¹ For full accounts, see Pina, Antonio López, ‘The Shaping of the Constitution’, in Penniman, Howard, ed., Spain at the Polls (Washington, D.C.: American Enterprise Institute, 1980)Google Scholar; and Pina, López, ed., La España Democrática y Europa (Madrid: Editorial Cambio 16, 1977).Google Scholar

² This paragraph is a condensation of complex historical changes treated at greater length by Ortí, Alfonso, ‘Política y Sociedad en el Umbral de los Años Setenta: Las Bases Sociales de la Modernizatión Politica’, in Cuadrado, M. Martinez, ed., Cambio Social y Modernización Politico (Madrid: Cuadernos para el Dialogo, 1970)Google Scholar and Pina, Antonio López, ‘Spain: An Anti-Model’, paper presented at the meeting of the Latin American Studies Association, Madison, 04 1973.Google Scholar

³ See Pina, Antonio López, ‘Desarrollo y Política en España’, in Sociologia Española de los Años Setenta (Madrid: Confederación de Cajas de Ahorro, 1970).Google Scholar

⁴ See Harrison, Joseph, An Economic History of Modern Spain (Cambridge: Cambridge University Press, 1978)Google Scholar; , Angel Martinez G.-Tablas, ‘Capital Extranjero y Oligarquia en la Crisis Económica’, El Carabo, XIV (1980), 41–71Google Scholar; Muñoz, Juan, Roldan, Santiago and Serrano, Angel, La Internationalización del Capital en España (Madrid: Cuadernos para el Dialogo, 1978)Google Scholar, and by the last three authors, ‘Les Mutations Economiques dans la Periode de Transition’, Le Monde Diplomatique (09 1979), 13–14.Google Scholar

⁵ Data are drawn from interviews conducted in June 1978 with a national sample of 3,004 Spaniards aged 16 years and older, and in December 1979 to February 1980 with 3,014 Spaniards in the same age group. Fieldwork was carried out by the firm Consulta of Madrid. The surveys were designed as a panel, with a cross-section added to the second wave to compensate for respondent mortality. The present report treats the two surveys as successive cross-sections. Truly dynamic analysis will be incorporated in subsequent articles.

⁶ For helpful compilations of social changes in Spain during the postwar period, see de Miguel, Amando, 40 Miliones de Españoles 40 Años Despues (Barcelona: Ediciones Grijalbo, 1976)Google Scholar, and de Miguel, , Recursos Humanos, Clases y Regiones en España (Madrid: Cuadernos para el Dialogo, 1977)Google Scholar. For an incisive yet historically rich overview, see Giner, Salvador, Continuity and Change: The Social Stratification of Spain (London: University of Reading, 1968)Google Scholar. In the light of the social mobilization evident in Spain particularly since the sixties, a considerable debate has grown up around the ‘top down’ as compared to the ‘bottom up’ forces behind the current political transition. For judicious analyses, see Maravall, Jose, Dictatorship and Political Dissent (London: Tavistock, 1978)Google Scholar, and Maravall, , ‘Transition a la Democracia: Alineamientos Politicos y Elecciones en España’, Sistema, XXXVI (1980), 65–105Google Scholar. Without denying the importance of labour militancy and the mobilization of mass support by leftist parties for a political amnesty in the early post-1975 period, a pivotal factor contributing to the relative success of the transition seems to have been the widespread feeling among political and economic leaders that authoritarian rule had outlived its usefulness in promoting aggregate growth and was, indeed, exacerbating labour unrest. The attraction of a democratic opening was enhanced by the political requirements for entrance into the EEC – a geopolitical contingency absent in, for example, the Brazilian case. See Kurth, James R., Spanish Public in Political Transition ‘Political Consequences of the Product Cycle: Industrial History and Political Outcomes’, International Organization, XXXIII (1979), 1–34CrossRefGoogle Scholar and Yglesias, Jose, The Franco Years (Indianapolis: Bobbs-Merrill, 1977).Google Scholar

⁷ See Pintor, Rafael Lopez and Buceta, Ricardo, Los Españoles de los Años 70 (Madrid: Editorial Tecnos, 1975)Google Scholar; Ridruejo, Dionisio, Escrito en España, 2nd edn. (Buenos Aires: Losada, 1964)Google Scholar; Payne, Stanley G., Falange: A History of Spanish Fascism (Stanford, Calif.: Stanford University Press, 1961)Google Scholar and Payne, , ‘Political Ideology and Economic Modernization in Spain’, World Politics, XXV (1972), 155–81CrossRefGoogle Scholar. The actual extent of mass political indifference in Spain during the various phases of the Franco period remains a matter of controversy. Compare, for example, Pina, Antonio López and Aranguren, Eduardo L., La Cultura Politica de la España de Franco (Madrid: Taurus, 1976)Google Scholar, and Guzmán, Eduardo Sevilla and de San Julián, Salvador Giner, Absolutismo Despótico y Dominación de Clase: El Caso de España (London: University of Reading, 1975).Google Scholar

⁸ See Almerich, Paulina et al. , Cambio Social y Religion en España (Barcelona: Editorial Fontanella, 1975)Google Scholar, and Blanco, José Jimenez et al. , La Consciencia Regional en España (Madrid: Centro de Investigaciones Soeiologicas, 1977).Google Scholar

⁹ Our objective, it should be noted, is not to elaborate a full-blown theory about the nature of political transition in Spain, or anywhere else. Although the Franco regime itself, especially during the post-1950 period, has been the focus of much scholarly attention especially in terms of the distinctive characteristics of developmental dictatorships, very little agreement has emerged about the transformation of authoritarian regimes into more democratic systems. For tentative explorations of this problem, scattered throughout a study of the reverse phenomenon of the decay of democracy into authoritarian rule, see Linz, Juan J., Crisis, Breakdown, and Re-equilibration (Baltimore: Johns Hopkins University Press, 1978)Google Scholar; compare Schmitter, Philippe C., ‘Authoritarian Experiences and the Prospects for Democracy’ (unpublished manuscript, University of Chicago, 1979).Google Scholar

¹⁰ Although the two are sometimes difficult to distinguish, these religious values stem in part from the historical association between the institutions of the monarchy and the Church in Spain, and in part from a cultural association between traditional piety and respect for hierarchy. Plainly, however, religious sentiment and the cultural weight of the King are not the only factors contributing to popular support for and acquiescence to the emerging political system in Spain. The Spanish social security system, developed under Franco, has almost certainly worked to dampen some potential political disruption through the allocation of pension funds, unemployment compensation, and so on. This linkage remains largely unexplored; but see Lancaster, Thomas D., ‘Toward an Assessment of the Spanish Social Security System’, paper presented at the Fourth European Studies Conference, Omaha, 10 1979.Google Scholar

¹¹ See Linz, Juan J. and de Miguel, Jesus M., ‘Hacia un Análisis Regional de las Elecciones de 1936 en España’, Revista Española de la Opinión Publica, XLVIII (1977), 27–68Google Scholar; Tezanos, José Felix, ‘Análisis Sociopolítico del Voto Socialista en las Elecciones de 1979’, Sistema, XXXI (1979), 105–21Google Scholar, and Tusell, Javier, Las Elecciones del Frente Popular (Madrid: Cuadernos para el Dialogo, 1979).Google Scholar

¹² The fact that the left–right cleavage is not utterly polarizing, and that the bases of support for the left are differentiated, should not be taken to signify a harsh discontinuity between the present period and the 1930s in Spain. For example, several analysts have documented correlations on the order of 0·50 between the leftist vote during the thirties and the corresponding vote in the recent elections of 1977 and 1979; the correlations between past and present votes for centre and right-wing parties are not much lower. See Maravall, José Maria, ‘La Alternativa Socialista: La Política y el Apoyo Electoral de PSOE’, Sistema, XXXV (1980), 3–48Google Scholar, and Linz, Juan J., ‘Il Sistema Partitico Spagnolo’, Rivista Italiana di Scienza Politica, III (1978), 365–414Google Scholar. Such continuity is impiessive, even if the correlations must be treated with caution in so far as they are based on small numbers of observations (usually less than the fifty provinces of Spain). Three points should be kept in mind. Firstly, the correlations are ecological and can therefore be no more than suggestive about levels of class consciousness and conflict. Secondly, strong negative correlations between the votes for leftist and rightist parties, whether synchronic or over time, do not necessarily or even typically indicate polarization. This is particularly the case if the number of parties or coalitions of parties is small. In the bi-party instance, the correlation between the votes for each of the parties is statistically bound to be – 1·0, and yet this can hardly be considered a demonstration of acute confrontation. Thirdly, intergenerational continuity in mass political currents can be fairly high, as it is in Spain and other parts of Europe, at the same time that the bases of electoral support become differentiated and more complex. In addition to the social mobility noted earlier, one simple indicator of probable electoral complexity in Spain is that the sheer size of the Spanish electorate has doubled from Republican times to the present: from 13·6 million to 26·8 million. For an excellent discussion of generational transmission in the formation of class loyalties, see Stephens, John D., ‘Class Formation and Class Consciousness: A Theoretical and Empirical Analysis with Reference to Britain and Sweden’, British Journal of Sociology, XXX (1979), 389–414.CrossRefGoogle Scholar

¹³ By ‘conventional measures of class’ we essentially mean standard indicators of occupation, one version of which is used in the present analysis. Some highly innovative attempts to reformulate such operationalizations from a Marxist perspective have appeared recently; see, for example, Wright, Erik Olin, ‘Class and Occupation’, Theory and Society, IX (1980), 177–214CrossRefGoogle Scholar, and Kalleberg, Arne L. and Griffin, Larry J., ‘Class, Occupation, and Inequality in Job Rewards’, American Journal of Sociology, XXCV (1980), 731–68CrossRefGoogle Scholar. One aspect of this approach which seems compatible with ours is the stress on the ‘non-automatic’ nature of occupation as a determinant of political attitudes and action, and the concomitant emphasis on formal organizations – that is, trade unions – as reinforcers of something approaching class-conscious behaviour.

¹⁴ The question reads: ‘For many people political attitudes are of the left or of the right. This is a scale that goes from left to right.’ (Respondents were then handed a card on which the scale was printed, with ten boxes separating positions on the continuum.) ‘Thinking of your own political attitudes, where would you situate yourself?’ The respondents then checked one of the boxes. Much later in the interview, respondents were asked: ‘When you were growing up, that is, between the ages of 10 and 15, how would you classify your father politically on the left–right scale? And your mother?’ The questions about parental left–right orientation were not asked in the second survey. It should be noted that the mean left–right placement of the respondents themselves, 4·4, is identical across the two waves of the panel.

¹⁵ Conceivably, non-response was brought on by reticence rather than ignorance. But ancillary evidence suggests that virtually all the ‘don't knows’ stem from lack of political information, not from circumspection. Firstly, we received almost no reports from the field of respondents refusing to answer questions out of fear, the exception being respondents in the Basque country. Secondly, an examination of non-response in a survey conducted in authoritarian Brazil has shown that, even in this situation, most non-response was genuine; for large portions of the Brazilian population the government (and politics) is distant and unfamiliar. Indeed, non-response for reasons of discretion was slightly higher among the better educated and economically well-off, who were better able to understand ‘the politics’ of the questionnaire, than the less educated. See Cohen, Youssef, Popular Support for Authoritarian Governments: Brazil under Medici (unpublished Ph.D. dissertation, University of Michigan, 1979)Google Scholar. Thirdly, Fraser encounted no reluctance to talk about politics when interviewing survivors of the Spanish Civil War during the last years of Franco's rule; see Fraser, Ronald, Blood of Spain: An Oral History of the Spanish Civil War (New York: Pantheon, 1979)Google Scholar. Finally, one of the few Spanish studies that takes a careful look at the incidence of non-response to a political survey (of 3,400 workers and 800 white-collar employees in 1978) demonstrates that the frequency of ‘don't knows’ soars among the least educated, and that it is largely a function of lack of information. See Diaz, Victor M. Perez, ‘Orientaciones Politicas de los Obreros Españoles Hoy’, Sistema, XXIX–XXX (1979), 159–79.Google Scholar

¹⁶ The fact that over a quarter of the respondents fail to place themselves on the left–right scale does not make Spain exceptional by cross-national standards. In the early seventies, the comparable figure for the United States was 32 per cent. The incidence of non-response to the left-right scale in Spain seems closest to the situation in Italy, where non-response reaches 26 per cent. What is striking about the Spanish case is the high frequency of non-response relative to the popular wisdom about the salience of the left-right cleavage in Spain. The figure for the United States is not surprising, given the comparatively ‘alien’ nature of left–right labels in American political culture. Data for the cross-national comparison are drawn from the eight-nation study conducted by Barnes, Samuel H., Kaase, Max et al. , partial results for which were reported in Political Action (Beverly Hills: Sage, 1979).Google Scholar

¹⁷ These final predictors were selected after a multi-variate analysis using an assortment of social background factors, in which the dependent variable was a dichotomy with one category including those who did not respond to the left–right scale and the other all those who did.

¹⁸ The correlation (Pearson's r) between sex and religiosity is 0·27; between sex and education, it is 0·17. While the difference in political awareness between males and females is not peculiarly Spanish, the widening gap at the lower levels of education may mirror the effects of the moral curriculum (educacion feminina) used in primary schools during the Franco era. These instructional materials were developed explicitly for female students. See de Miguel, Amando, La Piramide Social Española (Madrid: Editorial Ariel, 1977)Google Scholar and Threfall, Monica, ‘El Socialismo y el Electorado Feminino’, Sistema, XXXII (1979), 19–33.Google Scholar

¹⁹ Despite the tendency for less informed and apparently apolitical Spaniards to display conservative leanings, we think it is useful to maintain a distinction between ‘apathy defaulting to conservatism’ and ‘pure’ apoliticism. The inclination of occupationally marginal individuals with low levels of political information to follow conservative rather than progressive cues has been reasonably well-documented in various countries – for example, in France, as le marais and in Spain as the ‘neutral masses’. See Deutsch, Emeric, Lindon, Denis and Weill, Pierre, Les Familles Politiques Aujourd'hui en France (Paris: Editions de Minuit, 1966)Google Scholar and Robinson, R. A. H., ‘Political Conservatism: The Spanish case, 1875–1977’, Journal of Contemporary History, XIV (1979), 561–80CrossRefGoogle Scholar. Nevertheless, to equate what appears as political indifference with ‘objective’ reaction would overstate the relationship. The strength of a conservatism without a solid informational and organizational basis is dubious. Under certain circumstances, otherwise conservative strata may be susceptible to a leftist populism. See Schoultz, Lars, ‘The Socio-Economic Determinants of Popular-Authoritarian Electoral Behaviour: The Case of Peronism’, American Political Science Review, LXXI (1977), 1423–46.CrossRefGoogle Scholar

²⁰ Multiple Classification Analysis was used to produce the results. The technique is analogous to multiple regression analysis, with the added capacity to handle nominal scale predictors (such as region). In addition to the beta weights, comparable to regression coefficients, MCA generates mean values of the dependent variable for each category of each predictor, after correcting for the effects of all other predictors. See Andrews, Frank M. et al. , Multiple Classification Analysis (Ann Arbor: Institute for Social Research, 1973).Google Scholar

²¹ See Linz, Juan, ‘Patterns of Land Tenure, Division of Labour, and Voting Behaviour’, Comparative Politics, VIII (1976), 365–430CrossRefGoogle Scholar. The political consequences of regional attachment are so important in Spain that to consider the topic in depth would require book-length treatment. Although we do not dwell on the subject in this article, the force of the regional factor is detectable in this and other tables in our presentation.

²² Class itself – or rather, its surrogates as measured by various occupational scales – does not turn out to be a significant predictor of left–right orientations once the effects of other factors, such as subjective class identification, political interest, and religiosity, are taken into account. While it goes in the expected direction, the correlation between occupation, indexed by prestige scores derived from International Labour Organisation codes, and subjective class identification is fairly weak (R = 0·38). This reflects the fact that there is a good deal of attitudinal and political heterogeneity to class positions in Spain. The empirical results are much the same as those reported by Sani, Giacomo, ‘Old Cleavages in a New Society: The Mass Bases of Spanish Parties’ (paper delivered at the International Symposium on Spain and the United States, University of Florida, Gainesville, 12 1979).Google Scholar

²³ The great weight of religious sentiment on political orientations in Spain has been noted by several scholars; see, for example, Linz, Juan, ‘Europe's Southern Frontier: Evolving Trends Toward What?’ Daedalus, CVIII (1979), 175–209Google Scholar. Once again, however, Spain is not as distinctive in this respect as is sometimes supposed. Religiosity (or denominational differences) have been found to exert a powerful influence on partisanship in many countries of Western Europe. See Rose, Richard, ed., Electoral Behavior: A Comparative Handbook (New York: Free Press, 1974), pp. 13–20Google Scholar; Lijphart, Arend, ‘Religious vs. Linguistic vs. Class Voting’, American Political Science Review, LXXIII (1979), 442–58CrossRefGoogle Scholar, and Michelat, Guy and Simon, Michel, Classe, Religion et Comportement Politique (Paris: Fondation Nationale de Sciences Politiques et Sociales, 1977).Google Scholar

²⁴ For an excellent study of ideological tendencies in the Spanish public, see Martin, Francisco Alvira et al. , ‘La Ideologia Política de los Españoles’, in La Reforma Política (Madrid: Centro de Investigaciones Sociologicas, 1977).Google Scholar

²⁵ For the party identification variable, PSOE (Socialist) includes the PSC and PSP. For the 1979/80 survey, it also includes the PSA (Socialist Party of Andalusia). UCD is the Government party led by the Prime Minister, Suarez. PCE (Communist) includes the PSUC, its counterpart in Catalonia. ‘Others’ are composed mostly of extreme left parties without parliamentary representation. ‘Right’ includes the Alianza Popular, led by Manual Fraga Iribarne, and the Coalition Democratica, as well as the extra-parliamentary Fuerza Nueva, the Falange and the Carlist party. The largest local party is the Partido Nacional Vasco in the Basque region. The proportion of non-identifiers is exaggerated slightly but not significantly by our treatment of the entire sample of those 16 years of age and older. Omitting respondents under 21 would reduce the relative number of identifiers a bit. In the second survey, we asked those who did not volunteer a partisan choice if they could select a party on being shown a list of party names. Fifty per cent of these respondents then chose one or another of the parties. We retain the responses to the unprompted item because this is the form usually presented in cross-national comparisons. See Budge, Ian, Crewe, Ivor and Farlie, Dennis, eds., Party Identification and Beyond (London: Wiley, 1976), p. 1976).Google Scholar

²⁶ Responses to the monarchy/republic item constitute a minor exception to the pattern of stability. The percentage of respondents claiming to be indifferent to the issue rose by nearly 15 points between the two surveys. It would be misguided, however, to take this datum alone as a sign of a trend indicating increased ‘alienation’ from the political system. Despite the importance given to it in the press, the controversy has evidently not been uppermost on the popular agenda at any time during the post-Franco period, and the fact that the issue was settled by a referendum in the autumn of 1979 no doubt contributed to its decline in salience. The increase in indifference probably means that the monarchy-versus-republic debate has become even more of a non-issue than it once was.

²⁷ For a sampling of the numerous commentaries in this vein, see Guzman, Eduardo, ‘Los Peligros del Desencanto’, Diario 16 (29 03 1980)Google Scholar and the editorial entitled ‘El Desencanto’, El País (30 03 1980), p. 8Google Scholar. Just as in the case of left–right placement, it is useful to set the proportion of non-identifiers in Spain in comparative perspective. Cross-nationally, the incidence of non-identifiers is very high indeed. In the eight-nation data collected by Barnes, Kaase et at., the country that most nearly approximates Spain in this respect is Switzerland, with 41 per cent non-identifiers; the lower bound is represented by the Netherlands (17 per cent) and Great Britain (19 per cent). The figure for Italy, which of all these nations is perhaps closest culturally to Spain, is 27 per cent. Of course, since the Spanish party system was in a state of suspension for over forty years, a large number of non-identifiers may be expected. The fact remains, however, that over the eighteen-month span between the two surveys, the party system in Spain shows no signs of ‘binding-in’. Because the period is so short, it is impossible to test the expectations of party consolidation discussed by Converse, Philip E., ‘Of Time and Partisan Stability’, Comparative Political Studies, II (1969), 139–71.CrossRefGoogle Scholar

²⁸ We confirmed this inference by tabulating reported vote in the 1977 national assembly elections with partisan identification (not shown here). The strength of the UCD among non-identifiers should be understood in the light of the fact that the UCD was formed out of a scattering of parties and factions only a few weeks before the first elections. It is altogether likely that a significant fraction of its support continues to emanate from the unaffiliated who think of themselves as ‘Suaristas’ rather than as members of a formal organization. See Roskin, Michael, ‘Spain Tries Democracy Again’, Political Science Quarterly, XCII (1978–1979), 629–46.CrossRefGoogle Scholar

²⁹ For all of the personalities except Franco, the question was worded as follows: ‘We would like to know your feelings about some leading figures in Spain today. Here we have a scale that goes from one to ten. If you like this person, the highest score you can give him is ten. If you dislike him, the lowest score is one. If you feel absolutely neutral or indifferent about him, you can give him a rating of five.’ In the case of Franco, respondents were presented with the same ten-point scale and asked: ‘A one on this scale means that you are totally dissatisfied, and a ten means that you are totally satisfied. In general, how satisfied or dissatisfied were you with the last ten years of the Franco government?’

³⁰ An exactly comparable question about parliamentary politics has been put to citizens of the Federal Republic of Germany on several occasions since the mid-fifties, and the results provide an intriguing point of reference for the Spanish experience. In 1956, 69 per cent of the German public felt that ‘a parliament in Bonn with all those deputies’ was needed, a figure very close to the current profile of opinion in Spain. By 1978, the figure had risen to 84 per cent. This is, of course, the trend to be expected with the increasing legitimacy of a competitive political system. As mentioned at the outset of this article, however, it is hazardous to project Germany's recent history onto Spain; the economic prospects of the Iberian peninsula are less favourable now than conditions in democratic Europe during the ‘miracle years’. We are grateful to David P. Conradt for calling the German data to our attention; the source is Demoskopie Archive, Surveys Nos. 095, 1010, 1020, 1031, 1069, 2085, and 3061. See also Baker, Kendall, Dalton, Russell and Hildebrandt, Kai, Germany Transformed (Cambridge, Mass.: Harvard University Press, 1980).Google Scholar

³¹ It is also evident that the King and Franco are the only figures whose popularity has not declined between 1978 and 1979–80. A particularly troublesome sign for the UCD is that the popularity of Suarez has suffered a significant drop. Our estimates of the fall in approval for Suarez match data from opinion polls conducted from September 1976 to March 1980 for the magazine Cambio 16 and reported in the article ‘Baja la Popularidad del Presidente’, Diario 16 (29 03 1980), p. 3.Google Scholar

³² The item tapping satisfaction with the Franco government is the same as that translated in fn. 29. The item for the Suarez government is: ‘Taking everything into consideration, in general how satisfied or dissatisfied are you with the present government?’

³³ The correlations between satisfaction with the present government and satisfaction with the Franco government in 1978 and 1979/80 are, respectively, 0·41 and 0·30.

³⁴ The tendency of Spaniards to locate themselves toward the centre has been found in every survey with which we are familiar using left-right scales in post-Franco Spain. Yet, for all this consistency, an important caveat should be noted. The left-right scale is a single, supposedly generic indicator. It may conceal significant polarization of opinion on specific issues. Considered alone, it also fails to tap the possibility that political cleavages in Spain may be multidimensional. Regional divisions furnish a prime instance of this possibility.

³⁵ Since the results from the second survey are essentially the same, only the 1978 figures are presented.

³⁶ The socio-economic background factors taken together, including occupation, do not explain more than 5 per cent of the variance in the popularity of any of the political personalities. Such results are technically significant but they are not substantively very meaningful with the large number of cases under study. Neither is subjective class identification a prominent predictor. Only in the case of the right-wing leader Fraga Iribarne does it retain an independent effect, and even then it is rather insubstantial. In interpreting these data, it is helpful to realize that they represent summary results. Restrictions of space prevent us from recapitulating all stages of the analysis, beginning with the baseline socio-economic models. In addition to using a step-wise procedure to eliminate the insignificant predictors, we tested interactive models to check the appropriateness of the additivity assumptions of the Multiple Classification Analysis. These models were all multiplicative, and they added nothing in statistical terms to the MCA results. However, it is possible that further specification of the models may yield somewhat more powerful explanations of the popularity of political figures. For an exemplary analysis of this problem, see Franklin, Mark N. and Mughan, Anthony, ‘The Decline of Class Voting in Britain: Problems of Analysis and Interpretation’, American Political Science Review, LXXII (1978), 523–34CrossRefGoogle Scholar. A substantive qualification about the ‘failure of class’ should also be kept in mind. Although unemployment has been on the rise since at least 1973 (even if some of its effects have been muffled by the social security system), workers who have kept their jobs have seen their purchasing power increase during the transitional period; the share of the GNP allocated to wages has grown. During the period immediately following the death of Franco, the advantages won by labour resulted from a combination of working-class militancy and the disorganization and uncertainty of the business community. Since 1977, when all the major parties, including those on the left, entered into an economic accord with the government, wage settlements favourable to labour have been reached in exchange for political moderation. If there is any group, besides the unemployed themselves, that has been severely penalized by these policies, it is the members of the smaller middle class, who generally lean toward the centre-right. See Quintana, Enrique Fuentes, ‘La Crisis Económica Española’, Papeles de Economia Española, I (1980), 31–67.Google Scholar

³⁷ See Cooper, Norman, ‘The Church: From Crusade to Christianity’, in Preston, Paul, ed., Spain in Crisis (New York: Harper and Row, 1976)Google Scholar and the literature cited therein.

³⁸ To the extent that the age variable does capture generational and historical differences, the concentration of conservatives and right-centrists among the older cohorts can be taken as a hopeful sign for the left in Spain, since age-specific mortality rates would seem to favour the left. This extrapolation is favoured by Sani, , ‘Old Cleavages in a New Democracy’, pp. 6–9Google Scholar. However, efforts at forecasting become highly speculative when the party system itself is so young. Had we found a significant association between youth and sympathy with the socialists or the communists, the ‘time-will-tell’ argument would be more convincing.

³⁹ The results for the analysis of the determinants of Santiago Carrillo's popularity may seem especially puzzling, since ‘everybody knows’ that communist support is, or should be, concentrated among the urban working class. Yet none of the standard socio-economic predictors bears significantly on the popularity of the communist leader. On consideration, however, the importance of ‘political interest’ has great theoretical significance. Although identification with the communist party leadership and programme can scarcely be called widespread in Spain, such support does depend on the commitment of party militants and organizers, whatever their class background. It is this militancy, rather than any simple association between ‘class and communism’, which the MCA results detect. Compare Barnes, Samuel H., ‘Ideology and the Organization of Conflict: On the Relationship between Political Thought and Behavior’, Journal of Politics, XXVIII (1966), 513–30CrossRefGoogle Scholar; Mann, Michael, Consciousness and Action among the English Working Class (London: Macmillan, 1973)CrossRefGoogle Scholar; and Zipp, John F. and Smith, Joel, ‘The Structure of Electoral Political Participation’, American Journal of Sociology, LXXX (1979), 156–66.Google Scholar

⁴⁰ See for example the analyses by Maravall, José Maria, ‘Eurocomunismo y Socialismo en España: La Sociologia de una Competición Política’, Sistema, XXVIII (1979), 51–73Google Scholar and Tezanos, José Félix, ‘El Espacio Político y Sociológico del Socialismo Espanol’, Sistema, XXXII (1979), 51–75Google Scholar. If one examines the correlations between party vote and socio-economic variables, such as urbanization, across provinces in Spain (n = 52), the coefficients are often strikingly high, confirming that indeed the urban areas favour the left. These results generally wash out when the unit of observation becomes individuals, and the hoary dilemma of ecological versus individualistic inference sets in. The problem is acute in Spain because it is extremely difficult to find systematic data on structural indicators at levels intermediate between individuals and provinces. We were able, however, to obtain data, gathered by the Banco Banesto, on the per capita income of each of over 8,000 municipios (counties) in Spain for 1970. For the 130 municipios that compose the primary sampling units for our surveys, we calculated the mean left-right placement of individuals residing in these areas. The county-level correlation between per capita income and mean left-right placement turns out to be 0·07 – in the expected direction (the higher the income, the more to the right), but very weak.

⁴¹ See Przeworski, Adam, ‘A History of Western European Socialism’ (unpublished manuscript, University of Chicago, 1978).Google Scholar

⁴² The variable-sum quality to the relationship between religiosity and political orientations was hinted at in Table 8, which shows the determinants of support for the socialist leader Felipe Gonzalez. He is not markedly unpopular among the non-identifiers, who tend to be more religious than those with partisan identification.

⁴³ See Stephens, , ‘Class Formation and Class Consciousness’Google Scholar, and Lepsius, M. Rainer, ‘Soziale Ungleichheit und Klassenstruckturen in der Bundesrepublik Deutschland: Lebenslagen, Interessenvermittlung und Wertorientierungen’, in Wehler, Hans-Ulrich, ed., Klassen in der Europäischen Sozialgeschichte (Gottingen: Vandenhoeck und Ruprecht, 1980)Google Scholar, and Diaz, Victor M. Pérez, ‘Clase Obrera y Organizaciones Obreras en la Españia de Hoy: Política y Vida Sindical’, Sistema, XXXII (1979), 3–17.Google Scholar

⁴⁴ The somewhat greater loyalty of the UGT members to the PSOE, compared to that of the CCOO members to the PCE, may appear odd if it is assumed that nominally communist trade unions are generally more thoroughly politicized than socialist trade unions. In fact, however, there is little evidence to suggest that this is the case. Conclusions about the supposed political ‘tightness’ of communist unions in Spain are sometimes propagated by reversing the independent and dependent variables in arrays like Table II. See for example Maravall, , ‘La Alternativa Socialista’, p. 33Google Scholar. This procedure generally shows that, among PCE voters or sympathizers, a very high proportion come from the CCOO. Yet it does not indicate the reverse probability that a CCOO member, or anyone else, will vote for this or that particular party.

⁴⁵ The bivariate association depicted in Table II is sustained when relevant control variables are introduced. However, the direct association is attenuated and amplified, respectively, at different levels of individual political interest. To take the extreme cases: among those reporting no interest whatsoever in politics, 60 per cent of the UGT members claim identification with PSOE and 26 per cent of the CCOO members with the PCE. At the highest level of political interest, fully 91 per cent of UGT members identify with PSOE and 62 per cent of CCOO members with the PCE. Statistically, then, there is a strong interaction between organizational membership and political awareness; substantively, the psychological and organizational factors reinforce one another.

⁴⁶ This should not be taken to mean that all or even most Spanish peasants have homogeneously radical proclivities. The classic interpretation of political ‘schizophrenia’ in rural Spain is Malefakis, Edward E., Agrarian Reform and Peasant Revolution in Spain: Origins of the Civil War (New Haven, Conn.: Yale University Press, 1970)Google Scholar. However, it does appear to be the case that a large landless peasantry is a necessary condition for successful radical movements of the left; see Paige, Jeffrey M., Agrarian Revolution: Social Movements and Export Agriculture in the Underdeveloped World (New York: Free Press, 1978)Google Scholar. Nowadays, it is customary to look upon what remains of the European peasantry as a conservative force. See for example Lewis-Beck, Michael S., ‘Explaining Peasant Conservatism: The Western European Case’, British Journal of Political Science, VII (1977), 447–64CrossRefGoogle Scholar, and Weber, Eugen, Peasants into Frenchman: The Modernization of Rural France, 1870–1914 (Stanford, Calif.: Stanford University Press, 1976)Google Scholar. But the Spanish (and the Portuguese) peasantry are not equivalent to the small proprietors of Northern Europe; there is a considerable, although greatly reduced, number of landless labourers, especially toward the South, where anarchist and radical movements have traditionally been strong. See Gilmore, David D., The People of the Plain: Class and Community in Lower Andalusia (New York: Columbia University Press, 1980).Google Scholar

⁴⁷ Compare Barnes, Samuel H., Representation in Italy: Institutional Tradition and Electoral Choice (Chicago: University of Chicago Press, 1977).Google Scholar

⁴⁸ Only 13 per cent of our respondents report belonging to two or more organizations of any kind. Twenty-three per cent say that they belong to only one voluntary association, with the great majority – 64 per cent – saying they belong to none. (The most popular associations are athletic clubs: 11 per cent belong.) Although these figures probably underestimate the true extent of associational activity, since the questions were posed at the end of the interviews when response fatigue is likely to set in, they still seem exceptionally low.

⁴⁹ The role of the King is, of course, not merely psychological. If it is true, as the data amply show, that no single political party is hegemonic at the mass level in Spain, then the function of the King in providing an anchor during the transition process assumes great importance. The contrast with contemporary Brazil, where the incipient transition away from authoritarianism has followed a zigzag path, is instructive. In the absence of a figure comparable to Don Juan Carlos, the Brazilian military has continued to fill a vacuum as the ‘moderating power’ among competing interests. See McDonough, Peter, Power and Ideology in Brazil (Princeton, N.J.: Princeton University Press, 1981).CrossRefGoogle Scholar

⁵⁰ Compare Seiler, Daniel L., ‘Clivages, Regions et Science Politique: Application d'un Schéma d'Analyse au Cas de la Suisse et de la Belgique’, International Journal of Political Science, x (1977), 447–72.Google Scholar

⁵¹ These ‘new politics’ groups form a primary constituency behind the tentative movement to establish an organization along the lines of the Radical Party in Italy. See ‘Hay Espacio para un Nuevo Partido?’ El País (8 09 1979), p. 6Google Scholar; Dorronsoro, Javier A., ‘Partido Radical o Partido Revolucionario?’ El País (12 03 1980), p. 18Google Scholar, and Llorente, Luiz Gomez, ‘Saludo al Partido Radical’, El País (3 04 1980), p. 35.Google Scholar

⁵² See Crewe, Ivor, Särlvik, Bo and Alt, James, ‘Partisan Dealignment in Britain, 1964–1974’, British Journal of Political Science, VII (1977), 129–90CrossRefGoogle Scholar; Inglehart, Ronald and Hochstein, Avram, ‘Alignment and Dealignment of the Electorate in France and the United States’, Comparative Political Studies, v (1972), 343–72CrossRefGoogle Scholar, and Beck, Paul Allen, ‘The Electoral Cycle and Patterns of American Politics’, British Journal of Political Science, IX (1979), 129–56.CrossRefGoogle Scholar

⁵³ This is not to say that a leftist government would represent a mere change of the palace guard. It would almost certainly imply a greater role for the public sector – not a trivial transformation given the comparatively modest level of, for example, fiscal control by the Spanish government. See Esping-Andersen, Gösta, ‘Social Class, Social Democracy, and the State: Party Policy and Party Decomposition in Denmark and Sweden’, Comparative Politics, XI (1978), 42–58.CrossRefGoogle Scholar