Though perhaps not a conscious one, there is a connection between theories about the world and the way political actors in the world behave.  For example, leaders who focus on relations of power among nations have a more “realist” mindset, while those who emphasize cooperation and social welfare in world politics are more “idealist” or “liberal internationalist.”  Further, these theories about the world can often be classified as more inclusive ways (e.g., liberal internationalist) or exclusive ways (e.g., realist) of viewing human relations.  When reading interviews with leaders and actually interviewing members of parties in divided societies, it is striking how often one sees or hears a comment that sounds like a proposition or a conclusion from one of several theories about mobilization or intergroup behavior.  For this study, a content analysis method was created to link theories emphasizing different reasons for group mobilization and intergroup conflict with the kinds of strategies leaders actually use in their rhetoric (or propaganda).  The strategies are also classifiable on a scale from more to less exclusive (the key aspect in this article).
        Understanding the theory that informs a leader’s definition of the situation is crucial because any proposals to address the problem—here, any negotiation points or peace proposals—will not get serious consideration if they do not address the problems and appeal to the causal relationships that the leader sees as paramount.  In other words, a leader with a mostly Enemy Image mindset would be a lot less likely to accept economic aid as a solution than another with a predominantly relative deprivation perspective (see below).  This research design and coding is also a useful approach because it offers insight into the international and domestic conditions under which particular strategies are most likely to succeed.  If external actors wish to push a leader to define the situation more or less exclusively, they would need to know more about how they could manipulate the context to increase the chances of that happening.

a.  Identity Strategy:  Group reference and group delineation
        Because the objective is to compare which people/groups the leaders deem relevant and the labels or bases of identification that different leaders (and the same leaders over time) use to describe potential groups in the society, the coder should underline all the groups the leader mentions within each unit (topic).   For this project, instead of using the particular labels (“Catholics,” “politicians,” “working class”), a method more generalizable across time and cases is used to record the group delineation.  Lists of labels that tend to be used to describe people in each country were created inductively, and then the lists were categorized according to degrees of exclusivity:  very exclusive, moderately exclusive, moderately inclusive, and very inclusive.
        It is not suggested here that there is an objective definition of  “inclusive” and “exclusive,” but in order to observe differences between leaders and changes over time, it is necessary to assert a reference point.  This project approaches this issue by considering the group reference:  who are the people the leader refers to when using a particular label?  The second code is the group delineation, for which the labels are assigned degrees of exclusivity based on the “perspective” of the group reference used with the labels.  For example, consider two groups with the same group reference named by a leader:  “Nationalists in Northern Ireland” and the “working class folks in Northern Ireland.”  From the perspective of the Northern Ireland group reference, “Nationalists” is considered “very exclusive” (because it is completely exclusive of those claiming British identity) and “working class” is considered “moderately inclusive” (since all are not in the working class in this area, it is not coded as “very inclusive”).   For this study, the labels “Catholic” and “Protestant” are considered “moderately exclusive” because religion does not preclude membership in either state, where as “unionist” (and also “loyalist”) and “nationalist” (and also “republican”) are mutually exclusive.
        Therefore, using the rules below, the coder is to record which people the leader is grouping together (Group reference).  This is an attempt to be as objective as possible; the coding then records the different ways in which groups or potential groups in these areas are bounded and categorized (Group delineation).  This section explains the coding rules for these two variables.
        How the leader describes both ingroups and outgroups (or “self” and “others”) is of interest.  “Outgroup” is used here to mean any group that the leader distinguishes as an “other,” or not part of the “self.”  “Ingroup” is used here to mean the group with which the leader identifies. For example, President Clinton may use the following ingroup descriptions: “Americans,” “Democrats.”  He may name outgroups such as “French government,” “Republicans.”
        An important issue is that identities are often nested, so that some outgroups may be discussed as subgroups of a larger ingroup.  For example, “Irish from the Protestant tradition” is surely an outgroup when spoken by the leader of a predominantly Catholic party; but the fact that he is suggesting the shared dimension of “Irish” may be crucial to understanding a dynamic conflict.  From the perspective of Catholics, this portrayal is at least more conciliatory than “Unionists” alone, which emphasizes no shared dimensions of group membership.  The idea of nesting and presenting a picture of group as multidimensional and crosscutting may be a key way of encouraging multiple loyalties.  In this project, intercoder reliability could not be achieved when the coder is asked to code when an outgroup is talked about as a subgroup of a larger ingroup.  Still, some sense of how leaders posit multiple levels of identity for the ingroup can be understood in the analysis, by examining the various ingroup references made.
        Group reference:  This section reviews the discussion above with another example. Before the groups are coded in terms of labels, it is essential to know which people the leader is trying to label.  For example, when someone says “Protestants” it is important to know if this label is given to what is considered the political majority in Northern Ireland, or if the speaker is using the island of Ireland as a reference point and therefore seeing this group as a minority.  If the group reference is to people on the island of Ireland, “Protestants” is a very exclusive way of describing people.  On the other hand, calling the majority in Northern Ireland “Protestant” is moderately exclusive (since more exclusive labels are commonly used:  Orangemen, unionists, loyalists).  This is important because at different times a leader may use the labels “Protestant majority” or “Irish of a different tradition.” In these other terms, the leader is potentially including and excluding different people.
        For the Northern Ireland cases, possible group references include people living in the borders of Northern Ireland (which may be called British or Irish or may be broken down into Catholics/Protestants or Nationalists/Unionists, for example), those living in the Republic of Ireland, those in the island of Ireland, those living in mainland Britain, people in both areas, those in Europe, and people in parts of the world or universal references (so people in Northern Ireland may be referred to as Christians, which has no geographic reference).  It should be apparent that for the Identity strategy, the coding is set up to allow the data to speak for itself so that the analysis is as interpretive as possible in recording the way leaders suggest group boundaries be drawn.
        Group delineation:  The “label lists” were created from research on these cases and from trial coding on material that was not coded for the project.  For each set of cases, a scale was assigned to the label lists, from very exclusive definitions to more inclusive labels, so that for the delineation column the possible values for both ingroups and outgroups are 0-4 (O if it is too ambiguous to code).  The exception is when the Group Reference is to “people in the world.”  Because there is such a wide range of possibilities, it did not make sense to try to set up a scale.  Instead, a list of labels with numbers for each is given.  This will help us see which international actors are talked about and how.  For example, one party may always talk about international allies who are always called “oppressed people,” while another may often mention “democrats.”  Codes for Northern Ireland are given at the end of this appendix.
 

b. Nature of Strategies
        Of the five strategies, some are explicitly more exclusive than others—Enemy Image is considered more exclusive than Governance Positive, for example.  An exclusivity ratio is calculated for each leader by dividing the percentage of exclusive strategies used by the percentage inclusive strategies used in that case.  The exclusive strategies are Justice Negative, Governance Negative, Enemy Image, Storytelling, and Identity (“very exclusive” delineations for particular ingroup references, explained in full codebook).  The inclusive strategies are Justice Positive, Governance Positive, Ally Image, and Identity (“very inclusive” delineations for particular ingroup references).  The leader with the higher ratio is considered more exclusive on this measure, but the degree of exclusivity also depends upon the target of the strategy.   Indeed, an initial reaction upon reading this classification may be to say that an Injustice strategy can be used in a more exclusive or more inclusive way, depending on whether the leader is arguing that everyone in the society is subject to injustice, or if just a subgroup of society is.   Since, as stated before, the reference is the society under examination, the latter is more exclusive than the former.  This idea is not disputed here, but in order to study the differences between more exclusive Injustice strategies and more inclusive Injustice strategies, for example, a second measure of target (or subject) exclusivity is included.  The classification of strategies (not targets) is now explained briefly.
        The Enemy Image strategy emphasizes enemies or scapegoats, and therefore builds boundaries among the groups that are placed in opposition when this strategy is used.  On the other hand, the positive variant of this strategy emphasizes allies and other items two or more groups have in common.  Thus, the former is more exclusive than the latter.  The Justice Negative strategy emphasizes the poor situation in which the ingroup finds itself, and is considered exclusive because it emphasizes ingroup membership and plight, without reference to others.  The Justice Positive strategy is considered more inclusive because the leader is emphasizing gains and how a situation of injustice is improving.
        The use of the Governance Negative strategy can be considered more inclusive than no references to Governance (Positive or Negative).  This is because talk of governance factors such as institutions, laws, negotiation processes, etc. signifies that the leader is at least considering the institutional setting and negotiation with others and possibly is advocating working within the system or government.  This approach is in contrast to extreme groups whose only modus is violence, such as paramilitary organizations.  Since all leaders were found to use at least some of both the positive and negative variants, Governance Negative is considered with the more exclusive strategies and Governance Positive with the more inclusive strategies.
        The Storytelling strategy is considered to add to the exclusivity of leaders’ strategies.  Since this approach emphasizes an imagined community and the bonds among people from centuries past with those of the present and future, it is expected to make people of a group feel more exclusive toward those in other groups.  Indeed, when scholars emphasize this view they argue that leaders draw on the past and the imagined community to promote ingroup solidarity.  The treatment of the Storytelling strategy as a more exclusive strategy receives support from the empirical data as well; it tends to be used overwhelmingly with the Injustice, Enemy Image, and Governance Negative strategies.

c. Nature of Targets
        The “target” is the subject of the strategy, and how exclusive a leader is overall changes depending about whom he is talking.  For example, coding may reveal that both leaders have similar frequencies for the Justice Negative strategy but the targets are very different:  one may argue that the target of deprivation is the entire country while the other claims that the target is only the political minority.  The latter is considered a more exclusive strategy.   For each strategy, an exclusivity ratio for the targets is calculated as follows:  the percentage (of total group references) made to groups considered more exclusive is divided by the percentage of more inclusive references (for each country, the inclusive and exclusive references are specified and justified).   Finally, two summary ratios are created by adding the ratios for the targets of the strategies for which the subject tends to be an outgroup (Enemy Image and Ally Image) and an ingroup (the Storytelling, Justice, and Governance strategies).   These measures give us several different cuts into the data.

d. Exclusivity Index
        For an overall exclusivity score, the three scores described above are summed and weighted by the N (or number of groups) for the case:  nature of strategies, nature of ingroup targets, and nature of outgroup targets.  In all but the first case, the change from the previous time period is also given.  The higher the exclusivity ratio, the more exclusive the leader’s appeals.  Table A2 is an example that shows the following information.  Adams and Hume use a similar number of strategies (650 and 671 respectively), and Adams is a little more exclusive in the strategies he uses.  When it comes to ingroups, Hume mentions many more ingroups than Adams (422 and 344, respectively), but Adams’s ingroups are described in terms that are six times as exclusive as Hume.  Likewise, in terms of outgroups, Adams is much more exclusive.  What this reflects is the way in which Hume often argues in this time period that all people in Northern Ireland are suffering from violence and the lack of proper institutions.
 
 
 

People in world.  Those coded included:  political prisoners; terrorists; United States; proper names of particular parties (e.g. ANC, PLO), states (other than  US), or peoples, when no explicit link is made in terms of ideology; imperialists; the establishment; freedom fighters/anti-imperialists; socialists; nationalists; democrats; Europeans; parties from other countries when a link is made in terms of ideology; oppressed people; totalitarian regimes/repressive regimes; capitalists; Western powers/the West; small nations/developing world; superpowers; ordinary people/people of goodwill; divided societies; republicans.