Explanatory style as a mediator between childhood emotional abuse and nonsuicidal self-injury.
Nonsuicidal self-injury (NSSI) has been defined as the direct, deliberate, self-inflicted damaging of a superficial or moderate amount of bodily tissue that is performed without the intent to die (Favazza, 1998). Following Favazza's classification system (1998), NSSI is differentiated front self-injurious behaviors that are sanctioned by social norms (e.g., body piercings or cultural rites) and from self-injurious behaviors that damage a major, rather than a superficial or moderate, amount of tissue. Common methods of NSSI are cutting, burning, hitting self, biting self, and severe scratching (Claes, Houben, Vandereycken, Bijttebier, & Muehlenkamp, 2010; Gratz & Chapman, 2007; Lloyd-Richardson, Perrine, Dierker, & Kelley, 2007).
Recent studies have reported high NSSI prevalence rates, particularly among adolescents and young adults. For example, Lloyd-Richardson et al. (2007) found that 46% of high school students had performed NSSI in the previous year. Similarly, Aizenman and Jensen (2007) reported that 41% of college students engaged in NSSI at some point in their lives. In clinical settings the rates are even higher: 82% of adolescent psychiatric inpatients were found to have performed at least one incident of NSSI in the previous year (Nock & Prinstein, 2004).
Despite the prevalence of NSSI, many mental health professionals are believed to be ill-equipped to provide care for self-injuring clients (Allen, 1995; Walsh, 2007; White, McCormick, & Kelly, 2003). One problem may be a general uncertainty about the causes of NSSI. In recent years numerous studies have investigated the prevalence, methods, functions, and psychological correlates of NSSI (Aizenman & Jensen, 2007; Claes, Klonsky, Muehlenkamp, Kuppens, & Vandereycken, 2010; Klonsky, 2007; Nock & Prinstein, 2004). However, there has been scant attention to its etiology. For example, although it is commonly understood that NSSI is associated with a history of childhood abuse (Glassman, Weierich, Hooley, Deliberto, & Nock, 2007; Gratz & Chapman, 2007), there is little understanding of the pathways through which previous abusive experiences may be related to later NSSI (Yates, Tracy, & Luthar, 2008).
Furthermore, although the emotional characteristics of self-injurers have been studied (e.g., Gratz & Chapman, 2007), less is known about their beliefs, their cognitive characteristics. More information about their cognitions, as well as NSSI etiology, might inform treatment planning. This study sought to address these limitations. It tested a theory-driven mediation model of NSSI that featured pessimistic explanatory style as a cognitive mediator between childhood emotional abuse and frequency of NSSI in the previous year. Model design was guided by the literature on these variables.
The first pathway of the proposed model concerned the relationship between childhood emotional abuse and NSSI. Emotional abuse refers to hostile behaviors, verbal or nonverbal but excluding physical contact, directed by an adult toward a child that endanger the child's psychological or physical well-being (Keashly & Harvey, 2005; McGee & Wolfe, 1991). Within North American culture, examples of such childhood emotional abuse are humiliation, derogation, and intimidation (Nicholas & Bieber, 1997). Both qualitative and quantitative studies have supported a relationship between childhood emotional abuse and NSSI (Glassman et al., 2007; Harris, 2000).
The second pathway concerned the relationship between childhood emotional abuse and pessimistic explanatory style. Explanatory style is a cognitive characteristic that pertains to a person's habitual way of interpreting the causation of events (Peterson & Park, 2007). A pessimistic explanatory style views the causes of negative events as (a) stable (vs. unstable, or transient) in duration; (b) global (vs. specific, or circumscribed) in their generalization to various aspects of life; and (c) internal (vs. external) to the person (Peterson & Park, 2007). For example, after failing a major exam, those with a pessimistic explanatory style might identify the cause as their own "laziness" (an internal cause), which is perceived as impacting several different areas of life (a global cause) continuously (a stable cause). By contrast, an optimistic explanatory style would be exemplified by those who recognize that they failed the exam because of having had too little time to study (perhaps on account of helping a friend), which is a causal explanation that is external to the person, isolated as a rare occurrence, and unlikely to influence other areas of life.
The construct of pessimistic explanatory style is a theoretical extension of the learned helplessness literature (Abramson, Seligman, & Teasdale, 1978). In theorizing about how children develop a pessimistic explanatory style, Rose and Abramson (1992) argued that emotional abuse is a particularly strong contributor. Unlike physical and sexual abuse, which do not necessarily involve verbal exchanges between a child and an adult, emotional abuse consists largely of an adult's hostile verbal communications to a child. These communications, according to Rose and Abramson (1992), often transmit to the child stable, global, and internal causal explanations for negative events (e.g., "You are a terrible, worthless child and all of our difficulties are your fault"). Over time, emotionally abused children who accept these explanations as true may develop an explanatory style marked by stable, global, and internal causal explanations. For example, they may conceptualize their role in negative events by thinking, "I'm a terrible person who always messes everything up." This cognition reflects the pessimistic dimensions of internality ("I'm a terrible person"), stability ("always"), and globality ("everything").
Empirical studies have supported the link between childhood emotional abuse and pessimistic explanatory style but failed to find a significant relationship between pessimistic explanatory style and physical or sexual abuse (Gibb et al., 2001; Hankin, 2005). Hence, the proposed mediation model specifies only childhood emotional abuse as a predictor of pessimistic explanatory style.
The relationship between pessimistic explanatory style and NSSI, which comprises the third pathway of the model, had not been studied at the time of this writing. However, Suyemoto and MacDonald (1995) theorized that control-related cognitions were factors in precipitating NSSI. Self-injurers, they thought, tend to experience helplessness; they engage in NSSI to obtain more control over emotional experiences and the external environment. Pessimistic explanatory style, as a control-related cognition with theoretical ties to learned helplessness, may be among the beliefs associated with NSSI.
Support for this pathway may be found in research on similar cognitive characteristics. Glassman et al. (2007), for example, documented the association between NSSI and self-criticism, a cognitive characteristic related to the internality dimension of pessimistic explanatory style. Similarly, Aizenman and Jensen (2007) reported a significant relationship between NSSI and the belief of having reduced control over important situations. Qualitative researchers (Moyer & Nelson, 2007) have also found that self-injurers frequently report a sense of helplessness and loss of control over their lives.
The mediation model proposed here was designed to extend research by Glassman et al. (2007), who reported that, of various types of childhood mal-treatment, only emotional abuse had a significant relationship to both NSSI and self-criticism. Self-criticism also partially mediated the relationship between childhood emotional abuse and NSSI. The present study tested an additional cognitive variable, pessimistic explanatory style, as a potential mediator between childhood emotional abuse and the frequency of NSSI.
The research hypotheses of this study were that
1. Pessimistic explanatory style will be positively related to frequency of NSSI in the past year.
2. Pessimistic explanatory style will mediate the relationship between childhood emotional abuse and frequency of NSSI in the past year so that the relationship between childhood emotional abuse and NSSI will be significantly smaller when pessimistic explanatory style is included in the regression model.
We used a quantitative, correlational design to test the mediation model. Data were collected at one point in time by means of self-report surveys.
The sample (N = 390) was composed of students aged 18-25 at a large private university in the northeastern United States; their mean age was 20.3. Participants were predominantly female (66%), heterosexual (95%), and Caucasian (72%); 13% were Asian/Pacific Islander, 7% were African American, 3% were Hispanic/Latino/a, 3% were Multi-Ethnic/Multi-Racial, and 2% identified as Other. Most of the participants (92%) were not international students. Forty-nine percent reported their economic background as middle class, 29% as upper middle class, 16.2% as working class, 4.1% as lower class, and 1% as upper class.
Childhood emotional abuse was operationalized through use of the Emotional Abusiveness subscale from the Exposure to Abusive and Supportive Environments Parenting Inventory (EASE-PI; Nicholas & Bieber, 1997). This subscale consists of 19 self-report items scored on a 5-point Likert-type scale. Participants are asked to rate the frequency with which, during their childhood, certain emotionally abusive communications occurred in interactions with parents. This subscale has been found to be adequately reliable and there is documented support for its concurrent and predictive validity (Hoglund & Nicholas, 1995; Nicholas & Bieber, 1997).
For this study, the subscale was modified to assess whether emotional abuse was experienced from either parental figure, rather than separately analyzing abusive behaviors of mother and of father (Nicholas & Bieber, 1997), an approach following that of many researchers, who do not distinguish between maternal and paternal behavior (Bernstein, Fink, Handelsman, & Foote, 1994; Glassman et al., 2007). Additionally, to be more inclusive the term guardian was added to all references to mother or father. The reliability of the modified subscale was high (alpha = .94).
Pessimistic explanatory style was assessed using the Attributional Style Questionnaire for General Use (ASQ-GU; Dykema, Bergbower, Doctora, & Peterson, 1996). The ASQ-GU lists 22 negative hypothetical events (e.g., "a friend is very angry with you"; "you have financial problems"). Participants are instructed to picture the event happening to them, write down a possible cause for it, and answer three questions concerning that cause. The three questions, scored on a 7-point Liked-type scale, assess the perceived internality, stability, and globality of causes. Responses to the items are then combined into a composite score of pessimistic explanatory style. Adequate reliability has been reported for this measure, and its discriminant and convergent validity have also been supported (Dykema et al., 2996; Peterson & Vaidya, 2001). In this study, its reliability was found to be good (alpha = .87).
Frequency of NSSI in the past year was assessed by use of the Functional Assessment of Self Mutilation (FASM; Lloyd-Richardson et al., 2007). Although the FASM was developed for use with adolescents, researchers have used it in studies with young adults (Selby, Connell, & Joiner, 2010). We used only the first section, which asks whether participants have engaged in any of 22 different self-injurious behaviors over the past 12 mouths and, if so, how often. The behaviors were cutting skin, hitting sell pulling hair out, tattooing skin, wound picking, burning skin, inserting objects under skin, biting sell picking body until bleeding, scraping skin, and "erasing skin," which refers to the use of an eraser to remove layers of skin. Lloyd-Richardson et al. (2007) grouped these behaviors into two subscales: (a) moderate/severe NSSI (e.g., burning, self-tattooing, scraping, and using an eraser); and (b) minor NSSI (e.g., hitting oneself, biting oneself, inserting objects under nails or skin, picking at a wound, picking at body areas to draw blood, and pulling one's hair). However, these subscales are often combined as a single measure of NSSI frequency (Nock & Prinstein, 2004; Selby et al., 2010; Yates et al., 2008). For our purposes, two subscale scores were combined as a single measure of NSSI frequency.
In alignment with concerns voiced by Lloyd-Richardson et al. (2007), the item "picked at a wound" was not included because this appears to be normative. Moreover, following modifications by Yates et al. (2008), the FASM was altered so that participants could score the frequency of each self-injurious behavior on a 5-point Likert-type scale (0 times; 2 time; 2-5 times; 6-10 times; 22 or more times). Adequate reliability has been reported for this measure (Yates et al., 2008), and its convergent validity has been supported (Guertin, Lloyd-Richardson, Spirito, Donaldson, & Boergers, 2002).
Another modification pertained to the question of whether self-injurious behaviors were suicidal. In the original form of the FASM, after participants indicated the frequency of their engagement in self-injurious behaviors, they were asked whether any of them was enacted as a suicide attempt. However, with an affirmative response it was still not possible to determine whether some or all of the self-injurious behaviors were suicidal in nature. Therefore, in this study, before the list of self-injurious behaviors, there were instructions that participants indicate the frequency with which they engaged in each behavior "without the intent to die"; and the reliability of the modified measure was found to be adequate (alpha = .64). Finally, a questionnaire was used to collect demographic data.
Three thousand e-mail addresses were randomly selected from the student directory. These students were sent an invitation to participate that included a link to a secure online survey administration tool. The invitation also directed participants to an online informed consent form and an explanation of the measures the study would use. As an incentive to participate, the students were informed that they could enter a random drawing to win one of three gift cards. Within three weeks, two reminder e-mails were sent to all deliverable addresses to increase response rate. After screening, the final response rate was 14%, which was comparable to previous online research on NSSI among college students (Aizenman & Jensen, 2007).
Before the analyses, variable distributions were inspected for skew, kurtosis, and outliers. If skew or kurtosis values indicate that a variable distribution departs meaningfully from a normal distribution, transformation of data to a more normal distribution is recommended (Tabachnick & Fidell, 2007). Distribution for the pessimistic explanatory style variable was found to approximate a normal distribution. Distributions for the emotional abuse and NSSI variables, however, were positively skewed and positively kurtosed. A skewed distribution is one in which the distribution of scores is asymmetrical. For example, a positively skewed distribution is characterized by a cluster of low scores without a similar number of high scores (Miles & Shevlin, 2001). Kurtosis refers to a distribution that is too flat or too peaked compared to the normal curve (Miles & Shevlin, 2001). A positively kurtosed distribution departs from the normal curve by being too sharply peaked, that is, having too few scores at the ends of the distribution.
A non-normal distribution may also be marked by a higher number of extreme values, outliers, than the number of extreme values in a normal distribution (Tabachnick & Fidell, 2007). To assess whether outliers have an extreme impact on statistics, the mean and 5% trimmed mean values of each variable are compared. To calculate the 5% trimmed mean score, the highest 5% of scores and the lowest 5% are removed from the distribution. Ideally, the mean and 5% trimmed mean scores should be similar. Where they are dissimilar, data transformations are recommended to reduce the impact of extreme outliers (Tabachnick & Fidell, 2007).
Unlike the distribution for pessimistic explanatory style, both the emotional abuse and NSSI distributions contained outliers that exerted undue influence on the analyses because the mean and 5% trimmed mean scores were dissimilar, so for these two variable distributions a logarithmic transformation was conducted. In terms of skew and kurtosis, both tire emotional abuse and NSSI data, once transformed, approximated a normal distribution, and the mean and 5% trimmed mean scores were similar, which suggested that outliers were no longer exerting extreme influence on the statistics.
The relationships between NSSI and several demographic variables were first examined. For chi-square tests, NSSI data were recoded into a categorical variable. As in previous research (Lloyd-Richardson et al., 2007), two categories were defined: (a) those who had engaged in at least one self-injurious behavior in the past year; and (b) those who had not. The results showed no significant relationship between NSSI and biological sex, race/ethnicity, sexual identity, or economic background (Table 1). A Pearson correlation showed a statistically significant negative relationship between age and NSSI (r = -.21, p <.001); age was therefore controlled in the principal analyses. The means and standard deviations for participant scores are reported with the nontransformed scores to allow for comparison with previous research (Table 2).
Our first hypothesis was that pessimistic explanatory style will be positively related to NSSI frequency in the past year. To test it, a bivariate correlation analysis was calculated between pessimistic explanatory style and NSSI, and a small but statistically significant positive relationship was found (r = .26, p < .001). Thus, the first hypothesis was supported.
To test the second hypothesis (that pessimistic explanatory style had a mediating relationship in the childhood emotional abuse--NSSI relation), four conditions had to be examined (Baron & Kenny, 1986). Three multiple regression analyses were conducted to address these. In each, age was entered as a covariate in the first step.
The first condition of mediation required that the mediator (pessimistic explanatory style) and predictor variable (childhood emotional abuse) be significantly associated (Baron & Kenny, 1986). Here, after controlling for age, childhood emotional abuse explained 3.9% of the variance in pessimistic explanatory style, [DELTA]F (1, 387) = 15.73, p < .001. The unstandardized coefficient for childhood emotional abuse was 11.73, and the standardized coefficient (beta value) was .20 (p < .001). The direction of the beta value indicated that increased emotional abuse was associated with increased pessimistic explanatory style.
The second condition required that the predictor variable (childhood emotional abuse) and outcome variable (NSSI) be significantly related (Baron & Kenny, 1986). After controlling for age, childhood emotional abuse explained an additional 7.6% of the variance in NSSI, [DELTA]F (1,387) = 33.7, p < .001. The unstandardized coefficient for childhood emotional abuse was .07, and the standardized coefficient (beta value) was .28 (p < .001). The direction of the beta value indicated that increased emotional abuse was associated with increased frequency of NSSI.
The third condition required that the mediator (pessimistic explanatory style) and outcome variable (NSSI) be significantly related after controlling for variance contributed by the predictor variable (childhood emotional abuse). After age was entered in the first step, NSSI was regressed simultaneously on pessimistic explanatory style and childhood emotional abuse. Together (after controlling for age) they explained an additional 11.7% of the variance in NSSI, [DELTA]F (2, 386) = 26.78, p < .001. The results showed that pessimistic explanatory style was significantly related to NSSI after controlling for childhood emotional abuse, beta = .21, p < .001. As expected, the direction of the beta value indicated that increased pessimistic explanatory style was related to increased frequency of NSSI.
The fourth condition of mediation required a significant reduction in the association between the predictor variable (childhood emotional abuse) and the outcome variable (NSSI), compared to the strength of the association shown in condition 2, after including the mediator variable (pessimistic explanatory style) in the regression (Baron & Kenny, 1986). Pessimistic explanatory style was found to function as a partial mediator. Although the childhood emotional abuse--NSSI relation remained significant in the results of the third regression analysis (B = .06, beta = .24, p <.001), this relation was smaller than it had been for the second condition of mediation (B = .07; beta = .28, p < .001). A Sobel Test (Sobel, 1982) was conducted to determine the significance of this mediated effect, which was found to be significant, z = 3.97, p < .001. Therefore, the second hypothesis was supported. Based on the findings, a medium effect size ([f.sup.2]) of .14 was calculated for the contribution of both emotional abuse and pessimistic explanatory style to explaining the frequency of NSSI.
Calculation of Pearson correlation analyses for variables included in mediation analyses produced several statistically significant relationships. Age was found in a significant relationship only with childhood emotional abuse (r = -.153, p < .01) and NSSI frequency (r = -.21, p < .001). Childhood emotional abuse was also significantly related to both pessimistic explanatory style (r = .20, p < .001) and frequency of NSSI (r = .31, p < .001); and pessimistic explanatory style was significantly related to frequency of NSSI (r = .26, p < .001).
In this study 42.3% of participants reported no engagement in NSSI in the past year, and 57.7% reported one or more incidents. Similar to Gratz and Chapman (2007) we identified a frequent self-injurer as a person who engaged in a single form of NSSI more than five times in the past year or a person who engaged in at least three or more forms of NSSI in the past year; 39.5% of study participants qualified as frequent self-injurers.
In this study, the NSSI methods used most often were biting self (37.1% of sample), scraping skin (26%), and cutting or carving the skin (8.5%). Biting self was operationalized on the FASM (Lloyd-Richardson et al., 2007) as "bit yourself (e.g., your month or lip)." The least popular NSSI methods were erasing skin (2.3%) and self-tattooing (3.3%).
The purpose of this study was to examine the frequency of NSSI among young adults and the mediating role of pessimistic explanatory style in the childhood emotional abuse--NSSI relation. Within this nonclinical sample, 58% of participants reported engaging in NSSI at least once in the past year, and over one-third could be classified as frequent self-injurers. The most common NSSI methods used were biting self and scraping skin.
These findings corresponded to the results of previous studies that used the FASM to measure NSSI. Lloyd-Richardson et al. (2007), for example, found that nearly half (46%) of a sample of nonclinical adolescents had performed NSSI in the past year; by contrast, Nock and Prinstein (2004) showed that 82.4% of clinical inpatients did so. Thus, as expected, in this study the prevalence of NSSI was lower than in clinical settings and similar to the rate found in other nonclinical settings.
The results also suggested that NSSI was not significantly related to such participant demographic characteristics as race, gender, SES, and sexual orientation, which conformed fairly well to previous studies (Gratz & Chapman, 2007), although the research in this area is limited. Age was found to be negatively related to NSSI. Even within the limited age range for participants (18-25 years old), younger participants were more likely to report more NSSI than older participants.
In the principal analyses for this study, pessimistic explanatory style was found to relate significantly with NSSI: individuals with a more pessimistic explanatory style tended to report more NSSI. This finding adds support to the theory that control-related cognitions are implicated in the development of NSSI (Suyemoto & MacDonald, 1995). It was also consonant with research on the association of NSSI with related cognitive variables (e.g., reduced sense of control; Aizenman & Jensen, 2007). Taken together, these studies suggest that cognitive characteristics--control-related cognitions in particular--may be important predictors of NSSI.
Another significant finding concerns the mediating role of pessimistic explanatory style in the childhood emotional abuse-NSSI relation. Previous research had not considered pessimistic explanatory style or its role in developmental pathways leading to NSSI, but these results suggest that it partially explains the relationship between childhood emotional abuse and NSSI. This mediated effect was found even after controlling for variance in NSSI attributable to age. Since this article was accepted for publication, two studies (Guerry & Prinstein, 2010; Hankin & Abela, 2011) have reached similar findings on the relationship between NSSI and pessimistic explanatory style.
Support for the proposed mediation model was also consistent with the findings of Glassman et al. (2007) that self-criticism functioned as a partial mediator in the childhood emotional abuse--NSSI relation. Self-criticism is conceptually related to the internality dimension of pessimistic explanatory style. The latter construct, however, addresses additional domains (stability and globality) pertaining to the causes of negative events (Abramson et al., 1978). Building upon the research of Glassman et al. (2007), this study identified another cognitive mechanism, pessimistic explanatory style, that mediates the relationship between childhood emotional abuse and NSSI.
This study had several strengths, including its focus on pessimistic explanatory style in conjunction with NSSI--few studies have examined the association of NSSI with cognitive variables. The study was also strengthened by its use of random sampling and the application of theory in designing the mediation model.
However, the study was limited in a number of ways. Because all data were collected at a single point in time, it was not possible to establish the temporal precedence of the predictor variable, childhood emotional abuse, to the other variables studied. Thus, the investigation was not positioned to assess causal relationships between variables. Further, the external validity of the study was threatened by the relative demographic homogeneity of the sample. For example, since the vast majority of participants in this sample identified themselves as heterosexual (95%), it is uncertain whether these findings generalize to populations that have high percentages of gay/lesbian and bisexual individuals. Moreover, the response rate for this study was relatively low (14%), which introduces questions about whether the data were representative of the population of young adults at the sampling site.
Another limitation relates to the grouping together of both moderate/severe forms of NSSI and minor forms on the outcome measure of NSSI frequency. The aim of this study was to calculate a frequency score that covered a broad range of NSSI behaviors. This approach corresponds to that of other researchers who have used the FASM (Nock & Prinstein, 2004; Selby et al., 2010; Yates et al., 2008). Moreover, proposed revisions to the Diagnostic and Statistical Manual off Mental Disorders (4th ed., text rev.; DSM-IV-TR; American Psychiatric Association [APA], 2000) include a new diagnosis for NSSI that also groups together methods that Lloyd-Richardson et al. (2007) categorized as minor and moderate/severe (APA, 2010). However, researchers have identified differences between self-injurers who engage in minor forms of NSSI and those who engage in moderate/severe forms (Lloyd-Richardson et al., 2007). It is possible, therefore, that differences related to particular methods (e.g., moderate/severe forms of NSSI) introduced a confounding variable in the analyses.
Despite such limitations, the findings have valuable implications for research, considering how few studies there are on the etiology of NSSI (Yates et al., 2008). The study constituted a preliminary step in identifying cognitive pathways that partially explain the relationship between childhood emotional abuse and NSSI. The findings lay the groundwork for research into whether childhood emotional abuse does indeed contribute to pessimistic explanatory style over time, which would increase the likelihood of later NSSI.
Furthermore, the results suggested that pessimistic explanatory style served as a partial, rather than full, mediator of the relationship between childhood emotional abuse and NSSI. Studies should be directed to testing additional cognitive variables that might further explain the relationship. Social self-efficacy, for example, has been inversely related to NSSI and refers to an individual's beliefs about his or her ability to effectively respond to challenging social scenarios (Nock & Mendes, 2008). Assuming that social self-efficacy is also associated with childhood maltreament, this cognitive characteristic may play a mediating role similar to that of pessimistic explanatory style.
Other implications pertain to issues of instrumentation. Frequency of NSSI in the past year was assessed, rather than frequency over the course of participants' lifetimes ("lifetime NSSI"). Although measures have been devised to assess lifetime NSSI (Gratz, 2001) and researchers continue to explore this outcome variable (MacLaren & Best, 2010), the practical relevance of studies on lifetime NSSI is arguably less apparent. Items on a measure of lifetime NSSI may be endorsed by both individuals who have self-injured in the past year and those who self-injured in the distant past (e.g., childhood or adolescence) but have since stopped. It is unclear whether the latter group has any need for psychotherapeutic treatment. Theoretically, this type of person may not exhibit any current maladaptive behaviors. Note that proposed revisions to the DSM-IV-TR (APA, 2000) define NSSI as occurring within the past year (APA, 2010). Thus, investigations of recent engagement in NSSI may have more practical meaning for the counseling field.
The clinical implications of this study clearly pertain to the findings on pessimistic explanatory style. Since this was found to partially mediate the childhood emotional abuse--NSSI relation, counselors may want to consider interventions that target the explanatory style of self-injurers. Counselors are encouraged to examine whether self-injuring clients tend to attribute negative events to internal, stable, and global causes. If they manifest a pessimistic explanatory style, it is advisable for counselors to help them re-form these cognitions.
To do so, two intervention strategies may be particularly useful: cognitive therapy (Beck 1995; Beck, Rush, Shaw, & Emery, 1979) and narrative therapy (White & Epston, 1990). In general, cognitive therapy focuses on adjusting cognitions that adversely influence emotional experiences and the behavioral options clients perceive (Beck, 1995; Beck et al., 1979). In a controlled research design, Fresco, Moore, Walk, and Craighead (2009) documented the effectiveness of one cognitive intervention, Self-Administered Optimism Training (SOT), in changing the pessimistic explanatory style of college students (N = 112). SOT addresses aspects of pessimistic explanatory style by quickly (10 minutes) teaching clients about explanatory styles and how to describe a cause along the dimensions of internality, stability, and globality. As part of the training, clients are asked to journal about events for one month, listing five positive events and five negative events each week; determine their perception of the cause of each; and decide on the degree to which the causes are internal, stable, and global. Later, they are asked to generate alternate causal explanations for each event and assess the degree to which these causes are internal, stable, and global. As noted by Fresco et al. (2009), SOT focuses solely on the creation of new alternatives for thinking about a situation rather than on evaluation of the validity of a particular cognition. The results indicated that, after one month, participants who took part in SOT (n = 43) reported a less pessimistic explanatory style than the control group (n = 55). Counselors, therefore, are encouraged to consider elements of SOT (in-session training on pessimistic explanatory style and homework assignments of journaling about causal explanations) as an adjunct to counseling with self-injuring, pessimistic clients.
Narrative therapy (White & Epston, 1990) may also be useful for altering pessimistic explanatory style. Although to our knowledge no studies have examined the effectiveness of narrative therapy in changing pessimistic explanatory style, the central emphases and strategies of narrative therapy seem ideally suited to achieving this treatment goal. Narrative therapy is guided by the notion that the stories that clients tell about their lives are fundamentally inadequate and imprecise because they fail to depict the entirety of client experiences (White & Epston, 1990). Narrative therapists thus work to help clients rewrite their stories to integrate more complexity through two primary techniques: externalizing the problem and searching for unique outcomes (White & Epston, 1990).
Consider, for example, the client who recently failed an exam and is consumed by self-reproach. In manifesting a pessimistic explanatory style, the client may initially attribute this negative event to laziness, which indicates the internality dimension of the style. To externalize the problem, narrative therapists initiate a new pattern of discourse in which the problem is described as though it were external to the person (White & Epston, 1990). Externalization begins as the client ascribes a name to the problem, such as "Laziness" or more playfully the "Laziness Spell." The very process of naming often helps a client to conceptualize the difficulty as a force external, rather than internal, to the self. Externalization is then reinforced as clients are asked to write letters to "Laziness," to discuss ways in which "Laziness" influences their lives, and to plan ways to work against "Laziness" (White & Epston, 1990). By employing new language, clients gradually separate the self from the problem (White & Epston, 1990). Thus, clients may begin to counteract the internality dimension of a pessimistic explanatory style and assume more complex views of the causes of difficult situations.
To address the other facets of pessimistic explanatory style (the stability and globality of causes of events), the counselor may employ the narrative therapy technique of searching for unique outcomes. This strategy assumes that there are moments in the client's life when the problem is absent (White & Epson, 1990). By uncovering the unique outcomes of (a) times when the cause of this negative event was not present, and (b) areas of the client's life that are not affected by the cause of this event, it is possible to target and challenge a pessimistic explanatory style. For example, in addressing the perceived stability of "Laziness," clients can be encouraged to describe times when they exerted a great deal of effort on a project. Also, in confronting the characterization of "Laziness" as a global cause, clients may discuss other areas where they rarely feel unmotivated, such as athletic competitions or spiritual activities. Over the course of counseling, clients may evidence a shift to an optimistic explanatory style, which attributes negative events to external, unstable, and specific causes. For example, a client may begin to highlight external, systemic variables (e.g., lack of resources in a workplace or academic environment) that may contribute to negative events. More research will be needed, however, to determine whether such cognitive shifts actually reduce NSSI among young adults.
Abramson, L., Seligman, M., & Teasdale, J. (1978). Learned helplessness in humans: Critique and reformulation. Journal of Abnormal Psychology, 87, 49-74. doi: 10.1037/0021-843X.87.1.49
Aizenman, M., & Jensen, M. (2007). Speaking through the body: The incidence of self-injury, piercing, and tattooing among college students. Journal of College Counseling, 10, 27-43.
Allen, C. (1995). Helping with deliberate self-harm. Journal of Mental Health, 4, 243-250. doi: 10.1080/09638239550037523
American Psychiatric Association (2000). Diagnostic and statistical manual of mental disorders (4th ed., text revision). Washington, DC: Author.
American Psychiatric Association (2010). DSM-5 development: Non-suicidal self injury. Retrieved from http://www.dsm5.org/ProposedRevisions/Pages/proposedrevision.aspx?rid=443
Baron, R., & Kenny, D. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173-1182. doi:10.1037/0022-3522.214.171.1243 Beck, J. S. (1995). Cognitive therapy: Basics and beyond. New York, NY: Guilford Press.
Beck, A. T., Rush, A., J., Shaw, B. F., & Emery, G. (1979). Cognitive therapy of depression. New York, NY: Guilford Press.
Berustein, D., Fink, L., Handelsman, L., & Foote, J. (1994). Initial reliability and validity of a new retrospective measure of child abuse and neglect. American Journal of Psychiatry, 151, 1132-1136.
Claes, L., Houben, A., Vandereycken, W., Bijttebier, P., & Muehlenkamp, J. (2010). Brief report: The association between non-suicidal self-injury, self-concept and acquaintance with self-injurious peers in a sample of adolescents. Journal of Adolescence, 33, 775-778. doi:10.1016/j.adolescence.2009.10.012
Claes, L., Klonsky, E. D., Muehlenkamp, J., Kuppens, P., & Vandereycken, W. (2010). The affect-regulation function of nonsuicidal self-injury in eating-disordered patients: Which affect states are regulated? Comprehensive Psychiatry, 51, 386-392. doi:10.1016/j.comppsych.2009.09.001
Dykema, J., Bergbower, K., Doctora, J. D., & Peterson, C. (1996). An attributional style questionnaire for general use. Journal of Psychoeducational Assessment, 14, 100-108. doi:10.1177/ 073428299601400201
Favazza, A. (1998). The coming of age of self-mutilation. Journal of Nervous and Mental Disease, 186, 259-268. doi: 10.1097/00005053-199805000-00001
Fresco, D. M., Moore, M. T., Walk, L., & Craighead, L. W. (2009). Self-administered optimism training: Mechanisms of change in a minimally supervised psychoeducational intervention. Journal of Cognitive Psychotherapy: An International Quarterly, 23, 350-367. doi:10.1891/0889-83126.96.36.1990
Gibb, B., Alloy, L., Abramson, L., Rose, D., Whitehouse, W., Donovan, P., ... Tierney, S. (2001). History of childhood maltreatment, negative cognitive styles, and episodes of depression in adulthood. Cognitive Therapy and Research, 25, 425-46. doi:10.1023/A:1005586519986
Glassman, L., Weierich, M., Hooley, J., Deliberto, T., & Nock, M. (2007). Child maltreatment, non-suicidal self-injury, and the mediating role of self-criticism. Behaviour Research and Therapy, 45, 2483-2490. doi:10.1016/j.brat.2007.04.002
Gratz, K (2001). Measurement of deliberate self-harm: Preliminary data on the Deliberate Self-Harm Inventory. Journal of Psychopathology and Behavioral Assessment, 23, 253-263. doi:10.1023/A:1012779403943
Gratz, K., & Chapman, A. (2007). The role of emotional responding and childhood maltreatment in the development and maintenance of deliberate self-harm among male undergraduates. Psychology of Men and Masculinity, 8, 1-14. doi:10.1037/1524-9188.8.131.52
Guerry, J. D., & Prinstein, M. J. (2010). Longitudinal prediction of adolescent nonsuicidal self-injury: Examination of a cognitive vulnerability-stress model. Journal of Clinical Child and Adolescent Psychology, 39, 77 89. doi: 10.1080/15374410903401195
Guertin, T., Lloyd-Richardson, E., Spirito, A., Donaldson, D., & Boergers, J. (2001). Self-mutilative behavior in adolescents who attempt suicide by overdose. Journal of the American Academy of Child and Adolescent Psychiatry, 40, 1062-1069. doi:10.1097/00004583-200109000-00015
Hankin, B. (2005). Childhood maltreatment and psychopathology: Prospective tests of attachment, cognitive vulnerability, and stress as mediating processes. Cognitive Therapy and Research, 29, 645-671. doi:10.1007/s10608-005-9631-z
Hankin, B. L., & Abela, J. Z. (2011). Nonsuicidal self-injury in adolescence: Prospective rates and risk factors in a 2 1/2 year longitudinal study. Psychiatry Research, 186, 65-70. doi:10.1016/j.psychres.2010.07.056
Harris, J. (2000). Self-harm: Cutting the bad out of me. Qualitative Health Research, 10, 164-173. doi: 10.1177/104973200129118345
Hoglund, C. L., & Nicholas, K. B. (1995). Shame, guilt, and anger in college students exposed to abusive family environments. Journal of Family Violence, 10, 141-157. doi:10.1007/BF02110597
Keashly, L., & Harvey, S. (2005). Emotional abuse in the workplace. Counterproductive work behavior: Investigations of actors and targets (pp. 201-235). Washington, DC: American Psychological Association. doi: 10.1037/10893-009
Klonsky, E. (2007). The functions of deliberate self-injury: A review of the evidence. Clinical Psychology Review, 27, 226-239. doi:10.1016/j.cpr.2006.08.002
Lloyd-Richardson, E., Perrine, N., Dierker, L., & Kelley, M. (2007). Characteristics and functions of non-suicidal self-injury in a community sample of adolescents. Psychological Medicine, 37, 1183-1192. doi: 10.1017/S003329170700027X
MacLaren, V. V., & Best, L. A. (2010). Nonsuicidal self-injury, potentially addictive behaviors, and the Five Factor Model in undergraduates. Personality and Individual Differences,49, 521-525. doi: 10.1016/j.paid.2010.05.019
McGee, R., & Wolfe, D. (1991). Psychological maltreatment: Toward an operational definition. Development and Psychopathology, 3, 3-18. doi: 10.1017/S0954579400005034
Miles, J., & Shevlin, M. (2001). Applying regression and correlation. London, UK: Sage.
Moyer, M., & Nelson, K. (2007). Investigating and understanding self-mutilation: The student voice. Professional School Counseling, 11, 42-48.
Nicholas, K., & Bieber, S. (1997). Assessment of perceived parenting behaviors: The Exposure to Abusive and Supportive Environments Parenting Inventory (EASE-PI). Journal of Family Violence, 12, 275-291. doi:10.1023/A:1022848820975
Nock, M., & Mendes, W. (2008). Physiological arousal, distress tolerance, and social problem-solving deficits among adolescent self-injurers. Journal of Consulting and Clinical Psychology, 76, 28-38. doi: 10.1037/0022-006X.76.1.28
Nock, M., & Prinstein, M. (2004). A functional approach to the assessment of self-mutilative behavior. Journal of Consulting and Clinical Psychology, 72, 885-890. doi:10.1037/0022006X.72.5.885
Peterson, C., & Park, N. (2007). Explanatory style and emotion regulation. In J. J. Cross (Ed.), Handbook of emotion regulation (pp. 159-179). New York, NY: Cuilford Press.
Peterson, C., & Vaidya, R. S. (2001). Explanatory style, expectations, and depressive symptoms. Personality and Individual Differences, 71, 1217-1223. doi:10.1016/S0191-8869(00)00221-X
Rose, D., & Abramson, L. (1992). Developmental predictors of depressive cognitive style: Research and theory. In D. Cicchetti, & S. Toth (Eds.), Rochester symposium of developmental psychopathology, vol. IV (pp. 323-349). Rochester, NY: University of Rochester Press.
Selby, E. A., Connell, L. D., & Joiner, T. E. (2010). The pernicious blend of rumination and fearlessness in non-suicidal self-injury. Cognitive Therapy Research, 34, 421-428. doi:10.1007/ s10608-009-9260-z
Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in structural equation models. In S. Leinhardt (Ed.), Sociological methodology 1982 (pp. 290-312). Washington, DC: American Sociological Association.
Suyemoto, K., & MacDonald, M. (1995). Self-cutting in female adolescents. Psychotherapy: Theory, Research, Practice, Training, 32, 162-171. doi:10.1037/0033-3184.108.40.206
Tabachnick, B. G., & Fidell, L.S. (2007). Using multivariate statistics (5th ed.). New York, NY: Pearson.
Walsh, B. (2007). Clinical assessment of self-injury: A practical guide. Journal of Clinical Psychology: In Session, 63, 1057-1068. doi: 10.1002/jclp.20413
White, M., & Epston, D. (1990). Narrative means to therapeutic ends. New York, NY: W.W. Norton & Company.
White, V. E., McCormick, L. J., & Kelly, B. L. (2003). Counseling clients who self-injure: Ethical considerations. Journal of College Counseling, 47, 220-229.
Yates, T., Tracy, A., & Luthar, S. (2008). Nonsuicidal self-injury among "privileged" youths: Longitudinal and cross-sectional approaches to developmental process. Journal of Consulting and Clinical Psychology, 76, 52-62. doi:10.1037/0022-006X.76.152
Trevor J. Buser is affiliated with Rider University; Harold Hackney is professor emeritus, Syracuse University. Correspondence concerning this article should be directed to Dr. Trevor Buser, Department of Graduate Education, Leadership, & Counseling, Rider University, Memorial Hall 202-M, 2083 Lawrenceville Rd., Lawrenceville, NJ 08648. E-mail: firstname.lastname@example.org.
Table 1. Chi Square Tests for Independence: Demographics and NSSI Behaviors Asymptomatic Valid Significance Demographic Cases Measure Value (2-sided) Biological sex 389 Continuity Correction .73 .39 Race/ethnicity 388 Pearson Chi-Square 6.27 .28 Sexual identity 387 Pearson Chi-Square .494 .48 Economic class 388 Pearson Chi-Square 5.44 .26 Table 2. Means, Standard Deviations, and Ranges of Childhood Emotional Abuse, Pessimistic Explanatory Style, NSSI, and Age (n = 390) Variable M SD Range Emotional abuse 9.7 10.8 66 Pessimistic style 4.5 0.8 6 NSSI 13.0 3.8 20 Age 20.3 1.9 7