a ge for l eisu re? p olitical m ed iation an d th e im p ...ea3/age for leisure asr 2005.pdf ·...

Scholars have increasingly turned their atten-tion to the consequences of social move-

ments (cf. McAdam, McCarthy, and Zald 1988and Amenta and Caren 2004). Much of thiswork has focused on the external consequencesof movements, especially those relating tostates and struggles over legislation. Despitethis work, one reviewer (Giugni 2004) recent-ly argued that our knowledge accumulationon the subject has thus far been minor. Others

(McAdam 1999; Zald 2000) argue that thepolitical process and resource mobilizationmodels do not help to explain the consequencesof these movements. Other reviewers (Bursteinand Linton 2002) claim that quantitative analy-ses in this area have frequently been misspec-ified and that when public opinion is takeninto account challengers are found to have lit-tle direct influence on state-related outcomes.

In this paper, we seek to contribute to thisdebate by elaborating and appraising a politi-cal mediation theory of social movement con-sequences (Piven and Cloward 1977; Amenta,Carruthers, and Zylan 1992; Skocpol 1992;Amenta, Bernstein, and Dunleavy 1994;Fording 1997; Amenta, Halfmann, and Young1999; Lipset and Marks 2000). Instead of ask-ing whether movements are generally influen-tial or whether certain aspects of movementsare always influential, as others have done, we

AAggee ffoorr LLeeiissuurree?? PPoolliittiiccaall MMeeddiiaattiioonn aanndd tthhee IImmppaacctt ooff tthheePPeennssiioonn MMoovveemmeenntt oonn UU..SS.. OOlldd--AAggee PPoolliiccyy

Edwin Amenta Neal CarenNew York University New York UniversityUniversity of California, Irvine

Sheera Joy OlaskyNew York University

This article elaborates a political mediation theory of the impact of social movements onstates and policy, positing that the influence of mobilization and specific strategies ofcollective action depends on specified political contexts and the type of influence sought.Examining the influence of the U.S. old-age pension movement, which involved millionsof people, this article appraises the mediation model using state-level data from the1930s and 1940s on Old Age Assistance—the main support for the aged at the time—anda Senate vote for generous senior citizens’ pensions in 1939. Our models control forother potential influences, notably public opinion, which is often ignored in empiricalstudies and sometimes claimed to be responsible for causal influence mistakenlyattributed to challengers. We employ pooled cross-sectional and time series analyses andfuzzy set qualitative comparative analysis (FSQCA), which is especially suited toappraising the combinational expectations of the political mediation model. Both sets ofanalyses show that the pension movement was directly influential on the outcomes andprovide support for the political mediation arguments.

AAMMEERRIICCAANN SSOOCCIIOOLLOOGGIICCAALL RREEVVIIEEWW,, 22000055,, VVOOLL.. 7700 ((JJuunnee::551166––553388))

#2322-ASR 70:3 filename:70308-amenta

Direct correspondence to Edwin Amenta,Department of Sociology, New York University, 269Mercer Street, New York, NY 10003 ([email protected],[email protected]). For helpful comments and criticismon a previous version of this paper, the authors thankLaure Bereni, Jennifer Earl, Jeff Goodwin, David F.Greenberg, LaDawn Hagland, Mildred A. Schwartz,the NYU Politics, Power, and Protest Workshop,three anonymous ASR reviewers, and the ASR editor.

Delivered by Ingenta to :University of California

Tue, 13 Nov 2007 04:51:20

ask under what conditions are social move-ments likely to be influential. Our politicalmediation theory holds that political contextsmediate the influence of challengers’ mobi-lization and strategies. We argue that in somefavorable contexts mobilization may be enoughin itself for a challenger to exert influence andthat under more difficult political circum-stances more assertive strategies are needed. Inyet more difficult political contexts (which wespecify), a movement may not be able to exertany influence. Moreover, we argue that it takesa combination of favorable political contexts,mobilization, and assertive actions to bringabout far-reaching state outcomes. We speci-fy what constitutes long- and short-term favor-able and unfavorable political contexts, asdemanded by critics of previous social move-ment research (Goodwin and Jasper 1999);these differ from the standard four of the polit-ical opportunity model (McAdam 1996;Tarrow 1996). We also reconceptualize whatcounts as an assertive strategy (cf. McAdam1999; Kitschelt 1986), as previous definitionshave been too broad to address different meansemployed by state-oriented challengers.

We appraise the model and some alternativesby analyzing the impact of the U.S. old-agepension movement on old-age policy in itsformative years. Though largely forgottentoday, the old-age pension movement was amajor political phenomenon. The TownsendPlan,1 the largest mass pension organization,was formed in 1934 and in less than two yearshad organized two million older Americansinto Townsend clubs behind the slogan “Youthfor Work—Age for Leisure.” Townsend Plansupporters made up one of only about thirty

social movement organizations ever to attract1 percent or more of the U.S. adult population(Skocpol 2003). In 1936, the Townsend Planwas the subject of more than 400 articles in theNew York Times, placing seventh among alltwentieth-century social movement organiza-tions (see Table 1) in number of mentions intheir peak year. In addition to the TownsendPlan, many notable state-level pension organ-izations demanded generous support for theaged—senior citizens’ pensions—rather thanthe subsistence-level assistance or restricted,wage-related annuities provided or promised bythe 1935 Social Security Act. The old-age pen-sion movement case is a useful one to appraisepolitical mediation theory as the movementvaried greatly in its mobilization and actionsand operated in multiple political contexts,across states and times.

In addition, because the theory has impli-cations for different outcomes, we employ avariety of data on outcomes. The first data setconcerns state-level Old Age Assistance (OAA)programs, which were the main support forthe aged from 1936 through 1950. We exam-ine the generosity of OAA stipends and theprogram’s coverage among the aged popula-tion. These two outcomes add leverage to ouranalysis, because the pension movement’sclaims and strategies largely focused on theamount of benefit, rather than extent of cov-erage, and thus we would expect that the pres-ence and activity of pension organizationswould influence the amount of benefit morethan the coverage. We also examine whichsenators voted for a 1939 measure aiming totransform U.S. old-age policy into generoussenior citizens’ pensions for most adults over60 years. The political mediation model holdsthat effecting radical change requires moreextensive determinants than does influencingmore moderate programs, and the voting dataaddress these claims. We use two methods toappraise arguments: multiple regression analy-sis and fuzzy set qualitative comparative analy-sis (FSQCA), the latter of which facilitatesthe examination of complex and multiplecausal arguments (Ragin 1987, 2000), such asthose of the political mediation model. Eachtype of analysis supports our claims.

PPOOLLIITTIICCAALL MMEEDDIIAATTIIOONN AANNDD IIMMPPAACCTT OOFF PPEENNSSIIOONN MMOOVVEEMMEENNTT——––551177


1 Although most scholars, following Holtzman(1963), refer to this organization and phenomenon asthe “Townsend movement,” we call it the TownsendPlan first and foremost for historical accuracy. RobertEarl Clements, the organization’s initial leader,thought “Townsend Plan” was appealing and usedthat name. The newspapers did likewise. The NewYork Times index accordingly refers to the TownsendPlan and searching ProQuest for “Townsend move-ment” misses most articles on the Townsend Plan.Also, the Townsend Plan was a social movementorganization rather than a movement. In addition,“Townsend Plan” later became the official name ofthe organization.


Tue, 13 Nov 2007 04:51:20

TTHHEE SSTTAATTEE--RREELLAATTEEDD CCOONNSSEEQQUUEENNCCEESSOOFF SSOOCCIIAALL MMOOVVEEMMEENNTTSS AANNDDPPOOLLIITTIICCAALL MMEEDDIIAATTIIOONN TTHHEEOORRYY

MMOODDEERRAATTEE AANNDD RRAADDIICCAALL IINNFFLLUUEENNCCEE

A first step for scholars who seek to study state-related consequences of social movements is todefine “success” or “influence” for challengersmaking state-related claims. We follow in thefootsteps of most state-related research, focus-ing on new advantages (see Gamson [1975]1990; Amenta and Caren 2004; Meyer 2005).But we reject Gamson’s ([1975] 1990) defini-tion of “success”—whether a challenger’sclaims were mainly acted on—because of itslimitations. Challengers differ in how far-reach-ing their goals are, and thus a challenger mayfail to achieve its stated program, but still winsubstantial new advantages for its constituents(Amenta and Young 1999). There are also thepossibilities of negligible “successes,” such asa program that did not realize its intended

effects, and negative consequences, such asrepression or restrictions on movements (Pivenand Cloward 1977; McCarthy and McPhail1998).

For these reasons, we employ a wider conceptof influence based on collective goods, or group-wise advantages or disadvantages from whichnonparticipants in a challenge cannot be easilyexcluded (Hardin 1982). Collective goods canbe material, such as categorical social spendingprograms, but they can also be less tangible,such as new ways to refer to members of agroup. Most state-related collective action indemocratic political systems is aimed at majorchanges in policy and the bureaucratic enforce-ment and implementation of that policy(Amenta and Caren 2004). State social poli-cies are institutionalized benefits that providecollective goods routinely to those meetingspecified requirements (Skocpol and Amenta1986). Once enacted and enforced with bureau-cratic means, categorical social spending pro-

551188——––AAMMEERRIICCAANN SSOOCCIIOOLLOOGGIICCAALL RREEVVIIEEWW


Table 1. Top 25 U.S. Social Movement Organizations in the 20th Century, by Mentions in Articles in Peak Year,in the New York Times and the Washington Post

New York Times New York Times Washington PostOrganization (Peak Year) Articles Front Page Articles

01. American Federation of Labor (1937) 1,050 205 47602. Black Panthers (1970) 1,028 111 61703. CIO (1937) 786 186 32504. NAACP (1963) 762 128 44605. Ku Klux Klan (1924) 672 180 33906. Anti-Saloon League (1930) 409 99 9107. Townsend Plan (1936) 402 68 11808. Students for a Democratic Society (1969) 381 90 17409. Congress of Racial Equality (1963) 369 32 8610. America First Committee (1941) 280 24 12111. American Legion (1937) 263 70 20012. John Birch Society (1964) 255 32 12813. League of Women Voters (1937) 246 4 11714. American Civil Liberties Union (1977) 231 24 10215. Moral Majority (1981) 221 10 26816. Southern Christian Leadership Conference (1968) 215 36 14217. German American Bund (1939) 200 32 7118. Student Nonviolent Coordinating Committee (1966) 195 47 7619. Veterans of Foreign Wars (1950) 180 22 10420. American Liberty League (1936) 174 53 13621. Christian Coalition (1996) 170 52 25322. Association Against the Prohibition Amendment (1930) 168 56 3723. Weathermen (1970) 159 22 9224. Symbionese Liberation Army (1974) 157 23 9725. Jewish Defense League (1971) 145 31 91

Note: CIO = Congress of Industrial Organizations; NAACP = National Association for the Advancement ofColored People.


Tue, 13 Nov 2007 04:51:20

grams provide beneficiaries rights of entitle-ment. With bureaucratic reinforcement, an issuecan become privileged in politics, biasing thepolitical system in favor of the group. Benefitsthrough legislation can range from structuralbenefits that extend the political leverage of agroup, such as enhanced voting, associational,or civil rights, to one-shot pecuniary benefits,such as summer jobs, extensions of unemploy-ment insurance, housing vouchers, or bonuspayments. We argue that effecting more radicalchanges requires more extensive favorable con-ditions, both internal and external.

The old-age pension movement demandedgenerous and unrestricted grants to all nonem-ployed Americans over the age of 60 for theirlifelong service to the country—or what it called“pensions” for “senior citizens.” Pension pro-ponents in individual states demanded that OAAprograms be converted into pension programs.Their demands were not met. Our focus, how-ever, is on whether these groups had an impacton OAA, which permanently changed the rela-tionship between the state and the aged. Wealso address whether the movement affected aSenate vote for a pension alternative to existingold-age programs. Finally, we estimate the con-ditions under which such radical action mayhave been possible.

PPOOLLIITTIICCAALL MMEEDDIIAATTIIOONN TTHHEEOORRYY

Many scholars have developed (Piven andCloward 1977; Amenta et al. 1992; Skocpol1992; Amenta et al. 1994; Fording 1997;Amenta et al. 1999; Lipset and Marks 2000) ortested (Cress and Snow 2000; Soule and Olzak2004; Giugni 2004) political mediation modelsof social movement consequences. The basicidea is that challengers must engage in collec-tive action that changes the calculations of rel-evant institutional political actors and thusmobilize and adopt strategies in ways that fitpolitical circumstances. State actors must inturn see a challenger as potentially facilitatingor disrupting their own goals—for example,augmenting or cementing electoral coalitions,gaining in public opinion, or increasing supportfor the mission of governmental bureaus.Political mediation theory rejects the idea thatindividual organizational forms, strategies, orpolitical contexts will always influence chal-lengers, as is generally argued (see Amenta and

Caren 2004 for review). Instead, the theoryposits that different mobilizations and collectiveaction strategies will be more productive insome political contexts than in others.

The most extensive versions of the politicalmediation theory (Amenta et al. 1999; Amentaforthcoming) build upon arguments thatresource mobilization, strategies, and politicalcontext influence the consequences of move-ments. These versions argue that mobilizingrelatively large numbers of committed people isprobably necessary to winning new collectivebenefits for those otherwise underrepresentedin politics (Rucht 1998; Skocpol 2003; seereview in McCarthy and Zald 2002). In addition,making claims regarding the worthiness of thegroup (Tilly 1999b) and the plausibility of itsprogram (Cress and Snow 2000; Ferree et al.2002) is also necessary. Favorable political con-texts, both long- and short-term, are also help-ful (Jenkins and Perrow 1977; Kitschelt 1986;Kriesi 1995; Almeida and Stearns 1998; Meyerand Minkoff 2004). The political mediationmodel, however, helps to explain the impact ofsocial movements by examining mobilizationand strategies in combination with differentsorts of political contexts. In highly favorablepolitical contexts, all that should be required isa certain threshold of resource mobilization andminimally plausible and directed framing andclaims-making. In less favorable political con-texts, more assertive strategies of collectiveaction would be required for a social move-ment to have influence. In yet other politicalcontexts, where powerful systemic conditionswork against challengers, it may be impossiblefor the challengers to exert much influence. Inshort, the context must be extremely favorableand the mobilization and action extensive forchallengers to achieve the most radical goals.

EELLAABBOORRAATTIINNGG TTHHEE TTHHEEOORRYY

According to political mediation theory, theability of a challenger to win collective benefitsdepends partly on conditions it can control,including its ability to mobilize, its goals andprogram, its form of organization, and its strate-gies for collective action, including issue fram-ing and other claims-making. However, theimpact of even well-mobilized challengers alsodepends on political context. Political mediationtheory holds that political conditions influence




Tue, 13 Nov 2007 04:51:20

the relationship between a challenger’s mobi-lization and collective action on the one hand,and policy outcomes on the other. It holds thatmobilization and collective action alone areoften insufficient to effect changes in publicpolicy that would benefit a challenger’s con-stituency. The model posits relationshipsbetween mobilization and strategies and struc-tural and short-term political contexts. The argu-ment focuses on specific political contexts,taking into account criticisms that politicalopportunity models are often conceptualizedat too broad a level to be empirically tested(Goodwin and Jasper 1999; cf. Meyer andMinkoff 2004). Our political mediation argu-ment also differs from political opportunitystructure arguments (McAdam 1996; Tarrow1996) in that the latter tend to address the mobi-lization of challengers than their political con-sequences (Zald 2000). Also, our understandingof political contexts is based not on standardpolitical opportunity factors, but on contextsfound influential in altering social policy(Mayhew 1986; Skocpol 1992; Amenta 1998;Hicks 1999; Huber and Stephens 2001).

The political mediation theory holds that spe-cific long-term aspects of political and partysystems influence the productivity of chal-lengers’ action. First, the degree to which for-mally democratic institutions are bound bydemocratic practices is key (Amenta 1998; Tilly1999a). An extension of democratic rightsentails lowering the legal restrictions on insti-tutional political participation for the commoncitizens, including their ability to assemble anddiscuss issues. A highly democratized polity isalso characterized by meaningful choices amongparties or factions. By contrast, an underde-mocratized polity is one in which political lead-ers are chosen by way of elections, but in whichthere are great restrictions on political partici-pation, political assembly and discussion, vot-ing, and choices among leadership groups. Anunderdemocratized political system greatlydampens the impact of the collective action ofchallengers, assuming they are able to arise inthese polities. Underdemocratized political sys-tems are characterized by noncompetitive elec-tions and formal and informal restrictions onvoting, such as poll taxes, extensive eligibilitytests, harassment, and violence. The UnitedStates, especially the South, included manyunderdemocratized polities, as franchise barri-

ers were erected at the end of the nineteenth cen-tury and persisted through most of the twenti-eth century (Burnham 1974).

Second, patronage-oriented political parties—that is, autonomous, long-lasting, hierarchicalorganizations that seek to nominate candidatesfor a wide range of public offices and rely sub-stantially on material incentives (Katznelson1981; Mayhew 1986)—tend to deflect claimsfor collective benefits sought by pro-socialspending challengers. Granting automatic andlong-term entitlement claims to groups of citi-zens limits the sort of discretionary spending,such as for government jobs and contracts, thatmaintains a patronage-oriented political organ-ization. For these reasons patronage-orientedparties regard social movement organizations asa menace and consider programmatic spendingpolicies a threat to the individualistic rewards onwhich such parties thrive. Although it is notimpossible for movements to exert influenceunder these circumstances, these structuralimpediments make it difficult, and also oftenthwart the efforts of state actors and insurgentsin the party system to enact or enhance pro-grammatic public spending policies in favor ofchallenging groups. The United States, espe-cially in the Northeast and Midwest, was char-acterized by many patronage-oriented partysystems for most of the twentieth century(Mayhew 1986).

Medium-range and short-term political con-texts also influence the prospects of mobilizedgroups that hope to gain leverage in politicalsystems that are mainly democratized and arenot dominated by patronage-oriented parties.Domestic bureaucrats are key actors here.Bureaucrats whose mission is consistent withthat of a challenger—assuming they have ini-tiative, talent, and power—may provide admin-istrative rulings, enforce laws, or propose newlegislation that aids a challenger’s constituency,even within the context of an otherwise indif-ferent or opposed state (Skocpol 1992). Thesedomestic bureaucrats may advance such legis-lation further than they had intended if a chal-lenger summons a show of strength. Thesearguments are similar to Kitschelt’s (1986) aboutthe role of implementation capacities. But ourarguments are both wider, in considering the dif-ferent ways in which domestic bureaucraciesmight amplify the impact of challengers’ col-lective action, and more localized, as the rele-




Tue, 13 Nov 2007 04:51:20

vant bureaucracies will differ according to thechallenger and its constituency. Bureaucraciescentrally concerned with social and labor issueswere inaugurated largely in the 1910s in theUnited States; they varied widely in their ori-entation and power in the 1930s, and continueto do so today.

Another crucial factor is the partisanship ofthe regime in power. A new political regime orgovernment, hoping to add to its coalition, mayaid the constituency of social movements byproposing spending or other legislation thatfavors a certain group. A regime that is favor-able or open to the possibility of increased socialspending would be expected to amplify theimpact of a challenger’s mobilization and col-lective action, while a regime opposed to socialspending would dampen it. Often parties havelong-standing commitments to ideological posi-tions or groups whose interests and goals mayconflict with those of challengers (Klandermansand Oegema 1987). For state-oriented chal-lengers who seek collective benefits throughsustained public spending, the position of theregime on higher taxation is crucial. Since the1930s the U.S. Republican party and its repre-sentatives have tended to oppose automatic,programmatic spending claims because theyimply higher taxation, whereas the nationalDemocratic party and Democrats outside theSouth have tended to be “reform-oriented”—more open to policy claims requiring taxation(Amenta 1998; Hicks 1999). Regime condi-tions varied widely across the United States inthe 1930s, as some northern states followed thenational trend to elect pro–New Deal Democratsand others did not. Partisanship variations havepersisted to this day, with the South havinglargely turned reliably Republican in the decadesafter the Civil Rights and Voting Rights Acts ofthe 1960s.

The mediation model expects these contextsto affect the influence of movement activity. Ifthe political regime is open to the claims ofchallengers and the domestic bureaucrats areprofessionalized and favorably disposed to thechallenger’s constituency, limited activity onthe part of the challenger, even simply provid-ing evidence of mobilization, is likely to besufficient to produce increased collective ben-efits. The challenger needs mainly to demon-strate that it has support, such as through writingletters, holding rallies, petitioning, initiating

public awareness campaigns, staging limitedprotest, or even engaging in visible internalevents. Members of a reform-oriented regimeare likely to use such evidence of mobilizationas a confirmation of the beneficiary group’srelative importance in an electoral coalition. Ifa reform-oriented regime has many issues on itsagenda, it is more likely to address first theissues being pressed by mobilized groups.Domestic bureaucrats are likely to portray themobilization as indicating the need for the aug-mentation or greater enforcement of its pro-grams. If the regime hopes to add to its coalitionor if domestic bureaucrats have a mission thatis not yet realized, the best-mobilized groups arelikely to win the greatest benefits in public pol-icy for their constituencies.

By contrast, achieving collective benefitsthrough public policy is likely to be more dif-ficult if neither an open regime nor adminis-trative authority exists. When the regime isopposed to the challenger or sees no benefit inadding the challenger’s beneficiary group to itscoalition and when state bureaucracies in thearea are hostile, the sorts of limited protest list-ed above are likely to be ignored or to have anegligible effect. In the face of more difficultpolitical circumstances, more assertive or bold-er collective action is required to produce col-lective benefits. Here we drop the standarddistinction between “disruptive” and “assim-ilative” (Kitschelt 1986) and “noninstitutional”and “institutional” (McAdam 1999) strategies,instead focusing on variations in assertivenessof action, with “assertive” meaning the use ofincreasingly strong political sanctions—thosethat threaten to increase or decrease the likeli-hood of political actors gaining or keepingsomething they see as valuable (their positions,acting in accordance with their beliefs) or to takeover their functions or prerogatives. Sustainedpolitical action to unseat a representative, forexample, would be more threatening than, say,dispatching protesters to picket or to occupythe representative’s office. The institutional col-lective action of challengers works largely bymobilizing large numbers of people behind acourse of action, often one with electoral impli-cations. This collective action may be designedto convince the general public of the justice ofthe cause and influence elected and appointedofficials in that manner, but it can also demon-strate to these officials that a large segment of




Tue, 13 Nov 2007 04:51:20

the electorate is willing to vote or engage inother political activity on the basis of a singlekey issue.

If the political regime is not supportive of thechallenger’s constituency or issue, collectiveaction will be most productive if it focuses onelected officials. Such action might neutralizethose who would otherwise be hostile to legis-lation and win the support of those who wouldbe indifferent. Assertive action might includecontesting elections, such as endorsing and sup-porting the opponents of hostile incumbents orwinning promises from them and then provid-ing support. Newcomers elected with a chal-lenger’s endorsement and support would beespecially likely to support its program or otherprograms benefiting its constituents. Otherassertive action would include attempting tooverride legislative authority, as through directdemocratic devices such as the initiative, ref-erendum, and recall. Such displays of influencemight alter the views of legislators in whosestates the mobilizations took place, even whenthe actions fail. In the face of strong electoralsanctions, legislators previously opposed maycome to support moderate measures that bene-fit the group represented by the challenger.2

The mediation argument also addresses thecharacteristics of outcomes and legislation atissue. The more radical and far-reaching theoutcome, the greater the favorable conditionsrequired and the more the movement may haveto do to influence it. Thus for some outcomes,such as improving existing programs, merelymobilizing under favorable contexts would beenough to exert influence. Similarly, employingassertive sanctions when conditions are struc-turally favorable but the short-term context isunfavorable may be enough. Fundamentallyaltering policy, however, is likely to take bothstrong mobilization and extensive assertiveshows of strength. Even these may not beenough to create sweeping changes in policy;more favorable political conditions may alsobe necessary. The same is likely to be true for

bids to transform the structural position ofgroups, such as through new voting or civilrights.

To summarize, political mediation theoryholds that the influence of mobilization andstrategies of action are conditional on specificpolitical contexts. Some systemic political con-texts—an underdemocratized polity and apatronage-oriented party system—will deadenthe influence of challengers. Medium-rangeand localized political contexts will have a morevariegated effect on the relationship between achallenger’s collective action and state-relatedresults. Regimes and bureaucracies open tochallengers’ claims will tend to repay the chal-lengers’ mobilization. Even in democratizedpolities and non-patronage-oriented party sys-tems, however, regimes and bureaucracies maystill be unfavorable. In these circumstancessocial movements would need to engage inassertive collective action in order to win newbenefits. Before we employ quantitative analy-ses to appraise the individual influence of var-ious factors, and formal qualitative analysesto address the multiple theoretical interactionsand causal pathways, we briefly introduce theold-age pension movement and U.S. old-agepolicy.

TTHHEE OOLLDD--AAGGEE PPEENNSSIIOONN MMOOVVEEMMEENNTT,,UU..SS.. OOLLDD--AAGGEE PPOOLLIICCYY,, AANNDD OOUURRAANNAALLYYSSEESS

TTHHEE TTOOWWNNSSEENNDD PPLLAANN AANNDD TTHHEE OOLLDD--AAGGEE

PPEENNSSIIOONN MMOOVVEEMMEENNTT

The Townsend Plan was founded in January1934 by Dr. Francis E. Townsend, a laid-off, 66-year-old Long Beach medical assistant, andRobert Earl Clements, a 39-year-old real estatebroker (Holtzman 1963; Mitchell 2000; Amentaforthcoming). The purpose of the organizationwas to promote the enactment of the pension-recovery program Townsend had first outlinedin letters to the editor of the Long Beach PressTelegram in September 1933. The plan called for$200 monthly pensions to all nonemployed cit-izens over 60 years, excluding criminals, andwas designed to end the Depression and ensureprosperity through the mandatory spending ofthese pensions, as well as to end poverty amongthe elderly. Clements assembled a hierarchicalorganizing staff paid by commission, andTownsend was the organization’s symbol and



2 Assertive action directed at unfavorable bureau-cracies is also expected to be more productive thanmobilization and limited protest. Unfortunately, wewere unable to gain systematic information acrossstates regarding this sort of protest and so cannot testthat part of the argument here.


Tue, 13 Nov 2007 04:51:20

spokesman, akin to the role that Colonel Sandersplayed for Kentucky Fried Chicken. To maintainenthusiasm and mobilize resources, Clementsand Townsend inaugurated Townsend “clubs”—local affiliates with no formal decision-makingpowers, but that met regularly to hear speakers,collect donations, and act in political campaigns.The Townsend Plan grabbed national attentionin late 1935 when it was organizing clubs at therate of one every two hours. At that point, leg-islation it had endorsed would have providedalmost all senior citizens with about $60 permonth, less than $200 but far more than whatwas promised by the Social Security Act.

Although Townsend clubs and their mem-bers remained the backbone of the old-age pen-sion movement, they were joined by a numberof pension organizations later in the 1930s. Agroup known as Ham and Eggs won nationalattention in 1938 with its program to provide$30 every Thursday to aged Californians. Othernotable state-level pension organizations includ-ed the National Annuity League of Colorado andthe Old Age Pension Union of Washington. Indifferent states many coalitions formed brieflyaround specific initiatives, especially in 1938,when eight pension initiatives were placed onstate ballots. For the most part, these initiativesconcentrated on providing more generous ben-efits than existing programs, although the move-ment also stood for a wider extension of benefitsto the aged. The pension movement existed foralmost two decades under diverse national polit-ical circumstances and across state-level poli-ties. The Townsend Plan chose to ignore thestate level in the 1930s, arguing that improvingOAA programs in individual states would hin-der efforts to enact pensions at the nationallevel. But the Townsend Plan, too, eventuallybegan to demand changes in state old-age laws,placing propositions for $60-per-month pen-sions on the ballots of a few western states in1943. These efforts failed, and by 1950 theTownsend Plan had lost most of its membership.

UU..SS.. OOLLDD--AAGGEE PPOOLLIICCYY

AANNDD TTHHEE PPRREESSEENNTT RREESSEEAARRCCHH

The 1935 Social Security Act created two old-age programs. One was a national, proto-old-ageinsurance program, which was not scheduled tomake any payments until 1942. The second wasa federal-state matching program called Old

Age Assistance, which immediately providedbenefits to the aged in individual states. OAAwas the workhorse of old-age protection duringthe Depression, the Second World War, andimmediately afterward. It was upgraded in 1939and was not eclipsed by Old Age and SurvivorsInsurance (known now as “Social Security”)until the 1950s. We ascertain first whether theold-age pension movement influenced OAAprograms and, if so, in what ways and why.OAA programs are useful in a test of our argu-ments because the various state programs didnot converge in their generosity and coverage,and the pension movement sought to convertthem into pension-like programs with moregenerous benefits, often through initiatives. Wewould expect the pension movement to havegreater influence on the generosity of programsthan on the extension of programs.

We also seek to determine whether the pen-sion movement induced senators to vote for ameasure to replace the two programs with asenior citizens’pension and, if so, what types ofactivities were effective, by analyzing roll-callvotes (McAdam and Su 2002; Soule et al. 1999)through multiple regression analyses. We focuson 1939, the year that the Social Security Actwas amended and the only year that there werevotes on senior citizens’pensions. The so-calledLee amendment (S76-1061), though not specif-ically a vote on authorized Townsend Plan leg-islation, would have created a widespread andgenerous $40 per month benefit, twice as largeas the average OAA payment. The $40 figurewas also the median amount that public opin-ion polls indicated that the government shouldpay in monthly old-age benefits (Gallup 1936,1939). Although the amendment failed, by avote of 17 to 56, putting on the Congressionalagenda a radical alternative and lining up votesbehind it (Kingdon 1984) is one way for a chal-lenger to influence the political process andpublic policy, as it can induce opponents toaccept more moderate legislation. Indeed,observers suggest that this is what happened in1939 in the case of the pension movement(Huston 1939). The Lee amendment vote alsogives us an opportunity to compare the deter-minants of influencing existing programs withthose of altering policy in a more fundamentalway. Because the Townsend Plan was mainlyengaged at the national level and made plausi-




Tue, 13 Nov 2007 04:51:20

ble claims about pensions at the time, we expectit should be influential.

These state and national policy outcomesprovide a large number of cases, making it pos-sible to control for other variables that might alsohave influenced old-age policy-variables thatare often ignored in studies of the impact ofsocial movements (Earl 2000). Scholars ofsocial policy argue that democratized politicalsystems, left-wing regimes, powerful domesticbureaucracies, favorable public opinion, andeconomic and demographic developments mightall influence the policy changes that othersmight attribute to the presence or activity of asocial movement (Amenta, Bonastia, and Caren2001; Skrentny 2002). Worse, some of theseconditions, such as an increase in the aged or therise to power of the Democratic party, may havespurred both the pension movement and polit-ical action on old-age benefits (Amenta andYoung 1999). Our analyses address these issues.

EEXXPPLLAAIINNIINNGG OOLLDD AAGGEE AASSSSIISSTTAANNCCEE IINN IITTSS FFOORRMMAATTIIVVEE YYEEAARRSS

DDEEPPEENNDDEENNTT MMEEAASSUURREESS

We examine two basic components of Old AgeAssistance programs, their generosity and theircoverage, from their first year in operation in1936 through 1950, when the Social SecurityAct was amended to upgrade old-age insur-ance. The first dependent measure is the aver-age size of the OAA benefit, an indication of therelative generosity of states. (For details on theconstruction and source of each dependent andindependent measure, see Appendix.) In addi-tion, we examine each state’s OAA coverage—the state’s commitment to providing assistancebroadly to elderly residents. State legislationgenerally set the parameters controlling theseprograms. Benefit levels and coverage varieddramatically from state to state.

IINNDDEEPPEENNDDEENNTT MMEEAASSUURREESS AANNDD EEXXPPEECCTTAATTIIOONNSS

We appraise the main social movement andpolitical mediation arguments through a seriesof independent measures. Three capture differ-ent facets of the political institutional aspects ofthe mediation theory, and they vary little overtime. As an indicator of voting rights and poli-ty democratization, we consider the poll tax, akey, though far from the only, means to restrict

the franchise. Nine states had poll taxes at thestart of the period, but Florida and Georgiadropped theirs. We expect that underdemocra-tized political systems, characterized by restric-tions on the franchise, would discouragemovements for generous public spending pro-grams as well as OAA benefits and coverage.We also include a measure of patronage partyorganizational strength, with the expectationthat such party organizations would discour-age categorical public spending, resulting inless generous benefits and restricted coverage.We also include administrative strength, a meas-ure of the strength and structure of the statelabor commissions. Although they did not typ-ically control OAA, their existence indicatesoverall domestic bureaucratic development andpower, and we would expect them to have apositive influence on OAA outcomes.

Two measures address medium-term politi-cal and administrative conditions. We wouldexpect that control of the state government bya pro-spending party would lead to higher qual-ity OAA programs. We model democratizedDemocratic control by including a measure forcontrol of the governor’s mansion and bothhouses of the state legislature by the Democraticparty in states without poll taxes. Additionally,we include a measure of OAA county funding.We expect that higher county contributions toOAA would negatively influence OAA out-comes, as counties had fewer and more con-tentious taxing opportunities, mainly real estatelevies.

We also employ several control measures.Per capita income addresses how much socialspending states could afford; we expect thathigher per capita income would positively influ-ence OAA (Wilensky 1975). We also includepercentage black to take into account the poten-tially dampening impact of race on OAA ben-ef its (Quadagno 1988; Lieberman 1998).Percentage aged in each state is likely to spurdemands for old-age benefits (Mitchell 2000).We also include measures of pro-old-age pub-lic opinion. Public opinion is sometimes arguedto be the only direct influence on public policy(Burstein 1999) and often is absent in empiri-cal studies of movement consequences, possi-bly resulting in their models being misspecifiedand their conclusions faulty (Burstein andLinton 2002). We analyze two 1938 Galluppolls, the earliest polling efforts on old age that




Tue, 13 Nov 2007 04:51:20

survive in forms suitable for state-level analy-ses. We focus on the question “How much permonth should be paid to a single person?” in ouranalyses of OAA generosity. The median choicewas $40. For OAA coverage, we examine thequestion “Do you think pensions should begiven to all old people, or only to old people whoare in need?” About 20 percent chose “all.”Responses are aggregated to the state level.Although there is no systematic information onthe saliency of the issue, it is likely that old agewas a prominent issue through 1941, when theSecond World War began. We therefore includecontrol measures for the war and postwar timeperiods.3

Finally, we consider three pension movementmeasures, Townsend club activity, change inclub activity, and electoral initiatives. First, weoperationalize Townsend club activity4 in twoways. Given the expectation that high mobi-lization will have a lasting influence (McCarthyand Zald 2002), we measure club activity at its

peak value for a state. Second, we measure thechange in activity, which has been argued to bemore likely than the overall mobilization levelto influence politicians who are seeking infor-mation relevant to their reelections (Burstein andLinton 2002). We also include a measure for themost assertive movement strategy to influencestate legislators—the placing of pension propo-sitions on the ballot. These propositions wereusually designed to make OAA more generous,and less frequently to relax eligibility require-ments. While the vast majority of these effortsfailed, and the successes were countered bysubsequent legislation, we expect that proposi-tions would have a positive impact on OAAgenerosity. Propositions usually involve a polit-ical show of force: their proponents petition,dramatize an issue, and insert it onto the polit-ical agenda. In this instance, we expect thatthey would pressure politicians to prove theircommitment to their aged constituents.

Although it is not possible to test all the inter-actions in the political mediation argument withmultiple regression, the model does providedifferent expectations across the independentmeasures and across time. First, we expect thelong-term institutional factors to influence bothaspects of OAA, as these are general influenceson social policy. But we also expect the socialmovement measures to have a greater influenceon the average benefit than in coverage, becausethe claims and demands of the pension move-ment were largely focused on higher benefits.

MMUULLTTIIPPLLEE RREEGGRREESSSSIIOONN RREESSUULLTTSS

Our cross-sectional panel data set includes infor-mation on the 48 states, each over a period of14 years. Under these circumstances, with mul-tiple cases from the same state and multiplecases from the same year, we expect that theerror terms would not be independent and iden-tically distributed, making pooled OLS regres-sion inappropriate. Moreover, our data set iscase dominated, with many more cross-sec-tional cases than years, rather than temporarilydominated, as is typical in time-series cross-sectional research (Beck 2001). For these rea-sons we employ a GLS random-effects model,which allows for both time-varying and time-invariant variables (Western and Beckett 1999;Kenworthy 2002; Beckfield 2003). Additionally,we expect that spending by a state in a given year



3 The first Gallup poll on old-age benefits wastaken in 1935, but the data do not survive. Otherdata are often not usable, because Gallup’s questionswere often poorly worded. From the first poll forwardGallup found that overwhelming percentages ofAmericans were in favor of government providedold-age “pensions,” a word used indiscriminately atthe time to refer to all manner of aid to the aged(Schiltz 1971: chapter 2). Gallup occasionally askedrespondents about the saliency of issues, but theywere usually asked to volunteer answers, resulting innumbers too small to aggregate to the state level (seeWeakliem 2003) even if they had survived. For com-prehensive listings of the results of polls on old age,see Cantril and Strunk (1951:541–46) and Schiltz(1971).

4 Note that we cannot, unfortunately, appraisearguments about the strategic capacities (Ganz 2000)and movement infrastructures of challengers(Andrews 2001), as these did not vary greatly for theTownsend Plan, and it is difficult to get informationby state and year for other old-age campaigns. For themost part, the Townsend Plan would have to be con-sidered high on strategic capacities and infrastructuralresources. It also had an innovative form of organi-zation, combining aspects of real estate sales organ-izations and fraternal organizations in a socialmovement organization. To the extent that theTownsend Plan mobilization influenced policy it canbe considered a supportive result for these perspec-tives.


Tue, 13 Nov 2007 04:51:20

will be related to its spending in the previousyear, potentially producing serially correlatederrors. The Wooldridge (2002) test for auto-correlation in panel data reports significant evi-dence of first-order autocorrelation in ourmodels, and we employ the Baltagi-Wu (1999)estimator to remove this disturbance. For theaverage size of OAA benefits and coverage, wereport an initial model estimating the effects ofpolitical contextual factors and control meas-ures, and then a full model that adds socialmovement measures in order to ascertainwhether they add to the explanation.5

Table 2 presents the results for the averagesize of OAA benefit. The initial model, Model1, yields significant coefficients at the .05 levelor better for the political context measures, withthe exception of democratized Democratic con-trol. All the control measures, except for the warperiod, are significant. The full model, Model2, which accounts for almost 70 percent of theoverall variance, yields positive and significantcoefficients for Townsend club peak-level meas-ure at the .05 level and the proposition measureat the .10 level. Moving up a level in Townsendclub mobilization is worth $1.23 per month,and placing a proposition on the ballot is worthabout $1.37 per month. Both are substantialgains since the average payment in 1950 dollarsacross the entire period was about $37 permonth. In Model 2, moreover, administrativestrength and OAA county funding remain sig-nificant at the .05 level, and the poll tax meas-ure is significant at the .10 level. States with apoll tax spent approximately $3.90 less permonth on OAA, while those with a tradition ofdomestic administrative development spentabout $3.88 more per month, after controllingfor other factors. The public opinion measurealso remains significant and substantial. A dec-laration of an additional dollar for the appro-

priate amount of stipend was worth 30 cents.However, when we substitute the measure ofchange in Townsend activity for the peak valueof Townsend activity, the measure is insignifi-cant. This suggests that policy-makers are notresponding to new information about the chal-lenger. In addition, the coefficient for the WorldWar II period was positive, which is counter toexpectations given the drop in attention todomestic issues, though it falls short of signifi-cance.

Table 2 also presents the results for an initialand full model on OAA coverage in Models 3and 4, respectively. Patronage party strength,democratized Democratic control, and OAAcounty funding are significant at the .01 levelin Model 3. States with Democratic controlcover approximately 3.3 percentage points moreof their elderly population, and moving up onelevel in Mayhew’s five-level measure of patron-age party strength diminishes coverage by 2.9percentage points. Moving up 10 percentagepoints in public opinion to cover all the agedmeant an increase of one percentage point incoverage. These were substantial influences, asthe average coverage across all states and timeperiods was about 22 percent. Among the othercontrol measures, the coefficient for percentageaged is significant and negative, whereas the warand postwar period coefficients are both sig-nificant and positive. The war did not lead to thereduction of coverage—just the opposite. Thecoefficients for race and income are in theexpected directions, but insignificant. Model 4explains 30 percent of the overall variance.However, neither of the social movement meas-ures, when added, has a significant impact onspending coverage.

In summary, structural and short-term polit-ical contextual factors strongly affected bothOAA measures, and in the predicted directions,although each measure was not significant in allmodels. Among the control measures, publicopinion influenced both outcomes. The findingsfor the social movement measures were mixed.Townsend club activity and propositions sig-nificantly and substantively influenced the aver-age monthly old-age stipend, but they did notinfluence coverage under OAA. These results,however, fit with political mediation expecta-tions, as the pension movement’s claims andcollective action were focused largely on the size



5 For the full models, the Hausman specificationtest indicated that the efficient random effects modelwas not significantly different from the consistentfixed effects model, and we performed a Ramseyregression specification error test (RESET) for omit-ted variables for each year. The results (not shown,but available on request) were significant for only oneyear, which is additional evidence for the suitabilityof the random effects model.


Tue, 13 Nov 2007 04:51:20

of the benefit rather than coverage, and thus wewould expect a differential influence.

QQUUAALLIITTAATTIIVVEE CCOOMMPPAARRAATTIIVVEE AANNAALLYYSSEESS

The expectations of the political mediationmodel are combinational, and these sorts ofarguments can often be better assessed usingFSQCA, for which multicollinearity is not prob-lematic (see Ragin 1987, 2000). Here we employ

crisp rather than fuzzy sets, as most of the inde-pendent measures are nominal. We located 13states with GENEROUS OAA benefits.6 (In



Table 2. Average Size of the Old Age Assistance Benefit and OAA Coverage on Selected IndependentMeasures

Benefit Size Coverage

Model 1 Model 2 Model 3 Model 4

Institutional MeasuresPoll Tax –4.584* –3.897 –3.183 –3.100

(1.88) (1.61) (1.18) (1.14)Patronage Party Strength –1.220** –0.871* –2.880** –2.858**

(2.50) (1.74) (4.65) (4.45)Administrative Strength 3.918* 3.879* –0.941 –0.974

(2.15) (2.19) (0.44) (0.46)Democratized Dem. Control –0.067 0.299 3.265** 3.315**

(0.07) (0.30) (3.34) (3.33)OAA County Funding –0.104* –0.090* –0.123** –0.123**

(1.96) (1.70) (2.34) (2.33)Movement MeasuresTownsend Club Activity .— 1.229* .— 0.075

.— (2.15) .— (0.12)Pension Proposition .— 1.374 .— 0.625

.— (1.44) .— (0.74)Control MeasuresPro-Old-Age Public Opinion 0.306** 0.300** 0.100 0.097

(3.81) (3.83) (1.32) (1.27)Per Capita Income 0.007** 0.007** 0.001 0.001

(4.86) (5.10) (0.66) (0.71)Percentage Aged 1.347** 1.323** –0.957* –0.946*

(3.70) (3.66) (2.46) (2.42)Percentage Black –0.188* –0.119 –0.067 –0.061

(2.11) (1.29) (0.65) (0.55)War Period 0.971 0.557 2.262** 2.219**

(1.30) (0.73) (3.16) (3.03)Postwar Period 4.862** 4.354** 1.870 1.788

(4.93) (4.35) (1.94) (1.82)Constant 5.184 –0.781 32.778** 32.288**

(1.08) (0.14) (7.97) (6.54)Observations 672 672 672 672Number of states 48 48 48 48R2 0.69 0.70 0.30 0.31!2 410.83** 427.25** 59.51** 60.14**Df 12 14 12 14

Notes: Data shown are unstandardized coefficients from random effects regressions. The absolute values of zstatistics are in parentheses. For definitions of measures, see text and Appendix. * p < .05; ** p < .01 (one-tailed; except for percent aged, war period, and postwar period).

6 To determine which states provided generousOAA benefits, we average the residuals of a baselinemodeling, including only per capita income and per-centage aged for each state across our entire time peri-od. We designate the 13 states that had an observed


Tue, 13 Nov 2007 04:51:20

FSQCA notation, a measure written with allcapital letters denotes its presence, while onewritten in all lowercase denotes its absence.)Two of the institutional political measures arenominal. States that employed a poll tax dur-ing this period are designated POLLTAX.Those where the state’s labor commissionerhad rule-making authority over safety lawsthroughout the period are labeled ADMIN.States that were largely controlled byPATRONAGE party organizations are thosethat score either of the top two values onMayhew’s (1986) scale. States where theDemocratic party controlled the governor’smansion and both houses of the legislature forat least 40 percent of the time are consideredDEMOCRATIC. As for social movementmeasures, states that had reached the highestlevel of Townsend club presence at any time areconsidered highly MOBILIZED, and stateswhere the pension movement placed one ormore proposition on the ballot are said to havehad ASSERTIVE collective action.

We begin by comparing the theoreticalexpectations of the political mediation modelwith those configurations standard in the socialpolicy literature. The configurational theoret-ical expectations from the institutional politicsmodel of social policy (Amenta and Halfmann2000) for generous old-age spending, net ofeconomic controls, are as follows:

polltax*patronage*ADMIN*DEMOCRAT.

(In FSQCA notation, an asterisk (*) indicatesthe logical operator and; a plus sign (+) indi-cates the logical operator or.) This expressionreads as follows: States without poll taxes andwithout patronage-oriented parties and withstrong administrative powers and withDemocratic party regimes are expected to pro-duce generous social programs. By contrast,the political mediation argument suggests thatthe mobilization and collective action of chal-lengers can also spur policy, according to thefollowing expression:

polltax*patronage*(MOBILIZED*

(DEMOCRAT+ADMIN) + ASSERTIVE).

This means that in structurally conducive andpolitically favorable short-term situations, onlychallenger mobilization is needed to producecollective benefits. When short-term politicalconditions are less favorable, more assertiveaction is the best strategy. This type of activi-ty is sufficient to bring results.

While there are 64 theoretically possiblecombinations of the six dichotomous inde-pendent variables, only 21 combinationsdescribe the experiences of the 48 states dur-ing this period. The results indicate that six ofthe combinations consist of states that alwaysexhibited generous stipends.7 (See Table 3.)The six expressions in Table 3 account for 11of the 13 positive cases. These reduce to threethat encompass each of the successful casesand can be combined in one expression (seeTable 3). Necessary conditions for high OAAbenefits were the absence of poll taxes andpatronage parties. However, other conditionsalso had to be present to account for high OAAbenef its: administrative powers andDemocratic party control; Democratic partycontrol and mobilization; or assertive collec-tive action alone.

These results have implications for both thestandard institutional political model and thepolitical mediation model. First, as expected byboth models, democratic rights and the absenceof dominant patronage parties are necessary forhigh OAA spending. This suggests that undersome structural, systemic conditions the activ-ity of both institutional political actors andsocial movements are likely to be thwarted. Inpolitical situations where it is possible to pro-mote policy, the predictions of the standardinstitutional model and the political media-tion model are all borne out. The standardinstitutional model holds that in favorable sys-temic political circumstances, a favorableregime and administrative powers would belikely to produce generous social policy. One



value that was 10 percent larger than the predictedvalue, where there was a large break in the data, asgenerous spenders. The 10 states where the observedvalue is greater than predicted but less than 10 per-cent higher are coded as intermediate or “don’t care”cases, and the 25 remaining states are coded as zeros.

7 Contradictory combinations mainly consisted offailures and are treated as such.


Tue, 13 Nov 2007 04:51:20



of the reduced combinations indicates pre-cisely that.

The results also provide strong support forthe mediation model. Two of the hypothesizedcombinations expected to lead to generousOAA benefits are present. As before, onlyunder some long-term political contextual con-ditions is movement influence possible.Necessary conditions for influence include theabsence of both poll taxes and dominantpatronage parties. Under these circumstancesand a favorable short-term political circum-stance, the Democratic party holding power,only extensive movement mobilization is nec-essary to bring about high OAA benefits, andtogether they are sufficient. When there are nomedium- or short-term conditions in favor,neither a long-standing domestic administra-tive tradition nor a Democratic regime inpower, assertive action is sufficient to producehigh OAA benefits.

WWHHOO VVOOTTEEDD FFOORR SSEENNIIOORR CCIITTIIZZEENNSS’’PPEENNSSIIOONNSS IINN 11993399??

Next we turn to the Senate vote on an old-agepension measure, the proposed Lee amend-ment to the Social Security Act Amendmentsof 1939. Lee’s amendment would have pro-vided pensions of $40 per month to all agedAmericans and would have replaced existingold-age programs. Though not specifically avote on Townsend Plan-sponsored legislation,the Lee amendment would have provided agenerous, widespread, nationally financed pen-sion, which was the hallmark of Townsend’sproposal. Funding these pensions was a more

radical outcome and one demanded by the pen-sion movement. The amendment failed, 17–56.But forming a coalition behind a radical alter-native is a way to induce legislators to supportmore moderate legislation that they might nototherwise have favored. We analyze the vote forthe Lee amendment both by logistic regressionand FSQCA techniques. The latter are impor-tant because the mediation theory expects thatmany favorable circumstances are necessary toinfluence this sort of radical outcome.

For the regression analyses, we use many ofthe previous independent measures, adjustedfor the year. For partisanship, however, weemploy the party affiliation of the senator, not-ing whether he or she was a non-poll taxDemocrat or a member of a radical third party.Public opinion in these models is measured bythe state-level support expressed in a December1938 Gallup poll for the so-called Lodge bill.That proposal would have provided for $60monthly pensions for almost all agedAmericans, with $40 being provided by thefederal government. About 65 percent of thoseexpressing an opinion were in favor. We alsoemploy a measure of whether the senator wasendorsed by the Townsend Plan, which urgedclubs to support those whom it endorsed. Thepension measure may be more valid than theendorsement measure, as only a third of thesenators came up for election in 1938. Becauseof the smaller number and the different natureof the cases, we use a modified version of themodel to explain OAA outcomes. In the firstmodel, we include the poll tax and patronageparty measures, as well as the partisanshipmeasure. We also include the control measures

Table 3. Six-Measure FSQCA Results for Generous OAA Benefits

Individual Configurations (and numbers of states)—polltax * patronage * ADMIN * democrat * MOBILIZATION * ASSERTIVE (5) + —polltax * patronage * ADMIN * DEMOCRAT * MOBILIZATION * ASSERTIVE (2) + —polltax * patronage * admin * DEMOCRAT * MOBILIZATION * assertive (1) + —polltax * patronage * ADMIN * DEMOCRAT * mobilization * assertive (1) + —polltax * patronage * ADMIN * democrat * mobilization * ASSERTIVE (1) + —polltax * patronage * admin * DEMOCRAT * mobilization * ASSERTIVE (1)

Reduced Forms of Configurations—polltax * patronage * ADMIN * DEMOCRAT + —polltax * patronage DEMOCRAT * MOBILIZATION + —polltax * patronage * ASSERTIVE

—polltax * patronage * (DEMOCRAT * (ADMIN + MOBILIZATION ) + ASSERTIVE)

Note: For FSQCA notation and definitions of measures, see text and Appendix.


Tue, 13 Nov 2007 04:51:20

for income, percentage aged, and public opinion.8

In a second model, we add the three movementmeasures to see if they add anything to the expla-nation.

Table 4 shows two logistical regression mod-els of the vote in favor of the radical Lee amend-ment. Model 1, including all the non-movement-related measures, explains about 20 percent of thevariance and correctly predicts about 78 percentof the cases. Two of the political contextual meas-ures are significant and in the proper direction,with a negative influence of dominant patronageparties and a positive influence of democratizedDemocratic or third party affiliation on the oddsof voting for senior citizens’ pensions. The con-trol measure public opinion in favor of the Lodgebill has a positive effect, significant at the .10level, whereas the control percent aged has anegative effect.

Model 2 provides a significant improvementof fit. The pseudo R-squared jumps from about.20 to about .40, and the increase in predictivepower is from about 78 percent to about 84 per-cent, which is a shift from predicting 57 of 73cases correctly to 61—or a quarter of the remain-ing cases. Townsend club strength significantlyincreases the likelihood of a senator’s voting forthe bill. So does a pension initiative, though atonly the .10 level. Having a pension initiative ina state makes a senator almost three times morelikely to vote for the Lee amendment, from about9 percent to about 25 percent. The Townsendendorsement has a positive, but insignificant,effect, possibly because only one third of sena-tors were subject to being endorsed. The meas-ure of patronage party strength becomesinsignificant, perhaps because it was exertinginfluence by dampening the pension movement.Also, in the final model the measure of publicopinion is significant at the .05 level and sub-stantively important. A movement from 60 per-cent of the public supporting the Lodge bill to 80percent would increase a senator’s chance of vot-

ing for the Lee amendment from about 7 percentto about 25 percent.9

We now turn to examining combinations ofconditions leading to positive votes throughFSQCA. Again, because most of the independentmeasures are categorical, as is the dependentmeasure, we employ crisp-set analyses. We scorethose in favor as one, and those opposed as zero,and begin our analyses with the same independ-ent measures as before, though this time weinclude whether the senator had a Democratic orradical third party affiliation (DEM/THIRD) andomit the administrative variable, which is notapplicable nationally. Although the mobilizationmeasure is the same, we combine the measure ofinitiatives and endorsements (ASSERTIVE), treat-ing these statewide assertive activities as func-tionally equivalent, to reduce the complexity in theresults. Our expectations here are that it may takeboth high mobilization and assertive action inthe most favorable possible contexts to reach thismore radical result. Because of the low percent-age of positive votes, almost all truth table com-binations that include positive votes are“contradictory,” including one or more negativevotes. Because we are interested in understand-ing the conditions under which it is reasonablylikely for a senator to vote for a radical program,we reduce combinations in which at least half ofthe senators supported the pension amendment.

Three combinations provide the greatest sup-port for the amendment:

polltax*patronage*DEM/THIRD*

MOBILIZATION*ASSERTIVE +

polltax*PATRONAGE*DEM/THIRD*

MOBILIZATION*assertive +

polltax*patronage*dem/third*

mobilization*ASSERTIVE



8 We also included a control for union density—union members in 1939 as a share of the nonagri-cultural employed (see Amenta and Halfmann2000)—because it is often argued that unions spurold-age programs. This measure proved to be insignif-icant in our model (results not shown, but availableon request), and we omitted it.

9 We also engaged in a similar logistical regressionanalysis of voting for the Townsend Plan bill, HR6466, in the House of Representatives that year. HR6466 also called for senior citizens’ pensions for thenonemployed who were over 60 years old, based ona transactions (sales) tax and other taxes, and wasexpected to produce initial benefits of about $60 permonth. The bill failed, 306–101. We did not reportthese results (available upon request) mainly becausethey largely replicate the Senate results.


Tue, 13 Nov 2007 04:51:20

As before, a lack of voting restrictions is a nec-essary condition. The first combination, whichprovides the most positive votes, six, is also theone closely associated with political mediationthinking. It includes all possible favorable con-ditions: a Democratic or third party affiliation,a strongly mobilized pension movement, andassertive action. The second combination sug-gests that high mobilization alone in an alreadyfavorable political context can be influential. Thethird indicates that assertive action will pro-vide an alternative means of exerting influencein less favorable situations. However, the last twocombinations help to identify only three addi-tional senators voting for pensions. The resultssupport political mediation thinking, but alsosuggest that yet other factors than those in thisversion of the model may be needed to identi-

fy legislators most willing to support the pro-grams of state-oriented movements.

CCOONNCCLLUUSSIIOONN

To summarize, our main claim is that thecollective action of state-oriented challengersand their influence on public policy is politicallymediated in specific ways. Challengers controltheir strategies and, to be effective, must be ableto alter them according to political contexts.Under certain political institutional conditions,notably restrictions on democratic practices andthe entrenchment of patronage-oriented politi-cal parties, the impact of state-oriented chal-lengers is likely to be greatly dampened. In thefirst half of the twentieth century, only about halfof the state-level U.S. polities were structural-



Table 4. Voting for the Radical Lee Pension Amendment on Selected Independent Measures

Model 1 Model 2

Institutional MeasuresPoll Tax 0.196 4.602

(0.13) (1.85)Patronage Party Strength –0.477* 0.212

(1.92) (0.57)Non-Poll–Tax Democrat or Third Party Member 1.464 3.220*

(1.56) (2.11)

Movement MeasuresTownsend Club Activity 1.279*

.— (2.28)Pension Proposition 2.018

.— (1.64)Townsend Endorsement 1.035

.— (0.83)Control MeasuresPro-Old-Age Public Opinion 0.045 0.079*

(1.61) (2.09)Per Capita Income –0.001 –0.001

(0.37) (0.41)Percentage Aged –0.428 –1.255*

(1.69) (2.48)

Constant –1.606 –7.966 (0.54) (1.92)

Observations 73 73Pseudo R2 0.20 0.41!2 15.97* 32.24**Df 6 9

Notes: Data shown are unstandardized coefficients from logistical regressions. The absolute values of z statisticsare in parentheses. For definitions of measures, see text and Appendix.* p < .05; ** p < .01 (one-tailed; except for percent aged).


Tue, 13 Nov 2007 04:51:20

ly open to influence. Under these open condi-tions and more favorable conditions over themedium and short term, little more than mobi-lization is needed for social movements to haveinfluence, whereas in less short-term political-ly favorable conditions more assertive action isnecessary. Like scholars of framing, we arguethat the influence of a challenger is likely to beconfined to the issues that it plausibly engages.To achieve radical results, the most favorableconditions, mobilization, and assertive action arerequired.

The results on the development of Old AgeAssistance and on the Senate vote bear out theseclaims. All sets of results also provide somesupport for the views that high mobilization isa key to influence and that strategies matter. Inaddition, Townsend club activity seemed to havea continuing influence on OAA generosity,whereas changes in activity did not seem toinfluence policy-makers similarly. This sug-gests that organization and mobilization maypay longer-term dividends for challengers,though perhaps only so long as the movementas a whole remains viable. Assertive strategiesalso influenced old-age policy. Pension initia-tives had a signif icant influence on OAAstipends. The results also suggest that the stan-dard distinction between institutional and non-institutional and disruptive and assimilativeaction is too broad to address the sorts of col-lective action that matter in political processes.Also, the fact that the pension movement large-ly concerned itself with high benefits led to adifferential influence on OAA generosity andcoverage, with the movement spurring benefitlevels, but not coverage. These results suggestthat movement claims-making can limit theinfluence of challengers and that being flexiblein this area may make wider benefits possiblefor a movement’s constituency.

The formal qualitative results also support themediation idea: that challengers need to matchcollective action strategies to political contexts.Combinations of conditions associated withhigh OAA pensions were as expected by thepolitical mediation model. One of three com-binations included a Democratic regime andTownsend club mobilization, suggesting thatunder short-term favorable circumstances,mobilization was sufficient to bring gains inOAA spending. Another combination indicatedthat when short-term political conditions were

not necessarily favorable, the aggressive strat-egy of initiatives proved sufficient to bringabout high OAA benefits. On the vote to trans-form old-age policy, movement mobilizationand assertive action under favorable conditionsbrought positive results.

Our research should not be interpreted tomean that this or that variable should be expect-ed always to bring influence for challengers orto mean that social movements are usually like-ly to produce policy results. Our point is sim-ply that social movements can be influentialunder certain conditions. It seems likely thatmost movement organizations are not likely tobe highly influential, given that challengers startfrom a position of relative disadvantage in polit-ical power. The pension movement included afairly powerfully organized and mobilized set ofchallengers with widespread support, and theold-age issue was a prominent one in the 1930sand 1940s. The results we report here may per-tain only to the most significant and highlypublicized movements. That said, there is noreason to believe that policy results of the sortthat we find would be confined to a movementbased significantly on one large challengingorganization like the Townsend Plan.Decentralized challengers and coalition-basedmovements combining the same characteristicsmight achieve similar sorts of results.

The results also support the view that publicopinion influences public policy and move-ments can have a further indirect impact onpolicy by influencing the general public aboutits mission, program, or constituency. However,the strong version of the public opinion argu-ment finds little support. Adding public opin-ion measures did not mean that other causesfaded into insignificance. Although data limi-tations made it impossible to appraise directlythe influence of the relative saliency of publicopinion, the old-age issue had a high profile inthe late 1930s, and its saliency was due at leastin part to the pension movement. The Galluppolls taken in late 1935 and early 1936 werelargely a result of the rise of the Townsend Plan,and the polls taken in 1938 and 1939 were in aresponse to a resurgent pension movement(Amenta forthcoming). The results here alsoline up with the idea that the influence of opin-ion polls on political actors may be historical-ly variable and requires explanation in itself. Wehope that promising ideas about the contingent




Tue, 13 Nov 2007 04:51:20

impact of public opinion on policy (Jacobs andShapiro 2000; Burstein and Linton 2002; Manzaand Cook 2002) are set out more explicitly andempirically examined. Scholars of the impact ofsocial movements may need to model the influ-ence of challengers on opinion and from thereon state outcomes.

As the results show, the political mediationmodel, as currently constructed, seems onlypartially adequate to understanding the moreradical and difficult-to-achieve outcomes forsocial movements. Favorable shifts in publicopinion (Burstein 1999; Giugni 2004) or inno-vative framing (Cress and Snow 2000) or gain-ing ground in discursive struggles through themass media (Ferree et al. 2002) may be requiredin addition to factors identified by the politicalmediation theory to achieve fundamentalchanges demanded by challengers.

Also, as studies mount, both from socialmovement scholars and political sociologistsexamining state policy, the impact of move-ments on policy seems to be understood at leastas well as the determinants of mobilization,which seem considerably more controversial(cf. Goodwin and Jasper 1999; McAdam 1999;Ferree and Merrill 2004; Meyer 2004). It is nolonger enough for students of the policy con-sequences of movements to justify their researchas being on a novel subject. What we need aretheoretical refinements and advancements ofmore complex ideas and the types of investiga-tions, whether quantitative, formal qualitative,or historical, that can enable us to appraise the-ory and further our understanding of the influ-ence of movements on political outcomes. Wealso need to address whether political mediationideas are applicable or whether completely newtheorizing is needed to understand the manyattempts at influence by movements that arenot mainly directed at the state (Amenta andYoung 1999; Earl 2004).

Edwin Amenta is Professor of Sociology at New YorkUniversity and the University of California, Irvine.He is the author of When Movements Matter: TheTownsend Plan, the Pension Movement, and theMaking of Social Security (Princeton UniversityPress, forthcoming).

Neal Caren is a Ph.D. candidate in the Departmentof Sociology at New York University. His dissertationexamines urban political participation in modernAmerica. In the fall, he will join the Robert Wood

Johnson Foundation Scholars in Health PolicyResearch Program at the University of Michigan.

Sheera Joy Olasky is a graduate student in theDepartment of Sociology at New York University.Her research interests include political sociology,social movements, and urban sociology. She is cur-rently studying environmental justice and socialpolicy.

AAPPPPEENNDDIIXX

DDEEFFIINNIITTIIOONNSS,, DDAATTAA,, SSOOUURRCCEESS,, AANNDD

QQUUAANNTTIITTAATTIIVVEE MMEEAASSUURREESS

DEPENDENT MEASURES

Average Size of the OAA BenefitThe average OAA benefit in each state for

each year in 1950 dollars (U.S. Social SecurityBoard/Administration 1935–1950; U.S. Bureauof Labor Statistics 2004).

OAA CoverageThe percentage of each state’s residents over

the age of 65 years (Gardner and Cohen 1992)who received OAA benefits in a given year(U.S. Social Security Board/Administration1935–1950).

Radical Lee Pension AmendmentThe roll call of the Senate vote S76-1061, the

Lee amendment to the 1939 Social SecurityAct Amendments (Rosenthal and Poole 2000):one for each senator in favor, zero for opposed,and others excluded from the analysis.

INDEPENDENT MEASURES

Poll TaxA dichotomous measure that takes a value of

one for a period when states had a poll tax andzero for a period without a poll tax (Ogden1958).

Patronage Party Organizational StrengthA time-invariant expert-judgment measure

of the degree to which each state’s political par-ties had substantial autonomy, were long-last-ing and hierarchical, regularly attempted tonominate candidates, and relied on materialincentives to engage people to do organiza-tional work, ranging from five in states wherepatronage party organizations predominated toone in states where party organizations had lit-tle control (Mayhew 1986).

Administrative StrengthA time-invariant and dichotomous measure

of the strength and structure of the state gov-ernment that takes a value of one for states




Tue, 13 Nov 2007 04:51:20

where the labor commissioner had rule-makingauthority over safety laws and zero where thecommissioner did not (Brandeis 1935).

Democratized Democratic ControlA dichotomous measure for control in each

year of the governor’s mansion and both hous-es of the state legislature by the Democraticparty (Burnham 1992) in non–poll-tax states(Ogden 1958).

Non-Poll-Tax Democrat or Member ofRadical Third Party

A measure scoring one for senators fromnon–poll-tax states (Ogden 1958) andDemocrats or members of a third party(Rosenthal and Poole 2000) with others scoringzero.

OAA County FundingThe percentage of total OAA funding coun-

ties contributed in states by year (U.S. SocialSecurity Board/Administration 1935–1950; U.S.Social Security Board 1945).

Per Capita IncomeState income per person in each year, meas-

ured in 1950 dollars (U.S. Bureau of EconomicAnalysis 2001).

Percentage BlackThe percentage of state residents who are

African American in each year, based on thedecennial censuses in 1930, 1940, and 1950(Gardner and Cohen 1992), with values forintercensal years computed based on a lineargrowth for each state.

Percentage AgedThe percentage of state residents over 65

years of age in each year based on the decen-nial censuses in 1930, 1940, and 1950 (Gardnerand Cohen 1992), with values for intercensalyears computed based on a linear growth foreach state.

Pro-Old-Age Public OpinionThree measures from Gallup polls in 1938 in

which individual responses to the followingquestions were aggregated to provide state-levelestimates. Open-ended answers to “How muchper month should be paid to a single person?”are averaged in our analyses of OAA generos-ity (American Institute of Public Opinion1938a). For coverage, we compute the percent-age of respondents who answered “All” to thequestion “Do you think pensions should begiven to all old people, or only to old people whoare in need?” (American Institute of PublicOpinion 1938b). For the Lee vote, we compute

the percentage of respondents in favor of theLodge bill, which would have established anational $60 per month pension, with $40 com-ing from federal funds (American Institute ofPublic Opinion 1938b).

Townsend Club ActivityA measure of mobilization based on a con-

tent analysis of a sample of eight issues of theTownsend Weekly by year, according to men-tions of Townsend club activity by state, adjust-ed by yearly national membership (Holtzman1963), ranging from one to five, with five indi-cating a level of activity that would place it inthe top 20 percent of all state-years during thisperiod. Club activity is measured in two ways:frozen at its peak value once reached, andaccording to the change in activity betweenyears.

Townsend EndorsementA dichotomous measure scoring one for sen-

ators endorsed by the Townsend Plan in the1938 Senate election (Townsend NationalWeekly 1938), with others scored zero.

Pension PropositionA dichotomous measure of the presence in a

given state and year of a ballot proposition orinitiative designed to increase pension amounts,coverage, or both, constructed from contempo-rary accounts in the New York Times, theWashington Post, the Los Angeles Times, and theChristian Science Monitor.

RREEFFEERREENNCCEESS

Almeida, Paul and Linda Brewster Stearns. 1998.“Political Opportunities and Local GrassrootsEnvironmental Movements: The Case ofMinamata.” Social Problems 45:37–60.

Amenta, Edwin. 1998. Bold Relief: InstitutionalPolitics and the Origins of Modern AmericanSocial Policy. Princeton, NJ: Princeton UniversityPress.

———. Forthcoming. When Movements Matter:The Townsend Plan, the Pension Movement, andthe Making of Social Security. Princeton, NJ:Princeton University Press.

Amenta, Edwin, Mary Bernstein, and KathleenDunleavy. 1994. “Stolen Thunder? Huey Long’sShare Our Wealth, Political Mediation, and theSecond New Deal.” American Sociological Review59:678-702.

Amenta, Edwin, Chris Bonastia, and Neal Caren.2001. “US Social Policy in Comparative andHistorical Perspective: Concepts, Images,Arguments, and Research Strategies.” AnnualReview of Sociology 27:213–34.




Tue, 13 Nov 2007 04:51:20

http://www.ingentaconnect.com/content/external-references?article=0037-7791()45L.37%5Baid=897434%5D



Amenta, Edwin and Neal Caren. 2004. “TheLegislative, Organizational, and BeneficiaryConsequences of State-Oriented Challengers.” Pp.461–88 in The Blackwell Companion to SocialMovements, edited by David A. Snow, Sarah A.Soule, and Hanspeter Kriesi. Malden, MA:Blackwell Publishing.

Amenta, Edwin, Bruce G. Carruthers, and YvonneZylan. 1992. “A Hero for the Aged? The TownsendMovement, the Political Mediation Model, andU.S. Old-Age Policy, 1934–1950.” AmericanJournal of Sociology 98:308–39.

Amenta, Edwin and Drew Halfmann. 2000. “WageWars: Institutional politics, WPA Wages, and theStruggle for U.S. Social Policy.” AmericanSociological Review 65:506–28.

Amenta, Edwin, Drew Halfmann, and Michael P.Young. 1999. “The Strategies and Contexts ofSocial Protest: Political Mediation and the Impactof the Townsend Movement.” Mobilization56:1–25.

Amenta, Edwin and Michael P. Young. 1999. “Makingan Impact: The Conceptual and MethodologicalImplications of the Collective Benefits Criterion.”p. 22–41 in How Movements Matter: Theoreticaland Comparative Studies on the Consequences ofSocial Movements, edited by Marco Giugni, DougMcAdam, and Charles Tilly. Minneapolis, MN:University of Minnesota Press.

American Institute of Public Opinion. 1938a. “GallupPoll no. 130.” Princeton, NJ: American Institute ofPublic Opinion [producer]. New Haven, CT: RoperCenter for Public Opinion Research [distributor].

———. 1938b. “Gallup Poll no. 140.” Princeton,NJ: American Institute of Public Opinion [pro-ducer]. New Haven, CT: Roper Center for PublicOpinion Research [distributor].

Andrews, Kenneth T. 2001. “Social Movements andPolicy Implementation: The Mississippi CivilRights Movement and the War on Poverty,1965–1971.” American Sociological Review66:21–48.

Baltagi, Badi H. and Ping X. Wu. 1999. “UnequallySpaced Panel Data Regressions with AR(1)Disturbances.” Econometric Theory 15:814–23.

Beck, Nathaniel. 2001. “Time-Series-Cross-SectionData: What Have We Learned in the Past FewYears?” Annual Review of Political Science4:271–93.

Beckfield, Jason. 2003. “Inequality in the WorldPolity: The Structure of InternationalOrganization.” American Sociological Review68:401–24.

Brandeis, Elizabeth. 1935. “Labor Legislation.” Pp.399–700 in History of Labor in the United States,1896–1932, edited by John R. Commons. NewYork: Macmillan.

Burnham, Walter Dean. 1974. “Theory and VotingResearch: Some Reflections on Converse’s

‘Change in the American Electorate.’” AmericanPolitical Science Review 68:1002–23.

———. 1992. “Partisan Division of American StateGovernments, 1834–1985.” [Computer File] AnnArbor, MI: Inter-university Consortium forPolitical and Social Research.

Burstein, Paul. 1999. “Social Movements and PublicPolicy.” p. 3–21 in How Social Movements Matter,edited by Marco Giugni, Doug McAdam, andCharles Tilly. Minneapolis, MN: University ofMinnesota Press.

Burstein, Paul and April Linton. 2002. “The Impactof Political Parties, Interest Groups, and SocialMovement Organizations on Public Policy: SomeRecent Evidence and Theoretical Concerns.”Social Forces 81:380–408.

Cantril, Hadley and Mildred Strunk. 1951. PublicOpinion, 1935–1946. Princeton, NJ: PrincetonUniversity Press.

Cress, Daniel M. and David A. Snow. 2000. “TheOutcomes of Homeless Mobilization: TheInfluence of Organization, Disruption, PoliticalMediation, and Framing.” American Journal ofSociology 105:1063–104.

Earl, Jennifer. 2000. “Methods, Movements andOutcomes: Methodological Difficulties in theStudy of Extra-Movement Outcomes.” Research inSocial Movements, Conflicts, and Change 22:3–25.

———. 2004. “The Cultural Consequences ofSocial Movements.” Pp. 508–30 in The BlackwellCompanion to Social Movements, edited by DavidA. Snow, Sarah A. Soule, and Hanspeter Kriesi.Malden, MA: Blackwell Publishing.

Ferree, Myra Marx, William Anthony Gamson,Jurgen Gerhards, and Dieter Rucht. 2002. ShapingAbortion Discourse: Democracy and the PublicSphere in Germany and the United States.Cambridge, UK: Cambridge University Press.

Ferree, Myra Marx and David Merrill. 2004. “HotMovements, Cold Cognition: Thinking aboutSocial Movements in Gendered Frames.” Pp.247–61 in Rethinking Social Movements: StructureMeaning and Emotion, edited by Jeff Goodwin andJames M. Jasper. Lanham, MD: Rowan andLittlefield.

Fording, Richard C. 1997. “The Conditional Effectof Violence as a Political Tactic: Mass Insurgency,Welfare Generosity, and Electoral Context in theAmerican States.” American Journal of PoliticalScience 41:1–29.

Gallup, George. 1935. “Americans Reveal TheirHopes and Fears in Nation-Wide Poll.” WashingtonPost, December 15, p. B1.

———. 1936. “Nation 9 to 1 for Old Age Pensions;Against Townsend Plan.” Washington Post, January12, p. B1.

———. 1939. “Political Strength of PensionMovements Growing, Survey Shows.” WashingtonPost, February 26, pp. B2–3.




Tue, 13 Nov 2007 04:51:20









Gamson, William A. [1975] 1990. The Strategy ofSocial Protest. 2d ed. Belmont, CA: Wadsworth.

Ganz, Marshall. 2000. “Resources andResourcefulness: Strategic Capacity in theUnionization of California Agriculture,1959–1966.” American Journal of Sociology105:1003–62.

Gardner, John and William Cohen. 1992.“Demographic Characteristics of the Population ofthe United States, 1930–1950.” [Computer File]Ann Arbor, MI: Inter-university Consortium forPolitical and Social Research.

Giugni, Marco. 2004. Social Protest and PolicyChange: Ecology, Antinuclear, and PeaceMovements in Comparative Perspective. Lanham,MD: Rowan and Littlefield.

Goodwin, Jeff and James M. Jasper. 1999. “Caughtin a Winding, Snarling Vine: The Structural Biasof Political Process Theory.” Sociological Forum14:27–54.

Hardin, Russell. 1982. Collective Action. Baltimore,MD: Johns Hopkins University Press.

Hicks, Alexander M. 1999. Social Democracy andWelfare Capitalism: A Century of Income SecurityPolitics. Ithaca, NY: Cornell University Press.

Holtzman, Abraham. 1963. The Townsend Movement:A Political Study. New York: Bookman Associates.

Huber, Evelyne and John D. Stephens. 2001.Development and Crisis of the Welfare State:Parties and Policies in Global Markets. Chicago,IL: University of Chicago Press.

Huston, Luther A. 1939. “Congress Wind-Up on orNear July 15 Again Is Predicted.” New York Times,July 3, p. 1.

Jacobs, Lawrence R. and Robert Y. Shapiro. 2000.Politicians Don’t Pander: Political Manipulationand the Loss of Democratic Responsiveness.Chicago, IL: University of Chicago Press.

Jenkins, J. Craig and Charles Perrow. 1977.“Insurgency of the Powerless: Farm WorkerMovements (1946–1972).” American SociologicalReview 42:249–68.

Katznelson, Ira. 1981. City Trenches: Urban Politicsand the Patterning of Class in the United States.New York: Pantheon Books.

Kenworthy, Lane. 2002. “Corporatism andUnemployment in the 1980s and 1990s.” AmericanSociological Review 67:367–88.

Kingdon, John W. 1984. Agendas, Alternatives, andPublic Policies. Boston, MA: Little, Brown.

Kitschelt, Herbert P. 1986. “Political OpportunityStructures and Political Protest: Anti-NuclearMovements in Four Democracies.” British Journalof Political Science 16:57–85.

Klandermans, Bert and Dirk Oegema. 1987.“Campaigning for a Nuclear Freeze: Grass-RootsStrategies and Local Governments in the

Netherlands.” Research in Political Sociology3:305–37.

Kriesi, Hanspeter. 1995. “The Political OpportunityStructure of New Social Movements: Its Impact onTheir Mobilization.” Pp. 167–98 in The Politics ofSocial Protest: Comparative Perspectives on Statesand Social Movements, edited by J. Craig Jenkinsand Bert Klandermans. Minneapolis, MN:University of Minnesota Press.

Lieberman, Robert C. 1998. Shifting the Color Line:Race and the American Welfare State. Cambridge,MA: Harvard University Press.

Lipset, Seymour Martin and Gary Wolfe Marks.2000. It Didn’t Happen Here: Why SocialismFailed in the United States. New York: W. W.Norton and Co.

Manza, Jeff and Fay Lomax Cook. 2002. “The Impactof Public Opinion on Public Policy: The State ofthe Debate.” Pp. 17–32 in Navigating PublicOpinion: Polls, Policy, and the Future of AmericanDemocracy, edited by Jeff Manza, Fay LomaxCook, and Benjamin I. Page. New York: OxfordUniversity Press.

Mayhew, David R. 1986. Placing Parties in AmericanPolitics: Organization, Electoral Settings, andGovernment Activity in the Twentieth Century.Princeton, NJ: Princeton University Press.

McAdam, Doug. 1996. “Introduction: Opportunities,Mobilizing Structures, and Framing Processes—Toward a Synthetic, Comparative Perspective onSocial Movements.” Pp. 1–22 in ComparativePerspectives on Social Movements: PoliticalOpportunities, Mobilizing Structures, and CulturalFramings, edited by Doug McAdam, John D.McCarthy, and Mayer N. Zald. New York:Cambridge University Press.

———. 1999. Political Process and theDevelopment of Black Insurgency, 1930–1970. 2ded. Chicago, IL: University of Chicago Press.

McAdam, Doug, John D. McCarthy, and Mayer N.Zald. 1988. “Social Movements.” Pp. 695–798 inThe Handbook of Sociology, edited by Neil J.Smelser. Newbury Park, CA: Sage.

McAdam, Doug and Yang Su. 2002. “The War atHome: Antiwar Protests and Congressional Voting,1965–1973.” American Sociological Review67:696–721.

McCarthy, John D. and Clark McPhail. 1998. “TheInstitutionalization of Protest in the United States.”Pp. 83–110 in The Social Movement Society:Contentious Politics for a New Century, edited byDavid S. Meyer and Sidney G. Tarrow. Lanham,MD: Rowan and Littlefield.

McCarthy, John D. and Mayer N. Zald. 2002. “TheEnduring Vitality of the Resource MobilizationTheory of Social Movements.” Pp. 533–65 inHandbook of Sociological Theory, edited by




Tue, 13 Nov 2007 04:51:20







Jonathon H. Turner. New York: KluwerAcademic/Plenum Publishers.

Meyer, David S. 2005. “Social Movements and PublicPolicy: Eggs, Chicken, and Theory.” Chapter 1 inRouting the Opposition: Social Movements, PublicPolicy, and Democracy, edited by David S. Meyer,Valerie Jenness, and Helen M. Ingram.Minneapolis, MN: University of Minnesota Press.

———. 2004. “Protest and Political Opportunities.”Annual Review of Sociology 30:125–45.

Meyer, David S. and Debra C. Minkoff. 2004.“Conceptualizing Political Opportunity.” SocialForces 82:1457–92.

Mitchell, Daniel J. B. 2000. Pensions, Politics, andthe Elderly: Historic Social Movements and TheirLessons for Our Aging Society. Armonk, NY: M.E. Sharpe.

Ogden, Frederic D. 1958. The Poll Tax in the South.Tuscaloosa, AL: University of Alabama Press.

Piven, Frances Fox and Richard A. Cloward. 1977.Poor People’s Movements: Why They Succeed,How They Fail. New York: Pantheon Books.

Quadagno, Jill S. 1988. “From Old-Age Assistanceto Supplemental Security Income: The PoliticalEconomy of Relief in the South, 1935–1972.” Pp.235–64 in The Politics of Social Policy in theUnited States, edited by Margaret Weir, Ann SholaOrloff, and Theda Skocpol. Princeton, NJ:Princeton University Press.

Ragin, Charles C. 1987. The Comparative Method:Moving Beyond Qualitative and QuantitativeStrategies. Berkeley, CA: University of CaliforniaPress.

———. 2000. Fuzzy-Set Social Science. Chicago,IL: University of Chicago Press.

Rosenthal, Howard L. and Keith T. Poole. 2000.“United States Congressional Roll Call VotingRecords, 1789–1990: Reformatted Data.” CarnegieMellon University, Graduate School of IndustrialAdministration; Ann Arbor, MI: Inter-universityConsortium for Political and Social Research.

Rucht, Dieter. 1998. “The Structure and Culture ofCollective Protest in Germany since 1950.” Pp.29–57 in The Social Movement Society:Contentious Politics for a New Century, edited byDavid S. Meyer and Sidney G. Tarrow. Lanham,MD: Tarrow, Rowan and Littlefield.

Schiltz, Michael A. 1971. Public Attitudes towardSocial Security, 1935–1965. Washington, DC: U.S.Government Printing Office.

Skocpol, Theda. 1992. Protecting Soldiers andMothers: The Political Origins of Social Policy inthe United States. Cambridge, MA: Belknap Pressof Harvard University Press.

———. 2003. Diminished Democracy: FromMembership to Management in American CivicLife. Norman, OK: University of Oklahoma Press.

Skocpol, Theda, and Edwin Amenta. 1986. “Statesand Social Policies.” Annual Review of Sociology12:131–57.

Skrentny, John D. 2002. The Minority RightsRevolution. Cambridge, MA: Harvard UniversityPress.

Soule, Sarah A., Doug McAdam, John McCarthy, andYang Su. 1999. “Protest Events: Cause orConsequence of State Action? The U.S. Women’sMovements and Federal Congressional Activities,1956–1979.” Mobilization 4:239–56.

Soule, Sarah A. and Susan Olzak. 2004. “When DoMovements Matter? The Politics of Contingencyand the Equal Rights Amendment.” AmericanSociological Review 69:473–97.

Tarrow, Sidney G. 1996. “States and Opportunities:The Political Structuring of Social Movements.”Pp. 41–61 in Comparative Perspectives on SocialMovements: Opportunities, Mobilizing, Structures,and Cultural Framings, edited by Doug McAdam,John D. McCarthy, and Mayer N. Zald. New York:Cambridge University Press.

Tilly, Charles. 1999a. “Regimes and Contention.”Columbia International Affairs Online WorkingPaper. Manuscript.

———. 1999b. “Conclusion: From Interactions toOutcomes in Social Movements.” Pp. 253–70 inHow Movements Matter: Theoretical andComparative Studies on the Consequences ofSocial Movements, edited by Marco Giugni, DougMcAdam, and Charles Tilly. Minneapolis, MN:University of Minnesota Press.

Townsend National Weekly. 1938–1950. TownsendNational Weekly, Chicago, IL.

U.S. Bureau of the Census. Historical Statistics of theUnited States, Colonial Times to 1970.Bicentennial ed. U.S. Department of CommerceBureau of the Census. Washington, DC: U.S.Government Printing Office.

U.S. Bureau of Economic Analysis. 2001. “StatePersonal Income 1929–2000” [Computer File].Washington, DC: U.S. Department of Commerce,Economics and Statistics Administration.

U.S. Social Security Board/Administration.1935–1950. Annual Report of the Federal SecurityAgency. Washington, DC: U.S. GovernmentPrinting Office.

U.S. Social Security Board. 1945. Social SecurityYearbook. Washington, DC: U.S. GovernmentPrinting Office.

Weakliem, David L. 2003. “Public Opinion Researchand Political Sociology.” Research in PoliticalSociology 12:49–80.

Western, Bruce and Katherine Beckett. 1999. “HowUnregulated Is the U.S. Labor Market? The PenalSystem as a Labor Market Institution.” AmericanJournal of Sociology 104:1030–60.




Tue, 13 Nov 2007 04:51:20



http://www.ingentaconnect.com/content/external-references?article=1086-671x()4L.239%5Baid=4953738%5D





Wilensky, Harold L. 1975. Welfare State andEquality: Structural and Ideological Roots ofPublic Expenditures. Berkeley, CA: University ofCalifornia Press.

Wooldridge, Jeffrey. 2002. Econometric Analysis of

Cross Section and Panel Data. Cambridge, MA:MIT Press.

Zald, Mayer N. 2000. “Ideologically StructuredAction: An Enlarged Agenda for Social MovementResearch.” Mobilization 5:31–36.




Tue, 13 Nov 2007 04:51:20

http://www.ingentaconnect.com/content/external-references?article=1086-671x()5L.31%5Baid=5283003%5D

a ge for l eisu re? p olitical m ed iation an d th e im p ...ea3/age for leisure asr 2005.pdf ·...

Documents