The Relationship Between Stockholm Syndrome and Post-Traumatic Stress Disorder in Battered Women
2009, Vol. 1 No. 11 | pg. 1/1
Every year, 10-50% of women suffer intimate partner violence (Bargai, Ben-Shakhar, & Shalev, 2007). It is important to understand what conditions affect these battered women and how any resultant conditions interact with each other in order to help abused women work through the aftermath of their trauma. Two of the more common resultant conditions of abuse are Posttraumatic Stress Disorder (PTSD) and Stockholm Syndrome (Graham, Rawlings, & Rimini, 1988; Hughes & Jones, 2000).
The Diagnostic and Statistics Manual IV-TR (American Psychiatric Association, 2000) defines PTSD as the development of several characteristics following a traumatic experience where intense fear, helplessness, or horror is experienced. The symptoms include persistent reexperiencing of the event, persistent avoidance of stimuli associated with the event, a numbing of general responsiveness, and persistent increased arousal for more than one month (APA, 2000). Abuse by an intimate partner has been repeatedly shown to increase the abused person’s likelihood of exhibiting PTSD. (Hughes and Jones, 2000; Walker, 1991 both of which are explained later).In 1995, the National Research Council fulfilled a congressional request to investigate causes, consequences, and prevention of violence against women (Crowell & Burgess, 1996). The panel determined that PTSD is the most common disorder among survivors of domestic violence. The prevalence and treatment of PTSD was further explored by Hughes and Jones (2000) in a survey of California public programs for battered women. The survey included questions on the availability of support for survivors in the public programs’ facilities as well as the knowledge and information base the support programs utilized. The researchers found that although the programs were aware that the survivors were likely to suffer from PTSD, few programs actually provided support for PTSD. Only 50% of the domestic violence crisis centers provided support groups while less than 40% offered individual and group treatment for PTSD. This statistic indicates that although the existence of PTSD is acknowledged, in most cases support for symptoms of PTSD is nonexistent. The programs also have no basis for support for conditions other than PTSD. There is also the possibility that PTSD symptoms may be moderated by co-morbid conditions such as depression or the coping strategies utilized by survivors, one of which may be Stockholm Syndrome (Hughes & Jones, 2000).
Astin, Lawrence and Foy (1993) also examined the effect of PTSD on battered women. The study emphasized the risk and resiliency factors for PTSD with a self-report survey to assess violence, PTSD symptomatology, social support, intercurrent life events, religiosity, and demographics. The researchers found that PTSD was found in survivors of abuse and PTSD was also highly correlated with the severity of the abuse. Women who declared higher levels of social support and religiosity than the mean experienced fewer PTSD symptoms than those declaring lower levels of social support. The study supports the theory that PTSD is influenced by the situation in which the women find themselves. The exploration of PTSD and Stockholm Syndrome and their interactions may lead to valuable information about the relationships other conditions may have on PTSD which in turn benefits the assistance of survivors.
Stockholm Syndrome is the term for a condition that evolves between an aggressor and the victims in situations such as hostage negotiations, kidnapping, and abuse (Auerbach, Kiesler, Strentz, Schmidt, & Serio, 1994; Graham et al., 1988). The main symptom of Stockholm Syndrome is the development of positive feelings on the part of the hostages for their captors or abusers. Other symptoms may include reciprocal feelings from the aggressors or negative feelings on the part of the victim towards the authorities or outside influences (Auerbach et al., 1994). Stockholm Syndrome is a named after a bank heist in Stockholm, Sweden in which the hostages began to identify with their hostage-takers. During this event, one hostage became engaged to one of the aggressors (Kuleshnyk, 1984). Stockholm Syndrome results from a rather specific set of circumstances, namely the power imbalances contained in hostage-taking, kidnapping, and abusive relationships. Therefore, it is difficult to find a large number of people who experience Stockholm Syndrome to conduct studies with any sort of power. This makes it hard to determine trends in the development and effects of the condition.
However, Auerback, Kiesler, Strentz, Schmidt, and Serio (1994) empirically recreated a hostage situation which resulted in the formation of Stockholm Syndrome symptoms. Six FBI agents acted as terrorists and held several airline employees for four days. Stockholm Syndrome was observed in several of the airline employees. The researchers were also able to determine that Stockholm Syndrome was caused by the dynamic interplay of domination (the aggressors’ control of their hostages) and affiliation (the victims’ identification with the aggressors). The researchers concluded Stockholm Syndrome was a combination of defensive self-delusion and an effort to “get along” with their captors which is similar to some forms of coping such as wishful thinking or avoidance coping.
One problem with the Auerbach, et al. study (1994) is the morality of inducing Stockholm Syndrome in the participants. However, the participants all gave informed consent for a study to investigate the effectiveness of the training the airline workers had received about hostage situations. Stringent measures were also in place to ensure that participants were never physically harmed. The participants were also provided with extensive debriefing after the experiment. However, the Auerbach et al. study (1994) has been invaluable in observing the actual genesis of Stockholm Syndrome and has provided excellent data on what factors in hostage and kidnapping situations trigger the affiliation of the victims with the aggressors.
Kuleshnyk (1984) stated that any person who found themselves in a hostage situation was best served by fostering Stockholm Syndrome. Doing so allowed the hostages to better identify with the attackers and survive the encounter by working with the aggressors rather than against them. From this, it can be argued that Stockholm Syndrome is another method of coping with the stress and danger of the hostage situation while simultaneously allowing the victim to avoid the inherently dangerous option of resisting their captors by identifying with them. Stockholm Syndrome is similar to some forms of coping in that the participants do not directly address the problem but find a way to cope with the situation by identifying with the aggressor. Coping mechanisms such as these can have a large impact on PTSD.
There are several studies which have explored various aspects of PTSD and its relationship to coping mechanisms in women who had experienced sexual assault or domestic abuse. Walker (1991) surveyed over 200 victims of sexual and nonsexual assault (physical attacks or psychological abuse) on symptoms of PTSD and the effect of coping strategies. To measure the amount of PTSD symptoms and coping strategies, the participants completed self-report measures of PTSD and several coping mechanism measures. Walker found that the PTSD symptoms declined over time. However, the coping strategies, which consisted of Mobilizing Support (informing friends and family of the assault and requesting their support) and Positive Distancing (believing the trauma is past and it is time to move on), had no relationship to severity of symptoms.
However, Wishful Thinking (wishing the trauma had not happened, trying to determine their own responsibility for the trauma, or pretending the trauma had not occurred) was positively associated with severity of PTSD symptoms. This indicates that the fantasies of the Wishful Thinking coping strategy were related to the increased severity of PTSD symptoms. The Wishful Thinking form of coping may, in fact, be related to Stockholm Syndrome because Stockholm Syndrome includes a significant amount of self-delusion on the part of abused individuals. This indicates that Stockholm Syndrome may be a form of “wishful thinking” coping. However, Wishful Thinking is not the only form of coping which may relate to Stockholm Syndrome.
Lawler, Ouimette and Dahlstedt’s (2005) also explored coping strategies and PTSD among university students. The researchers used self-report surveys to determine PTSD symptomatology and avoidant vs. approach coping strategies. In avoidance coping, the participant is more likely to ignore the trauma and pretend the trauma did not happen whereas in approach coping the survivors are more likely to try and resolve their problems. The researchers found that students who utilized the health clinic on campus and who has high level of PTSD symptomatology also had higher levels of avoidance coping compared to approach coping. Avoidance coping is also related to Stockholm Syndrome as survivors who experience Stockholm Syndrome frequently delude themselves insomuch as they believe they experience less violence or refuse to acknowledge the abuse. Thus, Stockholm Syndrome may act as avoidant or wishful thinking coping in relationship to PTSD.
Graham, Rawlings, and Rimini (1998) advanced the exploration of Stockholm Syndrome by examining its relationship to women in abusive relationships. In a literature review, the authors compared the situation of hostages and kidnap victims to that of women in abusive relationships and found that the situations had many features in common, including violence, domination, seclusion, constant proximity to the “captor” and intimidation. Many of the women who are beaten and estranged from family and friends continued to claim to love and care about their abusers. It was this phenomenon that led the authors to the conclusion that many women in intimate violence situations may suffer from Stockholm Syndrome and that more research should be done on Stockholm Syndrome and battered women.
Graham et al. (1995) created a scale to identify Stockholm Syndrome in battered women. Six hundred undergraduate women participated in a self-report survey which included demographic and relationship information, a 127 item Stockholm Syndrome scale, the Borderline Personality Disorder scale, and a form of the Marlowe-Crowne Social Desirability scale. The last two scales were given in an effort to determine whether the traits on the Stockholm Syndrome questionnaire were more appropriately ascribed to Borderline Personality Disorder or the participants score on Social Desirability. After a rigorous course of tests to determine which questions most closely followed and represented the various facets of Stockholm Syndrome, researchers finalized a 49 question survey about three central factors of Stockholm Syndrome to identify Stockholm Syndrome in women who have suffered abusive relationships.
The three factors are Core Stockholm Syndrome (affiliation with the abuser), Psychological Damage (lasting effects from the abuse), and Love-Dependence (how the women feel they love or are dependent on their abusers). The three subsets were represented in questions such as “It is really hard for me to question whether my relationship with my partner is good for me (Core),” “Without my partner I would not know who I am (Love),” and “I feel down and blue (Damage).” These aspects once more support the hypothesis that Stockholm Syndrome shares several facets with avoidant coping such as being unable to look at the relationship and determine its value. Instead the battered woman chose to ignore the abuse, and as is suggested by the earlier studies, avoidant coping can have a strong relationship with PTSD symptomatology.
Thus, it is necessary to determine whether Stockholm Syndrome has any relation to PTSD because Stockholm Syndrome shares aspects with avoidance coping and is prevalent in populations of battered women. If PTSD and Stockholm Syndrome are correlated in populations of battered women, there may be new ways to approach therapy with the survivors. Favaro, Degortes, Colombo and Santonastaso (2000) is the only article to consider the relationship between Stockholm Syndrome and PTSD. The researchers studied PTSD in Italian kidnap victims. The participants were interviewed using a series of self-report surveys and questions about PTSD, including one question about Stockholm Syndrome. They determined that there was no relationship between Stockholm Syndrome and PTSD.
However, there are several problems with this conclusion. First, international kidnapping situations differ greatly from that of women in abusive relationships. Battered women were in constant contact with their significant others and were in a relationship which society generally considers acceptable. However, the kidnap victims were often isolated from their attackers and were in a situation which was not condoned by society. Thus, the correlations between Stockholm Syndrome and PTSD in a population of battered women may be considerably different than the correlation between Stockholm Syndrome and PTSD that was observed in Italian kidnap victims. Also, the scale used to identify Stockholm Syndrome was a single question about identifying with the kidnappers. If the participants described a positive bond with their captors, the researchers determined that the participant qualified for Stockholm Syndrome.
Not only does the method of identification lack statistical validity, the limitations of the interview question do not allow for the possibility of positive feelings towards the captors for reasons such as sympathy with the political ideology. The lack of validity combined with the difference between Italian kidnap victims and battered women allows for the possibility of a relationship between PTSD and Stockholm Syndrome in battered women. Stockholm Syndrome may be used as a coping mechanism in both situations but Favaro et al. (2000) fails to recognize this because of the poor validity of their identification measures of Stockholm Syndrome.
The current research explores the relationship between PTSD and Stockholm Syndrome in populations of battered women. It is hypothesized that PTSD and Stockholm Syndrome will be positively correlated because of Stockholm Syndrome’s similarity to avoidant methods of coping. Stockholm Syndrome contains elements of wishful thinking as well as a mental avoidance of the situation. As shown with avoidance coping, as the amount of avoidance increases, so does the amount of PTSD symptomology (Lawler, Ouimette, & Dahlstedt, 2005). It is also hypothesized that the amount of abuse suffered will be positively correlated with PTSD while there will be no relationship of abuse to Stockholm Syndrome. This is hypothesized because the kind of abuse does not determine the formation of Stockholm Syndrome (Auerbach, Kiesler, Strentz, Schmidt, & Serio, 1994).
Fifty women were recruited from women seeking assistance at two domestic violence shelters: one in a moderately sized community in Oregon and one in a slightly larger community in Pennsylvania. Of the fifty participants, two women failed to complete the survey as directed and were excluded. There were no significant differences between the women from each crisis center on any of the scales or demographic features except for race. One crisis center had significantly more Caucasians and Multiracial participants while the other crisis center had more African American participants, p < .01. However, PTSD, SSS, WAS, and demographic features did not differ significantly by race (See Table 1). The scores of the PTSD Checklist, SSS, and WAS were close to normally distributed across the sample (see Table 2).
The average age of the women was 36.69 years with a standard deviation of 9.94 years. The women had spent an average of 5.9 years in the abusive relationship (labeled Exposure in Table 1; SD = 5.2 years) and had been at their respective crisis center an average of 8.2 days (SD = 8.5 days). The women also had an average of 2.2 children (SD = 1.8) and had a median income bracket of $0 - $15,000 annual income. Demographic information about the participants can be found in Table 3. Women seeking the assistance of the crisis centers during the summer and fall were given the opportunity to participate in the survey on a voluntary basis. All participants were given ten dollars in compensation, regardless of the level of completion on the survey and the shelters were given a one hundred and fifty dollar donation for their assistance in the survey.
Stockholm Syndrome Scale (SSS). Graham, et Al. (1995) created a 49 item scale to identify Stockholm Syndrome in dating women. The SSS includes items such as “Other people see only my partner’s negative side; they don’t see all the small kindnesses he does for me that make me love him.” These items are to determine Core aspects of Stockholm Syndrome as well as to measure Damage (“I feel down and blue”) and Love (“Without my partner, I would not know who I am”). It is scaled on a five point scale from the letter A to E.
PTSD Checklist – Civilian version. Sypmtoms of PTSD were assessed with the PTSD Checklist – Civilian version (PCL-C; Walker, Newman, Ciechanowski, & Katon, 2002) which measures distress associated with individual PTSD symptoms experienced in the past month. The PTSD Checklist consists of 17 items rated on a five point scale from 1 (not at all bothered by this symptom) to 5 (extremely bothered).
Woman Abuse Scale (WAS). Intensity of abuse was measured with a 26 item measure which was in itself a modified version of the Conflict Tactics Scale (Straus, 1979). The researcher created the Woman Abuse Scale for Saunders (1995). Saunders (1995) was designed to find way to assist in the prediction of recidivism in abuse cases when determining child custody and other similar situations. The measure asked the participant not only to identify the kinds of abuse suffered (separated into psychological, physical, and life-threatening) but also how many times a month the abuse occurred (see Appendix A for a copy of the Woman Abuse Scale).
Demographic Survey. The participants also completed a demographic survey which included questions on the length of the abusive relationship, length of time away from abuser, length of time in the shelter’s services, age, number and age of any children, race, education level, and income level (see Appendix B for a copy of the survey).
After the women entered the shelters’ programs, the battered women were offered the chance to participate in the survey. The women were then given the consent form, which they signed and returned prior to completing the survey. The participants were given as much time as needed to complete the four sections of the survey and were able to discontinue participation if distressed. At completion of the survey the participants were then given an envelope containing a copy of the consent form to keep, a debriefing sheet with facts about PTSD and Stockholm Syndrome as well as $10 for their participation. The staff was available throughout the entire process to answer any questions the participants may have had.
The correlation between PTSD and Stockholm Syndrome in the 48 women yielded a Pearson r of -.101 (p > .05). The internal subsets of the SSS (Love, Core, and Damage) correlated with the total score of the SSS with a Pearson r of .841, .965, and .915 respectively, p < .01. The internal subsets of the WAS (Psychological, Physical, and Life-Threatening) correlated with the total score of the WAS yielded a Pearson r of .833, .926, and .746 respectively, p < .01. The correlations between the WAS, SSS, PTSD, Exposure, Service, age, number of children, income, race and education yielded no significant results (see Table 4). Correlations were also run with PTSD, the SSS, the WAS, and the demographic information and yielded no significant results (see Table 5).
To examine the correlation between Stockholm Syndrome, PTSD, and WAS, the SSS, PTSD, and WAS were divided along their medians (SSS median = 154, PTSD median = 58, WAS median = 18) The scores with high or low scores were then correlated among the PTSD, SSS, and the WAS scores resulting in only one significant relationship. When SSS is low, PTSD correlated with the WAS yielded a Pearson r of .512, p < .03. This is indicative of previous results in which the amount of abuse is directly correlated with the level of PTSD suffered by the abused women but also indicates that this is only the case when Stockholm Syndrome is low. This indicates that Stockholm Syndrome may have a mediating effect on the relationship between PTSD and abuse at higher levels. The other correlations between PTSD, SSS, and WAS resulted in no significant relationships, p > .05 (see Table 6).
Although it was not significant, a pattern did emerge from the median split of the PTSD and SSS measures. The correlation between SSS and PTSD in participants high in SSS yielded a Pearson r of .149 while the correlation between SSS and PTSD in participants low in SSS yielded a Pearson r of .253, p > .05. This may be indicative again of the mediating effect Stockholm Syndrome has with some women and the development of PTSD symptoms. The effect is not generalized to the whole population however, resulting in the lower correlation at higher levels of Stockholm Syndrome. Another pattern which emerged was the relationship between PTSD and the amount of abuse when PTSD was split along the median. The correlation between PTSD and WAS when PTSD was high resulted in an r of .227 while the result when PTSD was low was an r of .401. This indicates that the more abuse the women are suffering, the more strongly it is correlated with symptoms of PTSD.
There was an interesting, albeit nonsignificant pattern, which emerged when the Stockholm Syndrome, PTSD, and abuse measures were divided by their medians and correlated with each other. First, when Stockholm Syndrome is low, a correlation between abuse and PTSD emerged. This result provides for the possibility that Stockholm Syndrome somehow mitigates the relationship between PTSD and abuse even though Stockholm Syndrome is not explicitly correlated with either PTSD or the amount of abuse the women suffer. In addition, those low in Stockholm Syndrome showed a stronger correlation with PTSD than those high in Stockholm Syndrome. This correlational pattern indicates as well that there is the possibility that Stockholm Syndrome and PTSD may, in fact, have an inverse relationship to that previously hypothesized in this study. Stockholm Syndrome may be related to PTSD in a tri-modal relationship with abuse meaning that the three conditions have a complicated and interdependent relationship. However, further research with larger sample sizes and more rigorously tested materials is needed to determine whether this tri-modal relationship is the case.
The lack of significant correlations is most likely due to several factors. The first factor which contributes to the lack of significant results is the sample size of the survey. Forty-eight women is a relatively small number of participants which may have decreased the chances of finding a significant relationship between Stockholm Syndrome and Posttraumatic Stress Disorder because of low power. Other factors which may contribute to the lack of significant correlations are the measures themselves. The modified Woman Abuse Scale has not been tested for validity and reliability. In addition, the Stockholm Syndrome Measure is fairly new and has only been tested by its creators for validity and has not been used in other studies to this date. This may result in the measure being related to other conditions which have not been studied yet and leading to inaccurate scores. The SSS and WAS did show significant internal reliability which is a good indicator that the traits represented in the subsets do indeed have a relationship in survivors of domestic violence and Stockholm Syndrome.
Another reason for the lack of significant relationships between Stockholm Syndrome and PTSD is that the conditions may be unrelated just as Favaro et al. (2000) noted in their kidnapping victims. This may be because Stockholm Syndrome is an ineffective coping strategy for the trauma of abuse and kidnapping. Perhaps, the lack of relationship is an indicator of the temporal dependence of Stockholm Syndrome and PTSD. During the course of an abusive relationship, a woman may feel a wide range of emotions including love but those emotions may fade as she approaches the end of the relationship (Graham, Rawlings, and Rimini, 1988). Stockholm Syndrome is a condition that develops during the interactions of the aggressor and their captives and fades once that relationship is terminated (Auerbach et al. 2000) and PTSD develops after the traumatic event (DSM IV-TR, 2000). Therefore, PTSD and Stockholm Syndrome may show no relationship as they may be independent and unrelated conditions that develop at different times of the relationship. This study originally posited that the coping mechanism of Stockholm Syndrome would mitigate the later effects of the trauma, but this may not be the case.
Furthermore, if research on Stockholm Syndrome and PTSD fails to find a significant relationship between Stockholm Syndrome and PTSD, then perhaps other areas of research are required to better understand the nature of Stockholm Syndrome and Posttraumatic Stress Disorder and how these conditions may or may not affect battered women. If the two conditions are, in fact, entirely discrete, perhaps an exploration into the relationship of Stockholm Syndrome with other stress measures temporally unique to the relationship will show a relationship with Stockholm Syndrome. These may include depression measures and anxiety measures for the time during which they are were in the abusive relationship.
The kind and amount of stresses that battered women experience varies from the time they enter the relationship to the time they have exited the relationship. The stressors and conditions the abused women experience during the relationship may have a closer relationship to Stockholm Syndrome than to PTSD, which is developed after the trauma, rather than during the relationship. Perhaps Stockholm Syndrome may still be considered a coping mechanism, but one that instead affects the stress and anxiety level during the relationship. If the abused women believe that their partner loves them and the abuse the abused women suffer is their fault, then perhaps they experience less anxiety because they believe they can change for the better. The opposite may also hold true. The abused women suffer more stress and anxiety because they believe they actually have control over the abuse and nothing they do has any affect.
American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed.). Arlington: American Psychiatric Association.
Astin, M. C., Lawrence, K. J., & Foy, D. W. (1993). Posttraumatic Stress Disorder among battered women: Risk and resiliency factors. Violence and Victims, 8, 17-28.
Auerback, S. M., Kiesler, D. J., Strentz, T., Schmidt, J. A., & Serio, C. D. (1994). Interpersonal impacts and adjustments to the stress of simulated captivity: An empirical test of the Stockholm Syndrome. Journal of Social and Clinical Psychology, 13, 207-221.
Bargai, N., Ben-Shakhar, G., Shalev, A. Y. (2007). Posttraumatic Stress Disorder and depression in battered women: The mediating role of learned helplessness. Journal of Family Violence, 22, 267-275.
Crowell, N. A., Burgess, A. W. (Eds.). (1996). Understanding Violence Against Women. Washington, D. C.: National Academy Press.
Favaro, A., Degortes, D., Colombo, G., & Santonastaso, P. (2000). The effects of trauma among kidnap victims in Sardinia, Italy. Psychological Medicine, 30, 975-980.
Graham, D. L. R., Rawlings, E. I., Ihms, K., Latimer, D., Foliano, J., Thompson, A., Suttman, K., Farrington, M., & Hacker, R. (1995). A scale for identifying “Stockholm Syndrome” reactions in young dating women: Factor structure, reliability, and validity. Violence and Victims, 10, 3-22.
Graham, D. L. R., Rawlings, E., & Rimini, N. (1988). Survivors of terror; Battered women, hostages, and the Stockholm Syndrome. In Yllo, K., & M. Bograd (Eds.), Feminist perspectives on wife abuse (pp. 217-233). Thousand Oaks: Sage Publications, Inc.
Hughes, M. J. & Jones, L. (2000). Women, domestic violence, and Posttraumatic Stress Disorder. Family Therapy, 27, 125-139.
Kuleshnyk, I. (1984). The Stockholm Syndrome: Toward an understanding. Social Action and the Law, 10, 37-42.
Lawler, C., Ouimette, P., & Dahlstedt, D. (2005). Posttraumatic Stress symptoms, coping, and physical health status among university students seeking health care. Journal of Traumatic Stress, 18, 741-750.
Masho, S. W., & Ahmed, G. (2007). Age at sexual assault and Posttraumatic Stress Disorder among women: Prevalence, correlates, and implications for prevention. Journal of Women’s Health, 16, 262-271.
Saunders, D. G. (1995). Prediction of wife assault. In J. C. Campbell, (Ed.), Assessing dangerousness: Violence by sexual offenders, batterers, and child abusers. Interpersonal violence: The practice series, Vol. 8. (pp. 68–95). Thousand Oaks, CA: Sage Publications.
Straus, M. A. (1979). Measuring intrafamily conflict and violence: The Conflict Tactics (CT) scales. Journal of Marriage and the Family, 41, 75-88.
Walker, L. E. (1991). Posttraumatic Stress Disorder in women: Diagnosis and treatment of Battered Woman Syndrome. Psychotherapy, 28, 21-29.
Table 1: Comparison of Measures and Demographic Features Between States and Races
|Mean||Standard Deviation||Mean||Standard Deviation|
Table 2: Demographic Information
Table 3: Descriptive Statistics of the PTSD Checklist, Stockholm Syndrome Scale, and Woman Abuse Scale
Table 4: Correlations of Full Scale Scores with Scale Subsets
* denotes significance at p < .05
Table 5: Correlations of PTSD, SSS, and WAS with Demographic Features
Table 6: Correlations of PTSD, SSS, and WAS with Median Splits
*denotes significance at the p < .05 level