Cognitive Dissonance in Intervention Work

In my prior article, I took information from the book "Persuasion and Politics: The Social Psychology of Public Opinion," by Michael Milburn to consider methods of understanding attitudes held by people who choose abusive and violent behavior in relationships. As I make my way through the text, I'm finding potential in using the material for intervention work. Writing down my analysis is going to jump around as I process, and as current events spark my desire to explore something within the book.

Like any good psychology student, I remember initially learning about cognitive dissonance, and thought I understood it pretty well. In general, the theory proposes that when an individual has "belief dilemmas" where that person encounters conflict with new information, there's an effort to restore balance to beliefs by changing something within their cognition. Using a direct example from battering intervention work, if I want to control a situation and make my partner do something she does not want to do, and when I do so she becomes upset - and I notice and care about her response - then I will need to change something in my beliefs about controlling her in order to balance my desired result (that I get what I want and my partner goes along with that desire).

The theory is that part of the disconnect with people who choose abusive and violent behavior has to do with not noticing impacts, or caring about their partner's response. So within battering intervention work, we make a lot of effort to raise awareness of impacts on self and others, as well as try to get individuals to be more introspective and self-aware of how chosen behavior is abusive or violent.

But what if it's not that simple, and all these years that I've believed we just need to increase cognitive dissonance aren't exactly striking the chord of changing beliefs and behavior?

As far back as 1958, Fritz Heider proposed "Balance Theory (p. 90 of Milburn's book)" which hypothesizes that triads of relationships that have a positive or negative attribute (valance). He proposed that balance in belief systems needs an odd number of positive relationships (either one or three) to be balanced.

So using current events to illustrate, with the recent Twitter post by President Trump sparking debate about his directly racist statement, if someone supports Trump, but dislikes racism, then to create balance that person would need to either begin to disagree with or dislike Trump, or agree with or begin to like racism.

The problem is that while I learned about cognitive dissonance and balance theory enough to remember them easily, I did not remember the limitations and problems with these theories.

One limitation, and it's a big one, is that when fear or hatred is involved (very strong negative attitudes), individual's cognitions may persist as imbalanced. I cannot count the number of times I have worked with individuals in BIP who hate their ex-partner with such passion that they are unable and unwilling to consider how the damage they cause in that relationships directly damages relationships with their own children with their ex-partner. That hatred is so strong there is no motivation to consider personal choices that are abusive or violent, but rather there is a highly targeted focus on that ex-partner's behavior and why it is wrong.

This means that within intervention work, we need to more strongly consider methods of confronting hatred. Trying to convince someone to have empathy toward a person they hate is most likely going to be unsuccessful because the imbalance in cognition is going to be accepted. No matter the potential harm to themselves and others, an individual entrenched in their hatred will most likely be unable to shift their behavior toward respect and health.

A second limitation has to do with situations wherein an individual holds two strongly held beliefs that contradict each other. The text suggests that researchers were at a loss to account for this inconsistency in beliefs, but offers some suggestions on how people resolve belief dilemmas that may offer insight into how someone can maintain two strongly held beliefs that contradict.

Resolution of Belief Dilemmas:

1. DENIAL: This is the simplest way to eliminate inconsistencies in belief systems, and anyone who works within intervention understands this. You can enter into denial by changing the way one of the objects is valued, or by denying the relationship between the two objects of denial. Fortunately, this is also the weakest method of resolving belief dilemmas, and denial will break down if there are too many inconsistencies, or if there is too much conflicting evidence of other possible beliefs. In general, BIP does a decent job of confronting denial through both methods - introducing and reflecting on inconsistent beliefs, and by offering evidence of the impacts of abusive and violent behavior. 
2. BOLSTERING: When someone adds additional elements to an inconsistent pair of beliefs that serve to overpower another belief system, they bolster one side of the belief dilemma in such a way that the dilemma ends. This is a common challenge in BIP classes, and it is mostly framed as "collusion." When group participants support entitled belief systems, they often do so to bolster their individual sense of being right, and diminish the sense that their partner's perspective matters. Again, in general, BIP is decent at addressing bolstering behavior, and working to get class participants to hold each other to a higher standard - to discuss respectful and healthy beliefs, and bolster the side of the belief dilemma that supports changing behavior. It can be useful to be more cognizant of this process, and why individuals use it to continue hurtful behavior, and also to understand how a focus on discussing respectful and healthy alternatives serves to bolster in a positive way.
3. DIFFERENTIATION: A divide and conquer technique, this resolution involves separating two belief systems into a pair that is consistent, and a pair that is inconsistent - therefore creating an  illusion of balance. There are methods used in BIP to exploit differentiation, and I am not sure I fully agree with the technique, but the ManAlive approach is probably the easiest to describe. As a part of their curriculum, they have individuals in classes identify their "Hit Man" which consists of all the abusive, violent, entitled, and hurtful belief systems. Individuals in the class then compare that to healthy, respectful, and supportive belief systems in an attempt to diminish harm. I am concerned that this can potentially create that illusion of balance rather than actually creating balance by changing beliefs - but I am also willing to recognize that if someone is able to diminish hurtful belief systems through this analysis then that's important work.
4. TRANSCENDENCE: Methods of analyzing belief systems sometimes involve creating reasoning for the beliefs themselves. This in essence is transcendence of the inconsistencies themselves. The example Milburn uses invokes religious perspectives of God, and a dilemma that if God is perceived as pure good, how can God allow evil to exist? To transcend this dilemma, individuals explain this by considering the concept of "free will" and how it's not God allowing evil, but rather individual people choosing the path of evil. By coming up with this reason, it dissolves the dilemma. Consider how frequently individual participants in BIP want to come up with reasons for their behavior, and how often it focuses on a reason that blames others. In BIP, the methods of using transcendence could involve discussing entitlement and how believing you are better than others, believing others are less than you, and believing you deserve something from others allows individuals to be abusive and violent. If that is the reason for hurting others, then it is reasonable to address entitlement and begin to dismantle it to instead create support and care for a partner and for children. 

A caveat to these resolutions is that the researcher who founded the "Modes of Resolutions of Belief Dilemmas," R.P. Abelson, stated that for people whom believe politics are important will likely be more motivated to resolve belief dilemmas than those who do not. This is very true for BIP work, as facilitators will often be focused on change, responsibility, and accountability while participants may not be interested in any of those things. This is why taking a motivational interviewing approach can be so beneficial. 

Further discussion points out that attitudes that are important to an individual are more stable than those that are less important. So in essence, instilling a sense of importance to be respectful and healthy could go a long way toward motivating change in people who choose abusive behavior. The challenge is that often a sense of righteousness is much more important to entitled individuals than health and respect. This means that BIP facilitators need to be mindful of topics that participants are less knowledgeable of. Often this is in topics of respect and health, and while it is important to focus on and discuss abusive and violent behavior, individuals who have been abusive or violent often believe their innocence is the most important attitude, and will find several ways to prove that innocence and ignore identifying how they have been abusive or violent. If we can bolster health and respect, it is more likely that individuals who are closely tied to their belief of innocence will relax those beliefs enough to find methods of change. Researcher Jon Krosnick suggests that when there are two attitudes of equal importance, the above belief resolutions become possible, but in general people will only change their less important beliefs.

When considering how much we focus on belief change in BIP, we need to be much more aware of how this happens. Ableson suggested in his work on cognitive dissonance that beliefs are like possessions - that people hold on to them, value them, and are often reluctant to let them go. It's possible to directly influence changes in beliefs the more we can shift how people view what's important, and how they can connect with alternate perspectives. Entitlement is often very strong for people who choose abuse and violence, and as a result, this entitlement is also of high importance to them. It's possible to create a stronger importance in respect and health, and how we navigate those discussions can make all the difference.

Understanding Attitudes in BIP Work

I'm starting school again this fall, working toward my Ph.D. in Educational Research Methodologies at the University of North Carolina: Greensboro. A big part of that work will involve learning how to accomplish program evaluations, and in creating and understanding measurement tools. It's been my experience that much of the research in intimate partner violence has been heavily skewed toward using recidivism rates to determine "success," and that tools have been relatively poor for assessing abusiveness in a relationship.

As I prepare to get into the thick of things in school, I've been preparing myself in a number of ways. Mostly by taking time to relax before being incredibly busy and overwhelmed, but also by prepping my thinking about research and the problems I have seen by reading the book, "Persuasion and Politics: The Social Psychology of Public Opinion" by Michael A. Milburn. I'm not sure how I heard about the book, and despite its age (published in 1991) I thought it might be interesting to learn more about how political psychologists do research on public opinion and measure people's identification of their values. I think there's a great overlap in domestic violence work and political research. Take this following quote from the introduction:

"...one cannot even begin to understand where political attitudes come from and how they change unless one understands this fundamental truth: Attitudes and behavior are a function of an interactive process between the internal - or what people carry around with them: personality, knowledge, and belief structures - and the external - or what is brought to them from the outside: the influence exerted by other individuals in conversations and through the mass media. Thus there is a dialectical and ongoing interactive process between internal and external forces. Overemphasis or sole emphasis on either the internal or the external forces leads to an incomplete analysis of the dynamics of public opinion (Milburn, 1991, p. 1-2)."

For me, this quote fits perfectly into how good domestic violence intervention work begins. Both internal and external factors need to be discussed and understood both within individuals in the classes, but also from the group process as a whole. To check my biases, and to establish my focus while reading the book, I outlined some fundamental assumptions I have about people who choose abusive and violent behavior in relationships:

1. Most are decent and reasonable people in many layers of their lives;
2. Most will hide patterns of abusive/violent behavior out of feeling ashamed and having a fear of consequences;
3. An accumulation of hurtful behavior influences that individual's values and attitudes;
4. When confronted with the reality of hurtful choices and consequences in a humanistic way, individual people who have been abusive/violent may begin to shift their beliefs and patterns of harm.

In political psychology, there is a constant drive to seek out "public opinion" on various topics. The aggregate of several people's opinion are measured to discover an overall general opinion, and I believe we do this when we analyze participants in BIP in a broad manner. However, individual participants will have values and beliefs that are fluid based on individual relationship dynamics, and these beliefs may shift over time - creating some direct analytical similarities between measuring "public opinion" and in measuring overall understanding of people who are abusive. In fact, finding ways to measure attitudes of abusive individuals could lead to a better understanding of how to assess risk and overall patterns of harm.

Milburn discusses that a part of the difficulty of measuring attitudes is that we tend to assume that we can measure someone's attitude on a topic by simply asking. For some people, this may be true, particularly for attitudes on something an individual cares about, or is central to their personality. With many other people, however, attitudes cannot be measured, because they haven't made a verbal statement of them, even in their own minds. In BIP classes, we confront individuals with questions designed to tease out their attitudes, but in most cases these attitudes are generated following a cognitive review of their knowledge, information, and beliefs about a situation. We're asking people who choose abusive behavior to confront their thinking over behavior that may have had little conscious thought beforehand. Little recognition of beliefs, attitude, and patterns that drove that choice to abuse and to be violent.

Think what that means for research on domestic violence. If an individual abuser hasn't consciously considered personal choice and patterns of behavior, asking questions about abusive and violent attitudes aren't going to be particularly enlightening, because they will not have identified themselves in those terms.

I have seen this firsthand over the years. At Emerge, there is a tool called the "Violent and Controlling Behavior Checklist," which is used at program entry and at program completion. I did some brief measurement of number of identified behaviors at both points and the numbers showed approximately double identification of hurtful behavior after completing the program. Several reasons may be attributed to this, but the ones I tend to think of are that first off the list of behavior is really hard to acknowledge. So people are hesitant to check off everything they have done when first starting out. Also, many of the behavior on the checklist don't resonate with people upon entry - that is they don't fully understand them and don't see how they might have done those behavior, even when they consciously understand the terms being used on the form.

Milburn outlines some of the concepts behind measurement of attitudes by distinguishing between the direction of an attitude (positive or negative) and the strength of the attitude (strong or weak). Behind that there are three components of attitudes: cognitive (what a person believes about an object), affective (how the person feels toward the object), and behavioral (how the person expects to behave toward the object). These three components are linearly related to the behavior in which a person engages.

Think of the layering of attitudes in a BIP class session. Attitude toward the class? Probably a mix of negative and positive, strong and weak. Cognitive beliefs about the class might be that it's a waste of time, is all about the money, is helpful, makes you think, and several others. Affective beliefs might include curiosity, frustration, apathy, interest, and a blend of a multitude of other emotional responses based on the topic discussed. Behavioral beliefs (which I have heard participants describe over time) are expectations of just sitting through the class and saying nothing, thoughts that it's going to be a waste of time and the class will lead to a lot of arguing, plans on being compliant and just making the facilitator do what you want them to - and all these things because in our minds we have plans of behavior when facing unknown circumstances. Again, since attitudes are not established until after analyzing and reflecting on beliefs and making verbal statements about them - it takes time to even process attitudes about BIP classes as a whole. In a completely unscientific analysis at Emerge, I and my co-facilitators would estimate that on average, participants would take about 10-18 classes before they got to a point of understanding the point of the classes it to work to improve respect and health in relationships - not to shame, ridicule, or attack people. Yet, if one were to simply assess attitudes by asking questions to participants early on, chances are much of that understanding would be negative and discouraging about the impact BIP has on the participants.

In the "theory of reasoned action" by Fishburn and Ajzen (in 1975 and 1981), beliefs about the consequences of a particular behavior are linked to the value that a person puts on each outcome. When an individual is able to evaluate outcomes of behavior, and understand their beliefs behind those outcomes - they become more able to see their attitudes. However, when an individual believes that other people might approve or disapprove of a behavior, that person may or may not have motivation to comply due to a perceived subjective norm. In essence, since our society doesn't have strong attitudes against non-physical harms and control in relationships, many of those behavior don't get considered. It's when people choose to physically harm their partner that alarms go off. So these two dynamics indicate that for the most part perpetrators of domestic and sexual violence will easily ignore the outcomes of hurtful behavior unless it is extreme. And even then, there are such a pattern of other hurtful and controlling behavior that have built, the extremes don't seem to be about their beliefs - but rather the behavior of others.

Is it any wonder that the most critical point for intervention is at the start of any work?

The next factor to consider in work with people who have chosen abusive and violent behavior is the "state of consciousness fallacy (Bennett 1980)." In essence, many researchers fall prey to thinking groups of people are a fixed entity whose composition remains the same over time and for different issues. Often, researchers will assume that opinions and behavior of abusers are informed (during BIP classes), stable (in their opinions/values), or consistent (same for all abusers). In domestic violence research, I haven't seen many researchers consider that domestic violence offenders are incredibly ill informed about hurtful behavior in relationships (even during/after BIP classes), are very inconsistent in their attitudes and beliefs, and are highly changeable over time in several different ways. Asking questions on a Likert's style 1-5 rating test aren't going to give you outcomes that are particularly useful because of these inconsistencies.

Recently, on the AQUILA listserv, there has been some discussion about use of the term "batterer" and how that can be construed as name calling.The responses have done everything from reflecting on the history of addressing domestic violence, to the power of nuance, to the suggestion of methods to move beyond labeling.

Referencing the "state of consciousness fallacy," I think we're in a weird place on this issue. In general, most of the "public" are incredibly ill-informed about domestic violence, both in understanding victimization issues, but also in understanding perpetrators. Even among the community of professional intervention workers, our opinions are rather unstable between individuals, and in public the stable opinions I would imagine are dehumanizing about both victims/survivors and perpetrators. I can certainly admit that I fluctuate between an opinion of not using that term, but also in acknowledging that it is an accepted and readily used term within research and writing which leads to me using "batterer" or "abuser" (in fact, in this article I have fluctuated between both options several times). In general, opinions of the term are incredibly inconsistent between professionals as well.

The term is complicated, partly due to history, partly due to overall society having very few nuanced opinions about intimate partner violence. Even the term "domestic violence" is problematic the more we find that non-intimate partner domestic violence cases get referred for BIP.

It is going to be important to continue this dialog about terms we use, but it's also going to be critical to consider how attitudes fit into our work, and how many things we take for granted have several problems. If we ignore those problems, we're going to have a hard time moving forward in our work.

I plan on writing more about what I am finding in this book. The next article will focus a bit more on problems in domestic violence research stemming from sampling errors, response rates, question wording, and social factors in public opinion.