Introduction

            “In the public sector, conflict is necessarily resolved within a framework of policy making” (Stephenson & Pops, 464.)  The universe of text surrounding conflict in public administration is quite dense; however, as Zhiyong Lan observed, not many efforts have been published with systematic attempts to understand conflict resolution in the public administration realm (Lan, 1997.)  Further she observes that most of the studies in the public administration conflict universe describe “the conflictual nature of public administrator’s work or…microlevel (individual and organizational) conflict in public organizations” (Lan, 27.) 

            This being the case for macrolevel public administration conflict does not apply across the board with conflict resolution in general.  The nature of conflict and conflict resolution, in general, is well researched and documented.  Attitudes and responses to conflict range from: an implicit belief that conflict is a destructive force that should be closely managed and resolved to the belief that there are positive aspects to conflict.  Some positive aspects attributed to conflict in public administration include: establishing group identity, strengthening internal cohesion, stabilizing loosely structured organizations and creating freedom for individuals to choose their own course of action (Lan, 1997.)

            In this study I will not address the value-based nature of conflict.  As I see it, there is both good and bad in even the reasons listed above as positive aspects of conflict.  Cohesion and stronger group identity sounds like a promising aspect until taken to the level of fanaticism and bitter partisanship.  On the other hand, I agree with the sentiment that it seems the United States government functions best when conflict prevents much from getting done at the federal level.  The dialectical attitudes regarding conflict in public administration settings only serve to underscore the fact that conflict does exist in many forms and in the case of public administration conflict effects policy decisions.

            The United States’ checks and balances system of government inherently generates conflict by forcing conflict interaction between organizations.  All policy implemented carries consequences for the organization developing and carrying out the policy as well as consequences for those people the policy is intended to serve (Stephenson & Pops, 1989.)  Due to the fact that one branch of government (i.e. legislative) may establish policy for another (i.e. executive,) interaction between those departments is crucial to the effective governance of state business.  As Daniel Canary points out, “Interactions with other people occur in episodes that vary in terms of their conflict potential” (Canary, 520.)  He goes on to observe that several factors such as network loyalties and even the characteristics of the environment can influence the severity of conflict encountered in any given situation (Canary, 2003.) 

            Many factors play into why conflict arises as well as how intense conflict gets.  This is because underlying rules govern interactions with others.  These rules are generally not explicitly expressed, but are largely implied or even ingrained in the individual.  Littlejohn and Domencini observed, “A common rule is to defend yourself when under verbal attack” (Littlejohn & Domencini, 26.)  Outside of culturally derived understood rules, personality factors also seem to play an important role in skillfully managing conflict.  One important characteristic of skilled conflict managers is personal control.  Daniel Canary posits that people with an internal sense of control utilize cooperative tactics more readily and in the end enjoy better relational outcomes (Canary, 2003.)

            External factors also impact conflict primarily in the methods of resolution.  When examined in a framework of the decision mode established, certain methods of resolving conflict appear more likely to be successfully employed than others.  Among these external factors that impact conflict resolution are: stage of the policy process and the mandates created by existing law (Stephenson & Pops, 1987.)  In the 1990’s, the U.S. Congress observed the negative financial and relational impacts of adjudication within public administration and created a series of laws to address conflict resolution within the United States public administration organizations.  These laws supported, established or required federal agencies and departments to seek alternative measure to resolve conflicts and administrative disagreements.  These laws included: the Civil Justice Reform Act of 1990, the Administrative Dispute Resolution Act of 1990 and the Negotiated Rulemaking Act of 1990 (Lan, 1997.)

            By establishing these laws, Congress demonstrated an understanding of the need and desire to resolve disputes peacefully.  However, there may always a general worry- among the populace- expressed in a quote attributed to Ralph Waldo Emerson: “Democracy becomes a government of bullies tempered by editors” (www.quotationspage.com.) 

            A system is necessary to attempt to measure the level of trust that should be encouraged with regard to conflict resolution at the national level in the United States’ government.  The level of trust is a key factor in determining whether negotiation takes the route of positional bargaining or interest-based bargaining (Stephenson & Pops, 1987.)  It has been established in public administrative conflict resolution literature that at least three interested parties comprise the actors in conflict resolution.  Those parties have been labeled as the interested audience, the parties conflicting and those individuals who resolve the conflict (Lan, 1997.)  In my content analysis of a Congressional case study, I will attempt to establish a face work coding scheme based on ideas first expressed by Erving Goffman that may allow for a deeper understanding of what is taking place in any organizational conflict conversation.   Specifically, I looked at the contemporary issue of Iraqi refugees which has been labeled a crisis.  Because the universe of applicable texts is sparse concerning this type of system- applied directly to public administration- I consider this to be an exploratory study and have therefore offered no hypotheses.

 

Iraq Refugee Literature Review

            In 2006, it became apparent that a large number of people living in Iraq began leaving the country largely due to civil unrest.  Many of these people are now living in neighboring countries, such as Jordan Syria and Lebanon.  These mass migrations have resulted in hardships for those countries as well as the immigrants themselves.  It did not take long for eyes to turn toward the United States to see what we were doing to help Iraqi immigrants.  In some of the more well-publicized cases, the Iraqis immigrating out of Iraq were leaving because their lives were being threatened for having helped U.S. war efforts in Iraq (Committee on the Judiciary, 2007.)

            In fiscal year 2007, 1,600 Iraqi refugees were allowed to permanently resettle in the United States (Church World Services, 2007.)  This number represents a very small fraction of the 2 million Iraqis believed to have left the country and the 2.5 million Iraqis believed to be internally displaced within Iraq’s borders (Church World Services, 2007.)  For fiscal year 2008, the U.S. has committed to resettling 12,000 Iraqis largely in response to public criticism of fiscal year 2007’s efforts.  At the center of this U.S. conflict is the executive branch- represented by Assistant Secretary of State Ellen Sauerbrey- and the legislative branch- represented in this study by the Committee on the Judiciary.

            In December 2006, as this issue seemed to be getting heated up, Ellen Sauerbrey spoke primarily of reasons for why permanent Iraqi resettlement in the United States was taking such a long time and allowing so few individuals to immigrate here.  Some of the reasons given by Ms. Sauerbrey for the delays were attributed to coordination issues with the United Nations’ High Commissioner for Refugees organization responsible for identifying refugees for U.S. immigration and the logistics of the president determining the number of refugees permitted into the country every year (Federal News Service, 2006.)  The Assistant Secretary’s solution to this problem, in December 2006, seemed to rest on providing better assistance to the countries housing most of the immigrants and better identifying the most vulnerable Iraqis who may never return to their homeland (Federal News Service, 2006.) 

January 16, 2007 the United States Senate Committee on the Judiciary held a formal hearing on “the Plight of Iraqi Refugees.”  Assistant Secretary Ellen Sauerbrey testified at the hearing along with witnesses coming from varying perspectives of the refugee problem.  Included in the group of witnesses were: two Iraqi immigrants, an United Nations representative, two charity organization representatives and a U.S. marine formerly stationed in Iraq.  This hearing elicited 48 pages of spoken transcripts plus many more pages of documents submitted for the record (Committee on the Judiciary, 2007.) 

            Since the hearing, Senators Ted Kennedy and Carl Levin announced late in September of 2006 the Senate passed the Refugee Crisis in Iraq amendment (News Conference, 2007.)  Cited in their announcement as a reason for passing the amendment is the fact that “there’s been huge bureaucratic hurdles to [Iraqis] being accepted as refugees” (News Conference, 2007.)  As of the writing of this report, the House has not yet passed their portion of this resolution into law yet.

            This issue seems like a fruitful example of conflict in public administration because what is being debated between the executive and legislative branch here does not seem to be a highly contentious issue.  It is a humanitarian effort resulting from the war in Iraq.  The issue of who started the war is quite contentious but virtually no one in either organizations have come out saying the United States shouldn’t allow more Iraqis resettle within U.S. borders as long as certain conditions are met.  The seemingly universally accepted conditions for Iraqi resettlement into the U.S. are either: the Iraqi’s life is being threatened for having served the U.S. military (e.g. as a translator) or by being part of the Christian minority from Iraq (i.e. Chaldeans and Assyrians.)

 

Face Work

            To better understand what’s taking place during these U.S. Senate negotiations, I have borrowed a face work coding scheme developed for understanding FBI crisis negotiations (Rogan and Hammer, 1994.)  Before getting into the scheme, it is important to understand the concept of face.  Face is a concept that emerged from the work of Erving Goffman.  Goffman has explained elements of face utilizing a theatre metaphor (Goffman, 1959.)  Goffman asserts in this metaphor that, “When an individual plays a part he implicitly requests his observers to take seriously the impression that is fostered before them” (Goffman, 17.)  This “performance” takes place at all times.  It is the image projected by any individual interacting with other individuals. 

The projected image can be thought of as the face of the person performing.  Face, however, does not just involve the person performing but also requires another person to observe the performance.  Rogan and Hammer explain, “Face then is a concern for one’s projected image that is both immediate and spontaneous and is tied to the dynamics of social interaction” (Rogan and Hammer, 217.)  Necessarily tied to face then are feelings of honor and respect (Littlejohn and Domencini, 2001.)  A person gains self-worth by having their face (or performance) accepted by those with whom they interact.  When the performance of face is not accepted the individual performing is said to lose face (Rogan and Hammer, 1994.)

 

A Three-Dimensional Operational Model of Facework

            In Rogan and Hammer’s study of crisis negotiations, they established three dimensions related to coding facework behavior in crisis negotiations.  Their three dimensions included: locus of concern, face valence and temporality.  The locus of concern indicates the direction in which a face statement is aimed.  The locus of concern will either relate to the self or someone other than the self.  The face valence denotes whether face is being honored or threatened.   Finally, the temporality function proved problematic in their study because it required a coder to determine whether an honor face statement was intended to proactively defend face or retroactively restore face (Rogan and Hammer, 1994.)  Figure 1 illustrates their final coding scheme.

 

Crisis Negotiations Coding Scheme

Figure 1:  Three-dimensional model of facework (Rogan and Hammer, 218.)

 

 

            Their coding scheme resulted in an intercoder reliability kappa of .70 (Hammer & Rogan, 221.)  However, since the conversations taking place in the U.S. Senate are not crisis negotiations (as well as taking into account areas Rogan and Hammer said they struggled with the model) I made a few changes. 

            The first change I made was related to their discovery of the difficulty inherent in trying to determine temporality.  As practice- before attempting to code the entire Senate transcript- I elected to code two shorter senator transcripts utilizing the code shown in Figure 1.  I can confirm that temporality is incredibly subjective and difficult to identify in the course of an ongoing dialogue. 

In order to still retain the as much robustness as possible from their coding scheme, I elected to simplify the face valence dimension to just attack and defend.  This took away the need for the third dimension of temporality since all face restoring statements would be interpreted as defenses of face regardless of being retroactive or proactive. 

            I quickly discovered after briefly reading the transcripts of the Committee on the Judiciary’s hearing that a third dimension was needed.  I’ll call this third dimension the point of view.  Point of view relates to team concepts explored by Goffman (Goffman, 1959.)  Goffman suggested that “when two teams present themselves...they tend to stay in character” (Goffman, 167.)  As a group, they essentially build a collective face.  This may be related to Goffman’s concept of a performer taking on qualities associated with a task or vocation into their own performance while on duty.  As Goffman said, “[I]t often happens that the performance serves mainly to express the characteristics of the task that is performed and not the characteristics of the performer” (Goffman, 77.) 

            I observed in the Senate transcripts that many references referring to the self actually referred to the group that the performer was representing rather than them personally.  This dimension of point of view also cropped up in regard to statements referring to the other.  In some cases the other referred to wasn’t present in the room.  In some cases, these other statements seemed to refer to an entire organization that couldn’t directly be credited to any one individual at the hearing.  To alleviate these issues, under the third dimension point of view I added for the self references I or We to indicate a personal reference or a reference intended to refer to the organization that individual represented.  For the other statements, I added in the point of view dimension an In Room and Out of Room code.

            Finally, I found I was able to remove the neutral valence code as a matter of choosing a different unit of analysis that will be discussed later.  These changes are illustrated in Figure 2 below.

 

Congressional Face Work Coding Scheme

 

Figure 2:  Three-dimensional model of facework for congressional study

            The coding scheme in Figure 2 yields 8 codes: Attack Self Face I (ASI,) Attack Self Face We (ASW,) Defend Self Face I (DSI,) Defend Self Face We (DSW,) Attack Other Face In Room (AOI,) Attack Other Face Out of Room (AOO,) Defend Other Face In Room (DOI,) and Defend Other Face Out of Room (DOO.) 

            Attack statements in this scheme are not limited to obviously hostile statements that would make another person feel poorly about themselves, but also include any types of challenges to a person’s well being.  In many cases, the way that questions are addressed to another form an attack on the other individuals face because they can either challenge veracity or demand justification in situation where justification may not otherwise be offered.  Defense statements in the setting of a Senate hearing take place often in the form of ceremonial-type statements (e.g. “thank you chairman”) that support the face and formality of the Senate committee proceedings.

Methods

            The purpose of this study was to examine the frequency of facework as operationalized in the three dimensional model illustrated above.  Lacking existing research for this type of study I did not start with any hypotheses, however, I did want to try to answer the following research questions:

RQ1:  What facework behaviors are present in congressional interactions?

RQ2:  To what degree is the three-dimensional coding scheme used in this study a valid measure of facework behavior?

RQ3:  What significant behavior patterns emerge from this study?

 

            The transcript obtained for this study is publicly available on the Government Printing Office website (www.access.gpo.gov.)  The actual transcript is 218 pages long which include many pages of documents submitted for the record as well as written questions and answers that did not take place during the hearing.  For this study I only coded facework statements that occurred during the hearing face to face.

            The unit of analysis for this study was the uninterrupted consistent facework statement.  By unitizing this way I hoped to avoid two situations occurring.  The first situation I hoped to avoid a problem encountered by Rogan and Hammer in their crisis negotiation study.  In that study they indicated that the uninterrupted talking turn essentially contained too many facework statements making consistent accurate coding difficult (Rogan & Hammer, 1994.)  They suggested for future studies to utilize thought units as a unit of analysis; however this seemed unclear in segmenting the transcript and it also seemed to result in too many repetitious codes in a row.  Therefore, the uninterrupted consistent facework statement is intended to address a face tone from its beginning until it transitions into different dimensions.  Units of analysis yielded 938 units that broke down fairly evenly between congressmen (n = 471) and witnesses (n = 467.)

            Coding was completed by me alone; therefore, reliability of this coding scheme has not been verified and is a very strong indicator to treat all results from this study as tentative until the reliability of the scheme can be verified through the use of independent trained coders. 

 

Results

            The data were analyzed for frequency.  In follow up studies, I plan to analyze the data sequentially as well.  The frequencies of each type of facework for congressmen (no women of the Senate were present) and witness were computed combined and separately as groups for each code in the scheme and frequency was also calculated for face valance and locus of concern statements without point of view taken into account.

            As can be seen in Table 1, the most frequently used code of each of the eight codes for both congressmen and witnesses combined was Defend Other Face In Room (n = 222, 23.67%.)  The descending order of frequency for the remaining codes breaks down as follows: Defend Self Face We (n = 179, 19.08%,) Defend Self Face I (n = 152, 16.20%,) Defend Other Face Out of Room (n = 150, 15.99%,) Attack Other Face In Room (n = 110, 11.73%,) Attack Other Face Out of Room (n = 70, 7.46%,) Attack Self Face We (n = 50, 5.33%,) and Attack Self Face I (n = 5, 0.53%.)

 

 

 

 

 

 

TABLE 1

Frequency and Percentage of Facework Codes for both Congress and Witnesses

 

 

            Defend Face statements (n = 703, 74.95%) occurred more frequently when looking at both Congress and witnesses than Attack Face statements (n = 235, 25.05%.)  With point of view removed, Defend Other Face statements occurred most frequently (n = 372, 39.66%) followed by Defend Self Face statements (n = 331, 35.29%,) Attack Other Face statements (n = 180, 19.19%) and Attack Self Face (n = 55, 5.86%.)

            Table 2 presents frequencies of facework codes for the congressmen present only.

The most frequently used facework code for congressmen was Defend Other Face In Room (n = 146, 31.00%.)  The rest of the eight code scheme in descending order of frequencies were: Attack Other Face In Room (n = 101, 21.44%,) Defend Self Face We (n = 65, 13.80%,) Defend Self Face I (n = 63, 13.38%,) Attack Self Face We (n = 38, 8.07%) and Attack Other Face Out of Room and Defend Other Face Out of Room occurred with the same frequency (n = 29, 6.16%.)  There were no coded statements of Attack Self Face I attributed to the congressmen involved.

 

 

TABLE 2

Frequency and Percentage of Facework Codes for congressmen only

 

 

            Defend Face statements (n = 303, 64.33%) occurred more frequently than Attack Face statements (n = 168, 35.67%) for congressmen; however, the variance between frequency is less than when both congressmen and witnesses facework codes were combined.  With point of view removed, Defend Other Face statements (n = 175, 37.15%) occurred most frequently followed by Attack Other Face statements (n = 130, 27.60%,) Defend Self Face statements (n = 128, 27.18%) and finally Attack Self Face statements (n = 38, 8.07%.)

            Table 3 presents frequencies of facework codes for the witnesses only.  The most frequently occurring facework statement for the witnesses at the hearing was Defend Other Face Out of Room (n = 121, 25.91%.)  The other facework coded statements occurred in the following descending order: Defend Self Face We (n = 114, 24.41%,) Defend Self Face I (n = 89, 19.06%,) Defend Other Face In Room (n = 76, 16.27%,) Attack Other Face Out of Room (n = 41, 8.78%,) Attack Self Face We (n = 12, 2.57%,) Attack Other Face In Room (n = 9, 1.93%,) and Attack Self Face I (n = 5, 1.07%.)

 

 

 

 

 

 

 

 

TABLE 3

Frequency and Percentage of Facework Codes for witnesses only

 

            Defend Face statements (n = 400, 85.65%) occurred more frequently than Attack Face statements (n = 67, 14.35%) for witnesses; however, the variance between the frequencies for attack and defend was greatest for the witnesses.  With the point of view dimension removed, Defend Self Face statements (n = 203, 43.47%) occurred most frequently followed by: Defend Other Face (n = 197, 42.18%,) Attack Other Face (n = 50, 10.71%) and Attack Self Face (n = 17, 3.64%.)

 

Significant Findings

            Although reliability for this study is low due to the fact that there was only one coder, there are a couple of hopeful signs for construct validity.  In a congressional television spot study conducted by Brazeal and Benoit, campaign television ads were coded for the occurrence of acclaims, attacks and defenses by the congressional candidates.  Though their coding scheme doesn’t match mine perfectly, it isn’t too much of a stretch to say that acclaims and defenses in their study would nearly match up to Defend Face statements in my study, while attacks in their study would match the Attack Face statements in my study.  In this regard, I am simply matching the tendency for a congress person to utilize a particular face communication strategy without regard to locus of concern or point of view (it is to be assumed that if a congress person were to pay money for a television ad during campaign season that their locus of concern and point of view can be assumed.)  In Brazeal and Benoit’s study, they found that of all the spots they watched 64% contained acclaims and 35.5% contained attacks (Brazeal & Benoit, 2001.)  When congressmen’s data only in my study is washed for only valence, the data shows that 64.33% of the facework statements were Defend Face statements and 35.67% were Attack Face statements.

            Furthermore, in Rogan and Hammer’s crisis negotiation study negotiator and perpetrator facework statements were analyzed separately.  As I’ve alluded to earlier, there isn’t a lot of room for comparing my study with theirs; however, one way in which the studies to match up nicely is in the paradigm of controller and reactor.  In the Rogan and Hammer study, the negotiator was the controller and the perpetrator was the reactor (Rogan & Hammer, 1994.)  In my study, the congressmen were the controllers and the witnesses were the reactors.  When analyzed with this point of view, there is an interesting ordinal similarity with the frequency of facework codes. 

            Negotiators code frequency descended like this: Defend Other Face 63%, Defend Self Face 12%, Attack Other Face 0% and Attack Self Face 0%.  When analyzed without point of view involved, the congressmen’s code frequency is similar in an ordinal fashion.  The descending code frequency for congressmen is Defend Other Face 37.15%, Attack Other Face 27.60%, Defend Self Face 27.18% and Attack Self Face 8.07%.  The ordinal deviation in Defend Self Face and Attack Other Face may be explained by the fact that Rogan and Hammer did not code any Attack statements for the negotiators and that makes sense because they were dealing with highly volatile situations.  From the reactors perspective, the Perpetrator’s code frequency descends as follows: Defend Self Face 60%, Defend Other Face 7%, Attack Self Face .5% and Attack Other Face 0%.  The witness’ code frequencies again match up with minor deviation: Defend Self Face 43.47%, Defend Other Face 42.18%, Attack Other Face 10.71% and Attack Self Face 3.64%.  Again the ordinal ranking is not a perfect match, but it is understandable when it is considered that the negotiators and perpetrators were having a two party conversation while congress and the witnesses were conducting multiparty talks.  It seems that attack face statements would be less likely in a two party conversation.  The interesting aspect to note however, is that the top two facework codes for both controllers and reactors were the same in both studies even though percentages differed.

 

Discussion

            First, and foremost, the data yielded in this study is preliminary at best.  This study was conducted as an exploratory investigation of congressional facework behavior.  The findings are reported as descriptive statistics solely for the case studied.  Having stated that, I believe the answer to my second research question is that this study also offers some encouraging signs for the three-dimensional facework model employed.  The model seems to possess construct validity; however, it needs to now be further investigated for coding reliability before being utilized for future research purposes.

            My first research question was answered in the results section.  As for my third research question, the significant facework behavior patterns that emerged during this study center primarily around the nature of attack and defend facework utilized during the hearing.  It is interesting that there were no recorded instances of Attack Self Face I codes for any of the congressmen.  In turn, the congressmen’s heavy use of Defend Other Face In Room seemed to be a bit counter intuitive in relation to their heavy frequency of Attack Other Face In Room facework codes.  This is more understandable when it is realized that the most frequently recorded Defend Other Face In Room codes came from Ted Kennedy (n = 97) and he led the hearing.  He was also the primary supporter of resolving the Iraqi refugee problem on the committee and he had the most recorded unit (n = 215.)  Senator Arlen Specter had the second most recorded codes (n = 142) and he was also coded most frequently of anyone else utilizing Attack Other Face In Room statements.  Another interesting aspect of these proceedings was that the congressmen were coded using attack facework more than the witnesses and, in turn, witnesses used defend facework more than the congressmen.  It would seem this presents interesting implications in regard to principled bargaining and the ideals of democracy in the United States.

 

Implications for Future Research

            Any future research regarding this coding scheme must first focus on establishing coding reliability.  Assuming that the scheme is reliable though, it seems that this coding scheme could be used in conjunction with theories such as principled bargaining and democratic theory for levels of assessment.  On the more practical side, it has been suggested by Roloff, Putnam and Anastasiou that skilled negotiators avoid defend and attack spirals (Roloff, Putnam & Anastasiou, 2003.)  With a reliable coding scheme like the one presented here, negotiators and politicians could better understand their style of negotiation behavior in order to improve their success rate of conflict resolution.

References

Committee on the Judiciary, U.S. Senate.  (2007).  The Plight of Iraqi Refugees (S. HRG 110-3 Serial No. J-110-2).  Washington, DC: U.S. Government Printing Office.

 

Fisher, Roger & Ury, William.  (1981).  Getting to Yes.  New York, NY: Penguin Books.

 

Goffman, Erving.  (1959).  The Presentation of Self in Everyday Life.  New York, NY: Anchor Books.

 

Lan, Zhiyong. (1997).  “A Conflict Resolution Approach to Public Administration.”  Public Administration Review, 57 (1), 27-35.

 

Kelman, Steven.  (1992).  “Adversary and Cooperationist Institutions for Conflict Resolution in Public Policymaking.”  Journal of Policy Analysis and Management, 11 (2), 178-206.

 

Rogan, Randall G. & Hammer, Mitchell R.  (1994).  “Crisis Negotiations: A Preliminary Investigation of Facework in Naturalistic Conflict Discourse.”  Journal of Applied Communication Research, 22, 216-231.

 

Stephenson, Max O. & Pop, Gerald M. (1989). “Conflict Resolution Methods and the Policy Process.”  Public Administration Review, 49 (5), 463-473.