THIS CONCEPT RELATES to Anderson's "consensus," which he defines broadly as "the degree of congruence in the cultural orientations of various individuals and groups comprising an organization." He points out that parties' scholars are interested in the issues which obtain consensus, in the level of consensus obtained for different issues, and in the distribution of consensus across party organs (1968, pp. 396-397). Under this conception, consensus deals primarily with attitudinal agreement among party members. Not only is this type of data unlikely to be obtained through the library research procedures of the parties' project, but attitudinal agreement by itself appears to be too static a concept for the comparative analysis of party politics. The fact that party members disagree over an issue is undoubtedly important, but it is important to know how, if at all, that disagreement is expressed in intraparty politics. Therefore, we choose to focus instead on the concept of "coherence," defined as the degree of congruence in the attitudes and behavior of party members. In so doing, we follow the lead of Huntington, who introduces the concept of coherence, defines it in terms of consensus, and then suggests ways that coherence can be measured (1965, pp. 403-405).

The equivalent of coherence has been examined in studies of party "cohesion" and "factionalism." These studies have sought to identify the sources of coherence and to assess the consequences of varying degrees of coherence upon party effectiveness. Our data should be able to support both types of inquiry, as we operationalize coherence with reference to six basic variables: [Janda, 1980: 118]

The legislature constitutes one of the main political arenas in which one would expect to find evidence for or against party "coherence." A party that ranks high on coherence should demonstrate high cohesiveness, if not unanimity, in its legislative behavior. Although there are different forms of legislative behavior, one of the most important forms for comparative analysis is the voting decision. In a cohesive party, the party members in the legislature tend to vote the same way on issues before the chamber. Although studies have shown that certain characteristics of the political system-for example, a presidential or parliamentary form of government-strongly affect the degree of party cohesion in legislative voting (Ozbudun 1970), the lack of cohesion still reflects a lack of coherence--whatever the cause.

Comparative studies of legislative behavior have also disclosed great differences in the number of voting decisions rendered within legislatures of various countries. But it is not easy to judge the effect of these differences on the measurement of party cohesion, and we do not allow for them in our conceptualization. Wherever possible we seek to measure cohesion as demonstrated in "roll-call votes," which record the voting positions of individual legislators. Such votes are referred to variously as "divisions" or "record" votes in different countries.

Operational Definition. Suitable data for this measure were hard to obtain. Ideally, we sought data from which we could calculate the mean index of cohesion, a measure devised by Stuart A. Rice and calculated for a given vote as follows:

Index of Cohesion =

| N "Yes" - N "No" | | N "Yes" + N "No" |

= | % "Yes" - % "No" |

This measure ranges from .0 in the case of the members of the party splitting 50:50 in support and opposition to the bill to 1.0 in the case of all members voting the same, either all in favor or all opposed (see Anderson et al. 1966). The mean index is simply the "average" index for all bills voted on.

Unfortunately, the parties' literature rarely reported precise indices of cohesion for the legislative voting. When party divisions on legislative votes themselves were divulged, the index was calculated or approximated. In the complete absence of empirical data, we estimated party cohesiveness as shown in Table 10.1.

Other complications arose in handling this variable. The problem presented by one-party states was met by not scoring parties for legislative cohesion if they monopolized the legislature and if the legislature was not a forum for the expression of intraparty conflict. The presence of bicameral legislatures forced us to choose between chambers in assessing legislative behavior, with the lower house generally selected. Finally, because we had no sound basis for picking issues on which to base our measure of cohesion, we accepted votes reported on any issues.

The notion of party "factions" has figured prominently in the comparative analysis of political parties, and a party is assumed to lose coherence as its energies are dissipated in factional conflict. Zariski defines a faction as "any intra-party combination, clique, or grouping whose members share a sense of the common identity and common purpose and are organized to act collectively&endash;as a distinct bloc within the party&endash;to achieve their goals" (1960, p. 33). These goals may range across a variety of objectives. For the purposes of our analysis, we are especially interested in factions that arise in the pursuit of party "ideology" (the objective in variable 10.02), more discrete political "issues" (variable 10.03), political "leadership" (variable 10.04), and political "strategy" (variable 10.05).

The distinction between ideology and issues is often difficult to draw, but it hinges on the difference between an overall governmental philosophy, which characterizes an ideological faction, and discrete government policies, which provoke issues factionalism. Thus, southern Democrats would constitute more of an ideological faction in the U.S. Congress than an issue-oriented faction, which would be more appropriate to party divisions related to U.S. military action in Vietnam.

Zariski later qualifies his definition of a faction by requiring that the members maintain "a reasonably lengthy period of association&endash;several months, at the very least" (p. 37). Accepting this criterion of durability, we ignore fleeting "factions" as labeled by writers who are not using the term very precisely. Instead, we look for evidence of formalized interaction and joint consultation that exists over some minimum period of time, established as a year or more.

The main fact in assessing factionalism is, of course, determining the sheer existence of a faction. But factions can be more or less distinct, depending on whether or not they are labeled and whether or not factional membership can be determined. At the extreme, factions may be so distinct that they maintain their own organizations, with offices, staff, and officials, and operate largely independently of the party organization or in opposition to it.

Another important facet of factionalism is the size of the factional group, relative to the size of the party as a whole. In general, smaller factions are judged to be less threatening to party coherence than are larger factions, and the size of the factions is judged to be more important for party coherence than for the existence of formal organization .

One additional possible distinction in assessing party factionalism is whether the factional conflict erupts into a party split or purge. The occurrence of splits or purges over ideological matters has to be treated gingerly, however, for either may indicate the development of "coherence" in the party as well as indicating the existence of "incoherence." The distinction depends on the timing of the event. If it occurs at the beginning of the time segment under consideration, it leads to coherence; if it occurs later, it signifies incoherence.

The final issue in determining the existence of party factions is establishing where they may occur. In parliamentary democracies, there is the inclination to look for factions within the party's legislative membership, and this is a likely and important locus for factional activity. But factions also occur outside of the parliamentary party membership in parliamentary democracies as well as in other political systems. In general, we insist on identifying the "factionalists" as party members, but even this is difficult to require for nonmembership parties. In such parties, we require that members of the proposed faction be at least regarded as strong party identifiers.

Operational Definition. The higher the score on this variable, the greater the degree of ideological factionalism. The party is scored the highest applicable code from the following scale.

0	Ideological concerns are not subject to public debate and disagreement among party leaders.
1	Ideology is a matter of public debate and disagreement among party leaders but not enough to promote factional tendencies.
2	Ideology is a matter of debate and disagreement, factional tendencies are present, but factions are not clearly distinguished in the sense of labeled groupings with identifiable membership.
3	Ideological concerns have created a "small" faction within the party, but the faction does not have a formal organization of its own.
4	Ideological concerns have created a "small" faction within the party with some formal organization of its own or have provoked a split after the beginning of the period.
5	Ideological concerns have created a "large" faction within the party "large" defined as about 25 percent of the membership or more but the faction does not have a formal organization of its own.
6	Ideological concerns have created a "large" faction within the party with some formal organization of its own or have provoked a split after the beginning of the period.

The section under variable 10.02 discusses the conceptualization of factionalism in general within the specific context of "ideological factionalism." Variable 10.03 also refers to a specific type of factionalism, "issue factionalism." Issue factionalism is based on disagreement over one or more discrete substantive issues in politics. This stands in contrast to ideological factionalism, which is based on conflicting emphases in the context of an overall governmental philosophy. To differentiate between transient divisions on political issues and factionalism on issues, recall that a faction involves some formalized interaction or joint consultation over the issue that exists for a period of a year or more. We do not score a party high on issue factionalism simply by implication if it gets a high score on ideological factionalism. There must be independent identifiable evidence of one or more discrete political issues that divide the party as well.

Operational Definition. The higher the score on this variable, the greater the degree of ideological factionalism. The party is scored the highest applicable code.

0	There seem to be no long standing issues that are subject to public debate and disagreement among party leaders.
1	One or more long standing issues are a matter of public debate and disagreement among party leaders, but they are not sufficient to promote factional tendencies.
2	One or more long-standing issues are subject to debate and disagreement, factional tendencies are present, but factions are not clearly distinguishable in the sense of labeled groupings with identifiable membership.
3	One or more long standing issues have created a "small" faction within the party, but the faction does not have a formal organization of its own.
4	One or more long standing issues have created a "small" faction within the party with some formal organization of its own or have provoked a split after the beginning of the period.
5	One or more long standing issues have created a "large" faction within the party "large" defined as about 25 percent of the membership or more but the faction does not have a formal organization of its own.
6	One or more long standing issues have created a "large" faction within the party with some formal organization of its own or have provoked a split after the beginning of the period.

See the discussion under variable 10.02, "ideological factionalism," for the basic conceptual definition of factionalism. This variable deals with factions that form behind different political leaders primarily on the basis of the leaders' personal attraction. In many cases, ideological matters or issue orientations will supplement the attractive qualities of the leaders' personalities. In such instances, the party may be determined to display ideological and issue factionalism in addition to leadership factionalism. That determination is made separately for the party in terms of the two preceding variables. For this variable, the main conceptual distinction is the "personalist" basis of the following, often regarded in terms of charismatic leadership. In some cases, the charismatic basis may be lacking, but a personalist substitute may exist in the form of raw patronage for the supporters of the political leaders' factions. This, too, constitutes a basis for leadership factionalism.

Operational Definition. The higher the score on this variable, the greater the degree of leadership factionalism. A party is given the highest score that applies from the following scale.

0	Leadership contests for control of the party either do not occur or they are so covert or so "inside" that they do not engage large numbers of party members in their outcome.
1	Leadership contests for control of the party do emerge into view of party members, but factional tendencies are not evident.
2	Leadership followings are present, but factions cannot be clearly distinguished in the sense of labeled groupings with identifiable membership.
3	Followers of a political leader have created a "small" faction within the party, but the faction does not have a formal organization of its own.
4	Followers of a political leader have created a "small" faction within the party with some formal organization of its own.
5	Followers of a political leader have created a "large" faction within the party "large" defined as about 25 percent of the membership or more but the faction does not have a formal organization of its own.
6	Followers of a political leader have created a "large" faction within the party with some formal organization of its own.

The general phenomenon of factionalism is discussed at length under variable 10.02. "Strategic or tactical factionalism" is the fourth variety we consider. Members of political parties may agree on ideology and issues but disagree seriously on the strategy that the party ought to use in achieving its goal or perhaps on particular tactics that the party ought to follow within a given strategy. (See the discussion of the goal orientation cluster and variable 6.00 for a distinction between strategy and tactics.) Variable 10.05 is intended to express this basis of disagreement, or lack of coherence, within a party.

Operational Definition. A party is assigned the highest score that applies from the following scale, designed to measure the extent of factionalism concerning party strategy and tactics.

0	There is little or no disagreement voiced within the party concerning its appropriate strategy or tactics with regard to its goal orientation.
1	Disagreements over strategy or tactics do emerge into view of party members, but factional tendencies are not evident.
2	Advocates of certain strategies or tactics do agitate for their position, but factions cannot be clearly distinguished in the sense of labeled groupings with identifiable membership.
3	Adherents to a certain line of strategy or tactics have created a "small" faction within the party, but the faction does not have a formal organization of its own.
4	Adherents to a certain line of strategy or tactics have created a "small" faction within the party with some formal organization of its own.
5	Adherents to a certain line of strategy or tactics have created a "large" faction within the party&endash; large defined as about 25 percent or more of the membership but the faction does not have a formal organization of its own.
6	Adherents to a certain line of strategy or tactics have created a "large" faction within the party with some formal organization of its own.