¹ St. Martin's Press, New York, McMillan & Co., London, 1967, xiii + 375 pp. Further unreferenced page numbers refer to this work.

² Second edition, Armand Colin, Paris, 1965.

³ The exceptions are disappointing. Thus, Goguel has basically continued the Siegfried tradition without modification. Leites, who brilliantly draws attention to the psychic world of deputies, fails to indicate how he samples language and fails to relate the psychic world to other influences such as constituency pressures.

⁴ Boudon, Raymond, L'Analyse Mathematique des Faits Sociaux, Paris, Plon, 1966 Google Scholar.

⁵ Authors mentioned above and not cited are referenced in MacRae's bibliography.

⁶ The scales represent heuristic clusterings of roll-calls. All items in a scale have Yule's Q coefficients of .8 or better with one another. While better scaling procedures, particularly multidimensional scaling, might now be available, I present no critique on this topic given the development of data processing at the time MacRae carried out his research. Also, while MacRae largely selected roll-calls on the basis of L'Année Politique and while a better selection might have been made, I assume the improvement would have had only a marginal impact on the results.

⁷ Dimensions of Congressional Voting. University of California Publications in Sociology and Social Institutions, Vol. 1, No. 3. Berkeley, University of California Press, 1958 Google Scholar.

⁸ Leites, Nathan, On the Game of Politics in France, Stanford, Stanford University Press, 1959, pp. 29–32 Google Scholar.

⁹ This finding is held subject to the reservations about the multiple regression analysis made below. On the continued importance of religious practice for voting behavior, see Rosenthal, Howard, “The Popularity of Charles de Gaulle: Findings from Archive-Based Research,” Public Opinion Quarterly, Summer, 1967 Google Scholar.

¹⁰ Specific data on such local pressures is now available from a survey of Fourth Republic politicians conducted by the French Institute of Public Opinion (IFOP). See Michelat, Guy and Rosenthal, Howard, “Studies of Electoral Coalitions in the Fourth Republic,” Graduate School of Industrial Administration, Carnegie-Mellon University, mimeo. 1968 Google Scholar.

¹¹ These figures are estimates made from various tables as MacRae does not give precise figures. Even if separate analyses were carried out by party, it would have been useful to have a scale analysis of all deputies to treat the problem of unsealed deputies. Such an analysis would also have presented a dimensional portrait of the parties. Consequently, one would have been able to handle more precisely the question of the left-right position of the Radical, MRP, Moderate, and Gaullist blocs. One could also relate such a “mapping” to the various government coalitions. Michael Leiserson has made an initial attempt in this direction. See his “Political Coalitions,” unpublished doctoral dissertation, Yale University, 1966 Google Scholar.

¹² P. 96. On p. 181, MacRae unfortunately leaves the impression that only unanimity was a reason for a vote not being scaled. Such is obviously not the case; many roll-calls were not scaled because they did not fit into a cluster.

¹³ Based on Tables 3.6, p. 56; 4.5, p. 104; 5.5, pp. 146–147; 6.5, p. 174.

¹⁴ This categorization omits critical roll-calls during the life-time of a cabinet. Such roll-calls, which could be spotted by more detailed research and which would present great interest in the general study of instability, are not necessary to the points made here.

¹⁵ Although some roll-calls which appear on the graphs to contain 10% dissidence may not have been scaled because some of the dissidents did not stay within the group for the whole Legislature, such dissidents, as suggested above, bear close watching for “gaming” behavior.

¹⁶ In the entire Fourth Republic, the Socialists divided on only two “critical” votes, the Thorez and De Gaulle investitures.

¹⁷ P. 292.

¹⁸ Queuille's final government simply terminated with the elections for the Second Legislature.

¹⁹ P. 299.

²⁰ Rosenthal, Howard, “The Electoral Politics of Gaullists in the Fourth French Republic: Ideology or Constituency Interest?”, this Review, 06, 1969, pp. 476–487 Google Scholar.

Other limitations of MacRae's analysis of constituency relations are: (1) While he suggests that party militants limited bargaining at the local level (pp. 11, 289–302), study of apparentement alliances for the 1951 elections shows that the “highly organized” MRP formed the most coalitions. Another type of coalition, joint lists among two or more parties, in turn seems as related to the size of their vote as to their membership ratios. See Rosenthal, Howard, “Voting and Coalition Models in Election Simulation,” in Coplin, William, ed., Simulation and the Study of Politics, Chicago, Markham Press, 1968 Google Scholar; (2) Generally, the treatment of the role of electoral interest vs. that of ideology is too superficial. Thus, MacRae interprets (p. 176) an association between scales specific to issues and scales specific to cabinets as supporting evidence for the lack of a game and infers that high specificity was a characteristic of “ideological” parties. A “perfect” data point in this respect is represented by the Socialists' unique First Legislature scale which was entirely specific to one cabinet and to one issue, the electoral law. Yet MacRae only hurriedly investigates (pp. 290–291) the extent to which Socialists were simply voting their electoral interests on this issue. While awaiting such an examination, an alternative proposition can be advanced; namely, that since specificity seems to be related to the number of votes scaled, it is primarily related to the extent of party discipline rather than ideology. Since, when discipline breaks down, it may tend to do so during a brief period of time over a single question (which may be either ideological or tactical), cabinet and issue specificity may be associated. Hence, neither of these measures by themselves can be expected to reveal much about either ideology or coalition formation.

²¹ P. 9.

²² P. 53 and n.35, p. 63.

²³ Pp. 308, 313–315.

²⁴ Similarly, it is doubtful that the Mendesists were unskilled.

²⁵ P. 169.

²⁶ P. 161.

²⁷ With respect to ministerial rewards, MacRae has only considered differences in support between ministrables and non-ministrables. He thus has no measures on those who aimed at office but were unsuccessful. He also did not report on the relationship of voting to the allocation of posts over time.

In any case, his inference of a consensual system from ministrable support for cabinets can be questioned in that examination of Table 7.2 (p. 187) discloses that the association between ministrable status and cabinet support was strong only for the First Legislature. Furthermore, on p. 193, MacRae incorrectly interprets data in support of his hypothesis:

For one ministrable, Antoine Pinay, there is additional evidence of his support for cabinets of which he was a member. Pinay's friends con- stituted a well-known group … four of the seven are. in the top row on Scale II-2. They were somewhat more favorable to the cabinet than other members of the party.

We can hardly believe this since 43 of the total 63 Moderates scaled were also in the top row (see Table 7.4, p. 192).

²⁸ Actually 82 departments or dioceses, except for the Poujadist vote where there were fewer units. MacRae originally published these results in the American Journal of Sociology in 1958. While it may now appear subject to reservations, MacRae's study was undoubtedly instrumental in stimulating the use of multiple regression in French political studies.

²⁹ See Rosenthal, Howard, “Poujade's Voters: Suggestions for the Study of the Social Geography of the New French Politics,” Carnegie-Mellon University, Graduate School of Industrial Administration, mimeo., 1968 Google Scholar. This study further raises the point that beyond the particular case of “ruralism” discussed below, MacRae did not consider that Poujadism might represent a more complex reaction to modernization than the simple “defense” of “situations acquises” (p. 12).

³⁰ MacRae does not indicate whether his correlations are adjusted.

³¹ P. vii. The lack of rigor seems especially true for the comparisons MacRae draws between France and the United States. The qualitative differences between, say, lack of white-black communication in the U.S. and ouvrier-patron in France are so important that it is very risky to rely on the types of statistical tables now available for international comparisons and thereby establish such propositions as, “In personal relations and social structure, France is more divided than the United States” (p. 6).

³² March, James G. and Simon, Herbert, Organizations, New York, John Wiley, 1957 Google Scholar; Cyert, Richard M. and March, James G., Behavioral Theory of the Firm, New York, Prentice-Hall, 1964 Google Scholar.

³³ Davis, Otto A., Dempster, M. H., and Wildavasky, Aaron, “A Theory of the Budgetary Process,” this Review, 60, 1966, pp. 529–547 Google Scholar.

³⁴ Downs, Anthony, An Economic Theory of Democracy, New York: Harper & Row, 1957 Google Scholar; Riker, William, The Theory of Political Coalitions, New Haven, Yale University Press, 1962 Google Scholar.

³⁵ P. vii.