by Michael D. Anestis, M.S.
Recently, we have made a variety of posts regarding the various ways emotions can influence behavior. Amongst these articles were discussions on distress tolerance, rumination, dialectical behavior therapy, and the functional model of non-suicidal self-injury (NSSI). Essentially, these articles provided evidence that many individuals ultimately engage in problematic behaviors in an attempt to regulate their emotions, either by diminishing negative affective experiences or enhancing positive ones. The evidence for an emotion regulation model of behavior is rather compelling, but many such studies rely on an efficient but problematic methodology: retrospective self-report questionnaires.
Almost all researchers use at least some retrospective self-report questionnaires - questionnaires that ask participants to recall past events, trends, symptoms, and characteristics. Such methods are cheap, quick, psychometrically sound, and oftentimes the best available approach. I myself rely heavily upon such questionnaires in my research, so I am certainly not critical of others who do the same. The problem with this approach, however, is that is makes a lot of assumptions about participants' answers and this is particularly true in the case of questionnaires asking about emotions. Think about it. If I asked you what your mood was like yesterday, could you actually accurately report the duration and severity of each emotion you experienced throughout the course of the day? Could you tell me precisely when you began and concluded a behavior and what your emotional state was at each of those time points? Undoubtedly you could for some and not for others but we have no way of knowing when you were accurate and when you were wrong. Now imagine if I asked you to extend that recollection further back than a single day. Impossible, right? So, in addition to assuming that participants are honest and understand the question as asked, retrospective self-report questionnaires on emotions make the assumption that individuals can recall the precise pattern of their fluctuating emotions over an extended period of time. This, quite obviously, introduces a lot of error into empirical findings (and actually speaks to the magnitude of the effects, as researchers still manage to find statistically significant results consistent with their hypotheses).
In an effort to address this area of concern, some pioneering researchers have taken to utilizing ecological momentary assessment (EMA) procedures. EMA involves extending measurements outside the lab by giving participants hand-held computers on which they fill out questionnaires several times each day during the course of several days or even weeks. Participants fill out the questionnaires in response to a series of prompts at regular time intervals as well as immediately following the occurrence of particular behaviors relevant to the study (e.g., NSSI, binge eating). Detailed measures of the participant's mood are taken each time, removing the need to rely on retrospective reports.
Steve Wonderlich and colleagues have been amongst the most prolific researchers with respect to EMA methodology. Their research on eating disorders (see Engelberg, Steiger, Gauvin, & Wonderlich, 2007 and Smyth et al., 2007 for recent examples) has provided compelling evidence that emotional states play an important role in the initiation and reinforcement of binge eating behaviors. In a study just published in Behaviour Research and Therapy, this same group examined the role of emotions in NSSI in a sample of women with bulimia nervosa (BN; Muehlenkamp et al., 2009). 131 women who met DSM-IV-TR criteria for BN were given hand-held computers for a period of two weeks and instructed to fill out a series of questionnaires on the computer several times throughout the day. Signal-contingent recordings - responses prompted by a signal from the computer - occurred at six semi-random points throughout the day. Event-contingent recordings - responses prompted by the occurrence of particular events - were made immediately following the use of any of 19 specific behaviors (e.g., NSSI, binge eating). Interval contingent recordings - responses made at regular intervals - were completed prior to going to sleep. Over the course of two weeks, each participant thus provided a large number of data points, which allows for more complex statistical models of change. In this particular study, the authors wanted to see if NSSI episodes in women with BN were prompted by particular emotional states and whether engaging in NSSI changed the way the women felt in that particular moment. Their findings were highly supportive of an emotion regulation model for NSSI. Specifically, they found that negative affect increased and positive affect decreased immediately prior to self-injury episodes. In other words, the participants were most likely to self-injure in moments when they experienced high levels of negative emotions and low levels of positive emotions. Interestingly, they also found that, immediately following episodes of NSSI, the participants experienced sharp increases in positive affect while negative affect remained unchanged. In other words, immediately after the episode of self-injury, participants felt more positive emotions. Findings like this help explain why a behavior like NSSI can become a pattern. If an individual is upset and highly motivated to quickly feel better, a behavior that immediately causes that exact response will have significant value. In this case, NSSI resulted in individuals who had felt high levels of negative emotions and low levels of positive emotions experiencing sudden surges in positive emotions. The cycle is thus formed in which the individual will turn to self-injury when upset because it has proven itself successful in accomplishing their number one priority - relief - regardless of its long term consequences.
The fact that negative emotions did not decrease immediately following self-injury episodes is mildly surprising, as individuals often report feeling a sense of relief from aversive emotions following episodes of NSSI (Nock & Prinstein, 2004); however, the authors offered some interesting interpretations of these data. First, the authors state that the increase in positive emotions might simply make the experience of negative emotions more palatable. In this sense, it's not that NSSI decreases the intensity of the negative emotions, but it counteracts that feeling by increasing positive emotions. Additionally, the authors posited that the effects of NSSI on the endogenous opioid system may influence how individuals evaluate their emotions (Russ, 1992).
Another important finding in this study was that individuals with BN who did engage in NSSI had more severe BN symptoms than did individuals with BN who did not engage in NSSI. Additionally, this subsection of the sample also had more symptoms of borderline personality disorder. These findings indicate that the presence of NSSI in individuals meeting criteria for BN is a marker for more severe symptomatology potentially requiring more intensive therapeutic interventions (e.g., dialectical behavior therapy).
These findings are important for a variety of reasons. First, they provide compelling support for the emotion regulation model of NSSI proposed in many studies that have utilized retrospective self-report questionnaires to measure emotion. By eliminating the reliance upon participants' memories of their affective experiences, the authors increased confidence that prior findings accurately reflect the ways in which emotions relate to self-harm behaviors. Second, these findings lend further credence to the idea that self-injury is more often a result of a need to change one's own emotional state than an effort to manipulate others (Brown, Comtois, & Linehan, 2002; Klonsky, 2007). There is a distressing amount of misinformation widely available on the topic of NSSI, much of which erroneously depicts individuals who self-injure as manipulative, vindictive individuals for whom there is no effective therapeutic intervention. The hurtful nature of these false claims is deplorable and findings like these can go a long way towards combating misinformation and directing individuals who self-injure towards effective treatments.
If you would like to better understand non-suicidal self-injury, we recommend the following products:
Understanding Nonsuicidal Self-injury: Origins, Assessment, and Treatment
Cognitive-Behavioral Treatment of Borderline Personality Disorder
Understanding Borderline Personality Disorder: The Dialectical Approach
Mike Anestis is a doctoral candidate in the clinical psychology department at Florida State University
