these coaches were fired from their posi-
office personnel are far more important. tions as managers, presumably because
While judicious trades are helpful (here they failed to produce winning teams. Such
the field manager may be consulted but recruiting practices strongly suggest the
does not have the main responsibility), interchangeabilityof managers and the im-
the production of talent through a well-or- probability of explaining variance in team
ganized scouting and farm system is the performanceby anything the field manager
most important long-run determinant. The does.
field manager, who is concerned with day- If the field manager makes little differ-
to-day tactical decisions, has minimal re- ence, how does one account for the high
correlation which Grusky reports between
sponsibility for such management func-
team performance and rate of managerial
tions.
succession? The answer is quite simple:
In the short run, the supply of available The firing of the field manager is a classic
' Grusky suggests this possibility on p. 30, n. 23, example of ritual scapegoating. It is a con-
but he does not pursue its implications. venient, anxiety-reducing act which the

This content downloaded from 138.025.078.025 on August 03, 2016 03:48:15 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 AND DEBATES
COMMENTARY 71
participants in the ceremony regard as a ticularly appropriate for analysis for two
way of improving performance, even reasons: (1) mid-season changes tend to
though (as some participants may them- maximize disruptive effects and thus should
selves admit in less stressful moments) produce a fairer test for Grusky's theory;
real improvement can come only through and (2) there is less opportunity for influ-
long-range organizational decisions. ences other than the field manager's in-
Those involved have a strong stake in fluence to operate. In the gap between one
maintaining the myth of managerial re- season and the next, trades, the aging and
sponsibility. If the field manager himself retirement of present players, and the de-
denies responsibility for the team's failures, velopment of players from the farm sys-
then his claim for responsibility when the tem may serve to alter the available talent
team is successful is weakened. As for the with which the new manager has to work.
front office and the players, it is a happy If we compare the performanceof a team
excuse for what is really their own respon- during one season in a period prior to a
sibility. Finally, as Grusky notes, the fans managerial change and in the period fol-
of a poorly performing team can be ap- lowing the change, then we get a different
peased and their hopes of future success prediction from each of the above theories.
rekindled by this simple ritual act. (1) By common sense, we should expect an
Note that if we were simply to compare increase in team effectivenesses. (2) By
the performanceof the team after the man- Grusky, we should expect a decrease in
ager's dismissal with the performance be- team effectiveness due to the disruption of
fore we would surely find some immediate succession. (3) By the scapegoat theory,
improvement. Such a result is an artifact we should expect no change (after control-
of the conditions that produce the ritual, ling for the slump-ending effect).
namely, a slump in performance. If we We shall present data on twenty-two
compared average rainfall in the month mid-season managerial changes from 1954
preceding and the month following the per- to 1961.2 Team won-lost record was re-
formanceof the Hopi rain dance, we would corded at four points in time: (1) as of
find more rain in the period after. The approximately two weeks before dismissal;
dance is not performed unless there is a (2) as of the day of dismissal; (3) as of
drought, so such a comparison would be approximately two weeks after dismissal;
misleading. Nevertheless, this "slump-end- and (4) as of the end of the season (Ta-
ing" effect may help to account for the ble 1).
tenacity of belief in the effectiveness of the There are a number of comparisons of
ritual. interest. First of all, it is quite clear that
Our prediction would be that if the dismissals take place in periods of declin-
slump-ending effect is controlled, there will ing performance. Only four of the twenty-
be no difference in performance under the two teams had a higher percentage at the
old manager and the new manager. How- time of dismissal than in the previous time
ever, teams which perform poorly over a period. There is some indication of imme-
prolonged period will have more frequent diate improvement, as fifteen of the twen-
occasion to resort to the act of ritual scape- ty-two teams show better performance in
goating; hence the correlation between the next two weeks. In only one case, how-
team effectiveness and rate of succession ever, does a slumping team recover under
among field managers. the new manager to a point where the
team's won-lost percentage is higher two
A TEST OF THE THEORIES weeks after dismissal than it was two weeks
Managerial successions take place at before. Such data seem to document the
particular moments in time. Those that 2We are indebted to Andrea Modigliani for com-
occur in the middle of the season are par- piling the data reported here.

This content downloaded from 138.025.078.025 on August 03, 2016 03:48:15 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
72 JOURNALOF SOCIOLOGY
THEAMERICAN
existence of a slump-ending effect but are mary group relations of ball players, it ap-
equally consistent with all of the theories. parently does not lead to any visible de-
The most relevant test of the theories is terioration of performance.
the comparisonbetween won-lost record up Perhaps there is some slight improve-
to two weeks prior to the dismissal of the ment which can be attributed to the ritual
old manager and won-lost record for the itself. The new manager may have no effect
remainder of the season under the new through anything he does, but the players'
manager. Thus, the two weeks preceding belief in the efficacy of the ritual boosts
the dismissal of the manager are removed their morale and brings about a short-run
from the performancerecord to control the improvement.To test this, we can compare
slump effect. performanceup until two weeks before dis-
In thirteen of the twenty-two cases, the missal with performance for the remainder
TABLE 1 of the season starting two weeks after dis-
missal. In this comparison, both the peri-
TEAM PERFORMANCE AFTER CHANGING
ods immediately preceding and immediately
MANAGER, 1954-61
following are controlled. Here there is even
Im- Deteri-
less evidence of effect: in twelve cases, the
Comparison prove- ora- N team does better under the new manager,
ment tion in ten cases it does worse.
Two weeks prior versus time
of dismissal ............. 4 18 22 CONCLUSION
Time of dismissal versus two
weeks later .............. 15 7 22 The modest amount of data presented
Two weeks prior to dismissal here warrants no firm conclusion. How-
versus two weeks after dis- ever, Grusky's cavalier dismissal of the
missal ...... 1 17 18*
Two weeks before dismissal simple common-sense explanation seems
versus post-dismissal rec- unwarranted. We do not know if it is as
ord .................... 13 9 22
Two weeks before dismissal "scientific" as Grusky's more elaborate
versus record from two theory, but it proves the better predictor
weeks after dismissal until
end of season ............ 12 10 22 here. However, it still remains to be estab-
lished that the field managerhas any effect
* This comparison refers only to those eighteen teams
whose performancewas declining at the time of dismissal.
on team performance.Until it is, we prefer
the scapegoating explanation of the corre-
team performs better under the new man- lation between effectiveness and rate of
ager than it had under the old manager up managerial succession.
to two weeks prior to his dismissal. This is WILLIAM A. GAMSON
unfortunately not a very definitive result
for choosing between the common-sense NORMAN A. SCOTCH
and the scapegoating prediction, but it UNIVERSITYOF MICHIGAN
clearly goes against the Grusky prediction. AND
If managerial succession disrupts the pri- HARVARDUNIVERSITY

Reply

Professors Gamson and Scotch present as of "approximately"two weeks before or

data on twenty-two mid-season managerial after the manager's dismissal. In our repli-
changes from 1954 to 1961. All of their cation of their study we found twenty-
comparisonsrequired team won-lost record three cases that qualified when we defined

This content downloaded from 138.025.078.025 on August 03, 2016 03:48:15 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).