USING THE NEO PERSONALITY INVENTORY WITH COUPLES:

CAPTURING INTERPERSONAL DYNAMICS USING THE CROSS-OBSERVER ANALYSIS TECHNIQUE

Personality assessment is usually accomplished using just self-report data. Obviously, if you want to know something about someone, ask them! Individuals can be very good observers of their inner lives and can share information, under the right circumstances, that is very personal and intimate and not always available to others. Most personality scales are designed only to capture self-report data. However, gathering information from a person is really only one perspective on that person. It is limited by what the individual is aware of about themselves and by what they are willing to share. Fortunately, there are other perspectives (specifically three), less frequently taken, that can provide insights that a person may not be able to provide for him or herself. These other perspectives include life outcome data, which focuses on various outcomes a person has experienced over various time intervals (e.g., graduated college, number of children, civic memberships, etc.); test data, a reference to the use of instruments/machines that track intrinsic processes in ways that are nonobvious to the person providing the information (e.g., a heart or brain scan, projective test); and finally observer ratings, the collection of perceptions of an individual from those who know the person well. The use of observer ratings is an important element to assessment, as it provides information about a person that may not always be under their direct control, specifically their reputation. While we all attempt to positively engage with our social environment, it is ultimately how others see us that determines how our environment may react to us. Observer ratings have been shown to provide information about people over and above whatever information is acquired from the targets themselves. Because observers’ responses to personality questions are not influenced by the same motivations that can distort self-reports (e.g., acquiescence effects, presentation biases either to look good or bad, etc.), observer ratings can provide useful validity information about a scale. Cross-observer convergence is a term that indicates the extent to which scores on a self-report agree with what knowledgeable others have to say about the individual. When a self-report converges well with an observer rating, we know that this association is not confounded by correlated method error (i.e., the same person taking both tests), but instead reflects agreement obtained across different methods.

There are very few personality instruments that are designed to capture observer information; the NEO scales are one of the few to do so. The NEO PI-3, a measure of the Five-Factor Model (FFM) of personality, does have a validated observer report version of the scale. This is important, because there are instruments (e.g., Taylor-Johnson Temperament Analysis) that encourage the use of observer information, but do not have a validated form for use with observers. To be a validated observer form, it is essential that the scale has its own normative information that is based on the distribution of raters scores. The distributions of scores are different for self-reports and observer ratings; even when the observer scores are based on ratings of subjects whose scores constitute the self-report norms. For example, for the NEO PI-3, the combined normative mean self-report T-scores for the five factors are: 92, 121, 117, 108, and 109 for Neuroticism (N), Extraversion (E), Openness (O), Agreeableness (A), and Conscientiousness (C), respectively, The combined rater norms for these domains are: 94, 116, 105, 104, and 102, respectively (data from McCrae & Costa, 2010). These different values may reflect a number of issues, including the differing ways that self- and other-observers rate these qualities, including the observability of the dimensions, and personal style of the different raters. Thus, in using any scale with an observer version, it is important to know that there exists normative information for the set of observer scores. In this way, raw scores can be converted to standardized scores (e.g, T-scores, having a mean of 50 and standard deviation of 10) based on the each distribution’s statistics. This makes it possible to directly compare observer and individual scores using a common metric.

The value of cross-observer convergence is that it provides enhanced depth for making useful and insightful interpretations of personality. The use of these scores provides two sources of information that together can provide more insight than either score alone. I like to refer to this as an “interpretive parallax.” Parallax is a term often used to explain why humans have depth perception. When we view some visual stimulus, that information is acquired from both eyes, each of which sees the information from slightly different perspectives. If one were able to look at what each eye sees, it would be clear that the two images do not perfectly match up; there is a slight difference between the two images. We can see this for ourselves if we take our thumb and hold it straight out in front, and then shut one eye, open it and then close the other. You will see your thumb move back and forth, even though it is not moving! It is this difference, the parallax, that enables the brain to see depth. Combining these different images allows the brain to construct a sense of three dimensionality in our visual perceptions. So, too, with the use of both self and observer ratings. Each assessment gives a view of a common target (the self-rater) from different perspectives. Examining these differing ratings simultaneously can provide deeper insights into the personality of the self-rater. On scales where the two sets of ratings agree, one can be confident that those qualities do characterize the self-rater. Where the do not agree, there may be cause for doubting the validity of the self-rating. At a minimum, there is a need to further explore the dynamics that lead to this discrepancy: There may be a new insight to find!

Having two sets of perspectives provides some very interesting possibilities for assessment work with couples. People who have been in an intimate relationship with each other for an extended period of time would be expected to have a very clear understanding of their partners. After all, they have seen their partners in many different situations and contexts addressing a myriad of issues and tasks. They have seen their partners at they very best and their very worst. They have shared with their partners their own inner feelings and passions, what provokes them, what makes them whole. It is not surprising, then, that when examining cross-observer convergence, the highest levels of agreement are found for married couples (Costa & McCrae, 1992). So the key question becomes, “What happens when convergence is not found?” It is the answer to this question that is at the heart of an interpretive process for assessing couples I have labelled “Cross-Over Analysis” (COA). The value of COA is its ability to capture the interpersonal dynamics of each member of the relationship and to understand how their perceptions of their partner influences both their interpretation of their partner’s motivations as well as how the partner may be viewing and reacting to them.

The COA process begins with each member of the couple completing a NEO PI-3 for themselves and rating their partner. The resulting four profiles can be compared and contrasted to provide insights into those qualities that bring the couple together and those dynamics that may be distancing them from each other. The central interpretive core for this process is to compare the observer rating of the partner with the partner’s self-report. Differences between these two profiles speaks to the types of issues the rater is having with the partner. When the overall level of agreement between these two profiles is low, then the rater may be experiencing higher levels of dissatisfaction in the relationship. When the level of agreement is average or high, then the rater may not be having any issues. It is possible that only one partner may be having a problem, because their level of agreement is low even though their partner’s ratings of them is high. The former may be conflicted in the relationship while the latter may be quite happy.

The NEO PI-3 has 30 scales for making comparisons between the observer and self-ratings. When a significant difference on a scale is found, it may suggest the type of issue the rater is having with the partner. For example, a wife may rate her husband low on C (T-score = 40) while the husband rates himself high (T-score = 60). This difference is significant and indicates that the wife has issues with the husband not pulling his weight around the home. He may spend his time watching TV or engaging in personal hobbies, like golf or fishing, that take him away from the home for long periods. Thus, he does not spend enough time, in her mind, working on maintenance tasks at home or doing his chores, or helping with the children. His focus is more pleasure oriented than in accomplishing the tasks she believes he needs to do. She feels like she has to always be reminding him of the things he needs to do by creating “honey-do lists” or always having to badger him about what he needs to do. She may not be considering the fact that his job requires long work weeks of 60+ hours and that his work requires him to be very involved and engaged in order to maintain his professional and financial success. Exploring why he rates himself high on C and she low on the domain would help them to see the differences in expectations and experiences each has, and by exploring them come a better understanding and accommodation of each other’s needs.

There are two key features of COA. First, it is able to capture the more dynamic aspects of a relationship by simultaneously considering both peoples’ perceptions. COA contrasts the expectations each has of the other and can label these expectations and bring them out for discussion. “Why do you see your husband as being low on C? What does he do to give you this impression?” would be one good question to start a dialogue with this fictional couple. The husband’s response can help begin a process of clarifying the expectations each has of the other. COA reflects a couple’s styles of interaction and processes inherent to their relationship. Second, COA is not about blame or bad people. It is about contrasting these inner expectations and better understanding the perceptions each holds of the other. These are qualities that are amenable to change and improvement. COA promotes intimate dialogue and discussion. Each member of the couple can come away from this process with an enhanced understanding of where their partner is coming from.

In developing this interpretive process, I have come to realize that not all differences are necessarily bad. I have identified some differences that seem to promote intimacy  and some differences are toxic for a relationship (these differences vary for men and women). Nonetheless, this process can be very helpful for couples in building insight and understanding. It is a useful process for premarital counseling as well. COA recognizes that relationships involve interacting with personalities, and how those personalities frame and interpret the actions of their partners is key for success in the relationship.

Listed directly below are citations to two book chapters that I have written, along with a colleague, that outline the nuts and bolts to COA and its application with real couples. Clicking on these citations will bring up a copy of the chapter that you can download for your own library.

Piedmont, R. L., & Rodgerson, T. E. (2017). Cross-over analysis: Using the FFM and the NEO PI-3 for assessing compatibility and conflict in couples. In T. Widiger (Ed.), Oxford University Handbook of the Five-Factor Model. New York, NY: Oxford University Press.

Piedmont, R. L., & Rodgerson, T. E. (2013). Cross-over analysis: Using the FFM and NEO PI-R for assessing couples (pp. 375-394). In P. T. Costa and T. A. Widiger (Eds.), Personality disorders and the five-factor model of personality ( 3^rd ed). Washington, DC: American Psychological Association.

Moving the Paradigm Forward

In an effort to complement the COA process, my wife Rose and I engaged in a process of developing a measure that clinicians could use to quickly identify not only areas of relationship conflict, but to also evaluate the motivational dynamics behind them. This effort resulted in the creation of the Couples Critical Incidents Check List (CCICL). The CCICL was developed with a specific emphasis on content validity. To this end, a number of clinicians who worked with couples in crisis were asked to generate a list of specific behaviors that have been raised as issues of conflict among their clients. We also developed a list of specific behaviors that we be3lieved would reflect the behavioral tendencies of someone very high (or very low) on each of the five factors. This list was then reviewed by another group of clinicians to determine the appropriateness of each behavior item. Then the items were sorted into one of six categories. The first five, Emotional, Interpersonal, Flexibility, Cooperativeness, and Personal Reliability were designed to parallel the five major personality dimensions, respectively. Items included reflected potential problems that someone either high or low on that personality dimension would likely to have (e.g., someone very high on C may be seen by their partner as being too regimented, miserly, and lets work interfere with family time; someone low on C may be viewed as being sloppy/messy, lazy, unfaithful, unorganized, and/or self-centered). The final category, Relationship Context, contained behaviors that were seen as being particularly relevant to understanding conflict and dissatisfaction but could not be classified unambiguously into one of the other categories (e.g., physically abusive, sexual difficulties, gambler, mocks me in front of others). The scale also contains a single item rating concerning the degree of dissatisfaction the respondent is experiencing in the relationship.

Data on this instrument and evidence of its convergent and discriminant validity can be found in Chapter 5 of my NEO book (Piedmont [1998] The revised NEO Personality Inventory: Clinical and research applications. New York, NY: Plenum). Copies of the CCICL can be purchased on the Center for Professional Studies’ website. The validity data for the CCICL comes from the dissertation work of one of my students, Robert Kosek (Kosek, R. B. [1996] Criss-cross ratings of the Big Five personality dimensions as an index of marital satisfaction. Unpublished doctoral dissertation, Loyola College, Baltimore). The obtained data clearly demonstrate the value of the CCICL. The overall dissatisfaction rating correlated over r = .90 with scores from the Locke-Wallace Marital Satisfaction Scale, indicating the value of this simple rating for plumbing the levels of relationship satisfaction. More importantly, it was demonstrated that ratings on the CCICL did not correlate with the rater’s own self-report scores, indicating that these ratings were not the product of the rater’s issues or imagination. Second, ratings on the CCICL scales correlated significantly, and in appropriate ways, with both the rater’s ratings of his/her partner and with the partner’s own self-report.

The CCICL represents a simple, quick to administer and score instrument that can not only identify specific behavioral elements that are provoking distress in a relationship, but can also identify those personality motivations of the target person that may be contributing to these levels of distress. The above cited NEO book provides several case studies using both the COA method and the CCICL. Interested readers can obtain the chapters listed here and extend their understanding of the value of this approach. Also, the Center is developing a training seminar in the use of these techniques to complement its already popular NEO Training Session, which provides an introductory overview of the NEO, its applications and interpretive strategies. If you have any  further questions please feel free to contact me at drralph@thecfps.com.