Cultic Studies Review, 2(3), 2003, 203-224
This paper presents preliminary results for the adaptation of the Group Psychological Abuse Scale (GPA) (Chambers, Langone, Dole & Grice, 1994), a measure of group psychological abuse, to a Spanish population. This scale is unique in assessing the varieties and extent of psychological abuse in group contexts. The Spanish translation of the scale has been administered to 61 self-identified former members of diverse manipulative groups who had involvements with any of a total of 21 different groups. The findings on the psychometric properties of the Spanish version of this scale indicate that it is a reliable and valid instrument that reveals a structure of group psychological abuse composed of three factors: Compliance, Mind Control and Exploitation.
The existence of dogmatic groups with a certain paranoid component arouses considerable social unease (Jordán, 1991; Rodríguez, 1994). In these groups some people, who adopt beliefs, attitudes, and behaviors that would have been strange to them in the absence of psychological manipulation, place themselves at the service of a doctrine used as an instrument by the leader and/or group. In this context, a motivation and justification is found for the performance of sometimes illegal, social actions.
Given the “ambiguity” of the concept “cult”— "secta," "secte," "setta," "sekta" are the preferred terms in Europe— (Langone, 1994) and the caution that is advised when applying the term to a specific group (Langone, 2001), the concept of psychological abuse is crucial to delimit and differentiate coercive cults from other minor and legitimate groups that do not employ techniques of psychological manipulation. The concept of psychological abuse, according to Langone (1992), refers to practices through which people are treated as objects to be used and manipulated, instead of subjects whose mind, autonomy, identity, and dignity must be respected. Depending on the use, frequency, and intensity of abusive methods, distinctions can be made between manipulative and benign groups (Chambers, Langone & Malinoski, 1996).
The Group Psychological Abuse Scale (GPA) (Chambers, Langone, Dole & Grice, 1994) constitutes an empirical approximation of the characterization of group contexts as a function of the varieties and the extent of psychological abuse. This scale has been used with North American subjects, with former members of different manipulative groups (Chambers et al., 1994; Aronoff, Marshall, Whitney, Malinoski & Martin, 2001) and with former members of the same manipulative group together or compared to former members of non manipulative groups (Adams, 1998; Gasde & Block, 1998; Langone, 1996; Malinoski, Langone & Lynn 1999).
In order to generate the items of the scale, the investigators considered: (a) a Delphi study conducted by Dole and Dubrow-Eichel (1985) that examined the perceptions different experts had on cults; (b) Langone's (1992) theoretical analysis of psychological abuse; and (c) a review of the clinical literature on cults and thought reform programs. Afterwards, a factor analysis of the mail responses given by 308 former members of 101 different groups to 112 descriptive items was performed. Subjects were contacted through lists provided by associations that provided educational programs on cults, ex-member organizations, and professionals.
This paper presents preliminary data on the adaptation of the GPA to the Spanish population. Besides contributing to the need for a reliable instrument that could evaluate psychological abuse in group contexts, this study, by looking at group psychological abuse across different cultural groups, lays the groundwork for comparative studies that may provide a multicultural knowledge base, which would be more valuable than knowledge based on a single culture.
The Spanish version of the GPA scale was administered to 61 subjects from different provinces within Spain. These subjects were contacted through data provided by associations working in the area of information, education, and advice on group psychological manipulation (42.6%), mental health professionals, who were not necessarily in touch with the subject matter (21.3%), and some former members, who were taking part in this study and facilitated contact with other former members (36.1%). The inclusion criterion was that subjects identified themselves as former members of groups using manipulative methods to attract, indoctrinate, and keep members inside the group. This criterion implies a feeling of belonging or of having been a member of the group and excludes those who had just a brief contact.
Out of the 61 selected subjects, 68.9% (42 subjects) participated in the interviewer’s presence, whereas 31.1% (19 subjects) completed the Scale through ordinary mail. A total of 34 subjects (55.7%) were male (55.7%), and 27 female (44.3%). Participant’s mean age was 44.33 years (SD = 11.0). They had belonged to one of a total of 21 different groups. They got in touch with the groups at a mean age of 29.04 years (SD = 12.10) and joined them at a mean age of 29.58 years (SD = 11.13), ranging, in both cases, from a minimum age of 3 years to a maximum of 56 years. The average length of membership was 8.75 years (SD = 7.96), with a range of 0.33 to 31.25 years of membership since they first met their groups. The average length of time between exiting the group and participating in the study was 6.50 years (SD = 7.11), with a median of 3.58 years and a range of 0.25 to 33.33 years.
Spanish version of the Group Psychological Abuse Scale (GPA). Developed by Chambers, Langone, Dole & Grice (1994), the GPA Scale is a standardized measure used to evaluate psychological abuse in group environments. It is a self-administered instrument, easily understood and quickly completed by subjects. It consists of 28 items, 7 for each of the four subscales: Compliance, Exploitation, Mind Control, and Anxious Dependency. Each item is rated on a 1-5 Likert scale ranging from 1= not at all characteristic to 5 = very characteristic, with a possible range for each subscale from 7 to 35, and for the global measure a range from 28 to 140. Scores above the midpoint—21 for each subscale and 84 for the global scale—are considered positive, indicating that the subject perceived the group as abusive. The reliability coefficients identified in the original work were, according to the authors, satisfactory, varying from 0.70 for Mind Control to 0.81 for Compliance.
For the Spanish adaptation of the GPA scale, we followed the international methodological standards recommended by the International Test Commission (ITC) in order to adequately adapt instruments from one culture to a new one (Hambleton, 1994; Hambleton, 1996; Muñiz & Hambleton, 2000). Specifically, we followed these steps:
The evaluations were carried out in 13 different provinces inside Spain, corresponding to the different places of residence of the subjects who took part in the study. Certain characteristics of some members of the sample, i.e., difficult to get in touch with, a reluctance to attend an interview in person, geographical distance, made us consider two
Figure 1: Adaptation process
different ways to take part in this study: one was physical attendance, and the second was by ordinary mail. In both cases, a qualified interviewer informed the subjects about the characteristics of the study (those who participated via ordinary mail were informed by phone), and once they gave their informed consent to take part in it, they were provided with the necessary instruments to be completed and returned.
Statistical analyses were carried out using the statistical program SPSS-PC, version 10.0 for Windows, as follows: First, the Student’s t distribution helped us examine the existence of statistically significant differences among the global scores of the Attending and Mail samples. The following analysis focused on the psychometric characteristics of the Spanish version of the scale, taking as a reference the original structure. Next we used the Principal Components Factor Analysis with Oblimin rotation to examine the construct validity of the scale for the Spanish sample and observed psychometric properties for the new factorial structure. Finally, taking the average scores for each item in the Spanish and North American samples, we used the Mann-Whitney U test to compare the mean of the average scores related to the new subscales found in the Spanish sample.
The 61 participants were initially assigned to two different groups depending on the kind of participation: Attending (n=42) and by Mail (n=19). As no significant differences (t=1.150; p=0.255) were found between both groups in the global scores for the GPA scale, from then on both groups were considered together for the rest of the analysis.
We analyzed the 28 items of the original GPA scale. In Table 1, there is an English language description of the items and the average, standard deviation and the item-total correlation for each item. Items 1, 5, 21, 22, 25 and 26, which inversely evaluate the construct they refer to, are shown in this table previously recodified. As it can be observed, three of the items show a low correlation with the total of the scale, in particular item 5 (“medical help okay”), whose index (-0.044) indicates that it is barely related to the rest of the items, and that this almost non existent relation goes in the opposite direction to that of the rest of the scale. Similarly, items 6 (“politics major goal”) and 22 (“can think critically”) take the values of 0.16 and 0.17 respectively. According to the original authors of the scale, these three items are related to the following subscales: Item 5 to Anxious Dependency, Item 6 to Exploitation and Item 22 to Mind Control.
Correlations between each item of the test and its corresponding subscale were also calculated, showing an average item-subscale correlation of 0.45 for Compliance, 0.40 for Exploitation, 0.32 for Anxious Dependency, and 0.43 for Mind Control. The average item-global scale correlation was 0.48, which is higher than the same correlation obtained for the original North American sample (0.42).
Taking the original structure of the instrument, and in order to examine its reliability, we calculated the internal consistency using Cronbach’s Alpha index. The values found show, in general, an adequate internal consistency of the scale for the Spanish sample (α =0.87), which is even higher than the consistency showed by the original scale (α =0.81). As far as the subscales are concerned, values are, in almost all cases, less than those obtained by the authors of the original scale, especially for the subscale Anxious Dependency (α =0.59), which shows an alpha index considerably lower than the same subscale in the original North American sample (α =0.72). The Mind Control subscale, however, shows a higher value for the Spanish sample (α =0.76) with regard to the original (α =0.70).
Pearson correlation coefficients were also calculated between the subscales and between subscales and the global scale. In both cases, the correlations reached the significance level of 0.01.
The construct validity, or the organization and precision of the constructs analyzed within a coherent theoretical framework (Muñiz, 1998), was examined through an exploratory factor analysis. This analysis helped elicit latent variables, which were lying under the group of observed variables (García, Gil & Rodríguez, 2000). This method has become one of the most frequently used alternatives when there is an absence of a pattern or criterion, as in our case. The extraction method used was Principal Components and Oblimin rotation.
Table 2 shows, first, the factors obtained, as well as the percentage of the variance attributable to each of them, and second, the eigenvalues of the rotated factors. As can be observed, three factors accounted before rotation for 42.3% of the total variance. Factor 1, labeled Compliance, explains 26.6% of the variance; Factor 2, Mind Control, explains 8.7%; and Factor 3, Exploitation, explains the remaining 7.01%. Table 2 shows as well the results of the factor analysis. When considering the direction in which items saturate, it is necessary to remember that items 1, 5, 21, 22, 25 and 26 inversely evaluate the construct they refer to. The direction of the items of the second factor, known as Mind Control, shows that it refers to the positive aspect in the sense of psychological autonomy. Regarding item 9, with almost identical saturation for factors 1 (0.410) and 2 (-0.408), we have decided to keep it in Factor 2. Item 5 which, as previously noted, was scarcely related to the rest of the items has been removed from Factor 2, where it shows a slightly higher saturation, because of its discordant behavior with the rest of the items which compose Factor 2.
Given the new subscales, correlations between each item of the test and its corresponding scale and subscale were calculated. Table 3 shows the range of item-scale correlation values (out of 27 items, not including item 5) as well as item-subscales, average and number of items belonging to each of the subscales (n). Item-subscale correlations were all over 0.20 for all of the items, with the average values higher than those obtained in the original distribution of the scale.
We also analyzed the reliability of the three subscales by calculating the internal consistency using Cronbach’s Alpha Index. Table 3, contains the internal consistency values obtained for each subscale. As it can be observed, all the values are above 0.70, showing an adequate internal consistency for each of the three subscales. On the other hand, the final 27-item scale (once item 5 was removed) produced an alpha index of 0.88.
The average score of this 27 item Spanish version of the GPA scale was 99.82 (SD=18.66) for the sample of 61 Spanish subjects. As far as the 9 item subscales are concerned, Mind Control showed the highest average score (37.21; SD=6.65) followed by Compliance (36.91; SD=8.34), and Exploitation with the lowest average score (25.75; SD=8.07).
Taking the average scores for each item in the original North American and Spanish samples, we used the Mann-Whitney U test in order to analyze whether there were statistically significant differences amongst means of the average scores obtained in the North American and Spanish samples in the Spanish version of the scale and subscales. Figure 2 shows the means obtained in both samples and the Mann-Whitney U test values for the global scale and for each subscale. There were no statistically significant differences either among means of the average scores obtained in both samples for the total scale or for any of the subscales. Although the means for the Spanish sample were slightly lower, a similar pattern can be observed for the varieties of psychological abuse.
Results obtained in the present work, though preliminary, show an adequate validity and reliability for the Spanish version of the GPA scale, with enough empirical evidence to support its appropriateness for evaluating psychological abuse in group contexts.
Of the three subscales found for our sample, the first, Compliance, refers to a relationship of submission where the subject bows to the group’s will and/or the leader, who favor a strong dependency among members. This dependency is emotionally expressed towards the leader, who is regarded as a superior being, not subject to criticism or negative
Figure 2: Differences between mean scores of the Spanish and North American samples.
expression. Fear of the consequences if they dare to leave the group generates insecurity and becomes a factor that justifies any instruction given by the leader. In this way, compliance becomes a behavior in the shape of servitude towards the leaders and their instructions, even with sacrifice of personal goals and belongings, which leads to final acceptance and dissemination of the doctrine. The second subscale, Mind Control, refers to members' regarding themselves as a part of a special elite group and to the use of coercive persuasion by the group, that is, the use of deceit and psychological pressure, members' inability to make their own decisions or independent, critical judgments, and the use of thought-stopping techniques. The third subscale, Exploitation, describes the relationship with the outer environment of the group. In this sense, there is a clear isolation from the outside; groups are closed; members live together; and even intimate relationships, if allowed, are controlled. This subscale also includes a component of deviation from the norm, that is, the behavior of the group regarding people who are not related to it (threats, violence) or the environment (breaking the law). So far as behaviors inside the group are concerned, this subscale refers to a certain despotism towards the members, who are criticized or led to certain activities or attitudes such as prostitution or rejection of medical care.
When comparing this factor structure to the one found by the original authors, there are similarities in the dimensions for both versions. The subscale Compliance in the Spanish sample is mostly the product of elements of two subscales, Compliance and Anxious Dependency, in the original version. In this sense, we understand that both original subscales are conveniently expressed under the term Compliance, taking into consideration that the construct refers both to items describing submission and obedience to an authority figure (leader/s or group), as well as extreme dependency to it. Thus, in the new subscale, Compliance, we find the four items from Anxious Dependency that mainly accounted for the covariance in the original subscale, those with the highest values of item-total correlation (Items: 7, 8, 10 and 23; Values: 0.49, 0.34, 0.54 and 0.39 respectively). The remaining three items of Anxious Dependency: 5, 16 and 17, had lower values for the item-total correlation (0.03, 0.19 and 0.25). Item 5 was excluded from the scale as it hardly related to the global scale. We are of the opinion that item 16 (“exercises remove doubt”), that has been moved to the subscale Mind Control for the Spanish sample, refers better to this one than to Anxious Dependency because, although the exercises described in the item are driven by negative emotions (fear, guilt) experienced by the members, the item does not ask about them, but refers to the purpose of those exercises (cognitive control). Finally, we also consider that item 17 (“no medical help”) is better placed in the subscale Exploitation, to which it is moved for the Spanish version, as it represents an action carried out by the group to the detriment of its members. As far as the subscale Mind Control obtained in the Spanish version is concerned, it corresponds almost perfectly with the same subscale in the original instrument. Finally, the subscale Exploitation has remained almost the same as in the original.
In our study we have observed an important fact, which is the similarity of the varieties of psychological abuse shown by former members of different manipulative groups in two cultural contexts: North American and Spanish. When comparing the average scores for both samples, we found that in both cases the subscale with the highest average score was Mind Control (Spanish average: 37.21; North American average: 39.70), followed by Compliance (Spanish average: 36.91; North American average: 37.36) and finally the subscale Exploitation (Spanish average: 25.75; North American average: 30.13).
All this strengthens the hypothesis that the most consistent factor in abusive groups is mind control, leaving at a second level other aspects like isolation, which is more characteristic of other kind of organizations or of brainwashing processes, mentioned by Lifton (1961) in his initial studies.
Among the limitations of this study, we could mention the difficulty in verifying the representativeness of the sample regarding the universe of different situations of people who have been involved in manipulative groups. All this is due to the difficulty, usual in this kind of study, of getting in touch with former members of manipulative groups in a way that produces adequate sample sizes and permits the use of probabilistic methods, especially when, as in our case, 68.9% of the sample were interviewed, with just a few subjects being surveyed after information was provided by phone (Almendros, 2001; Langone, 1994). Our sample of 61 subjects is not the most adequate in size for the kind of analysis we have used, especially if we consider the large number of items in the original scale (28 items). Therefore the data obtained should be viewed as preliminary. However, the ratio subjects/items in our study is similar to the one in the original study (308/112).
In future studies, it will be helpful to analyze the diagnostic validity of the Spanish version of the scale, using as comparison a group of former members of non manipulative groups. This has been done in two published North American studies (Adams, 1998; Langone, 1996; Malinoski, Langone & Lynn, 1999) and one unpublished study (Mascarenas, 2003).
Adams, D.L. (1998). Perceived Psychological Abuse and the Cincinnati Church of Christ. Brief Report. Cultic Studies Journal, 15(1), 87-88.
Almendros, C. (2001). Investigación en el ámbito de las sectas coercitivas. In Asesoramiento e Información sobre Sectas (Ed.), Jornada trastorn de la dependència grupal en els grups de manipulació psicològica o sectes coercives (pp. 39-50). Barcelona.
Aronoff, J., Marshall, R., Whitney, J., Malinoski, P.T., Martin, P.R. (2001). Psychological Research. AFF 2001 Annual Conference. Video production.
Chambers, W.V., Langone, M.D., Dole, A.A., & Grice, J.W. (1994). The Group Psychological Abuse scale: a measure of the varieties of cultic abuse. Cultic Studies Journal, 11(1), 88-117.
Chambers, W.V., Langone, M.D., & Malinoski, P.T. (1996). The Group Psychological Abuse Scale. Presented to Division 36 (Psychology of Religion) American Psychological Association Annual Meeting. Toronto, Canada. Retrieved November 7, 2001, from http://www.csj.org/studyindex/studyresearch/study_ gpa.htm
Dole, A.A., & Dubrow-Eichel, S. (1985). Some new religions are dangerous. Cultic Studies Journal, 2(1), 17-30.
García, E., Gil, J., & Rodríguez, G. (2000). Análisis Factorial. Cuadernos de Estadística,7. Madrid: Editorial La Muralla.
Gasde, I., & Block, R.A. (1998). Cult Experience: Psychological abuse, distress, personality characteristics, and changes in personal relationships reported by former members of Church Universal and Triumphant. Cultic Studies Journal, 15(2), 192-221.
Hambleton, R.K. (1994). Guidelines for adapting educational and psychological tests: a progress report. European Journal of Psychological Assessment, 10(3), 229-244.
Hambleton, R.K. (1996). Adaptación de tests para su uso en diferentes idiomas y culturas: Fuentes de error, posibles soluciones y directrices prácticas. In J. Muñiz (Coord.), Psicometría. Madrid: Universitas.
Jordán Villacampa, Mª.L. (1991). Las sectas pseudorreligiosas. Ministerio de Justicia. Secretaría General Técnica. Centro de Publicaciones. Madrid.
Langone, M.D. (1992). Psychological abuse. Cultic Studies Journal, 9(2), 206-218.
Langone, M.D. (1994). Investigación sobre los cultos destructivos. In Asesoramiento e Información sobre Sectas (Ed.), II Congreso Internacional sobre Grupos Totalitarios y Sectarismo (pp. 63-83). Barcelona.
Langone, M.D. (1996). Study reveals cultic group’s abuses Boston Church of Christ / International Churches of Christ Movement. AFF. Cultic Studies Study Resources. Research. Retrieved November 7, 2001, from http://www.csj. Org/ studyindex/studyresearch/study_reshboston.htm
Langone, M.D. (2001). Cults, psychological manipulation, and society: International perspectives - an overview. Cultic Studies Journal, 18, 1-12.
Lifton, R.J. (1961). Thought reform and the psychology of totalism. New York: W.W. Norton.
Malinoski, P.T., Langone, M.D., & Lynn, S.J. (1999). Psychological distress in former members of the International Churches of Christ and noncultic groups. Cultic Studies Journal, 16(1), 33-51.
Muñiz, J. (1998). La medición de lo psicológico. Psicothema, 10(1), 1-21.
Muñiz, J., & Hambleton, R.K. (2000). Adaptación de los tests de unas culturas a otras. Metodología de las Ciencias del Comportamiento, 2(2), 129-149.
Rodríguez, P. (1994). Tu hijo y las sectas Madrid: Ediciones Temas de Hoy.
Zablocki, B. (1997). Paper presented to a conference, “Cults: Theory and Treatment Issues,” May 31, 1997 in Philadelphia, Pennsylvania
Esta escala ha sido diseñada para evaluar determinadas características de los grupos religiosos, psicoterapéuticos, políticos y comerciales, entre otros. Por favor, evalúe lo mejor que pueda el grado en que las siguientes afirmaciones caracterizan el grupo que está usted considerando. Evalúe retrospectivamente cada afirmación de acuerdo con su experiencia y sus observaciones sobre cómo funcionaba el grupo en la práctica. Si su grupo tenía distintos niveles o jerarquías de pertenencia, con claras diferencias entre ellas, por favor, aplique sus evaluaciones al nivel sobre el que tenga un mayor conocimiento. Piense cuidadosamente cada respuesta de modo que proporcione la evaluación más apropiada. Rodee con un círculo el número correspondiente.
A.2. Response choices:
1 – Nada característico; 2 – Poco característico; 3 – No podría decir/ No estoy seguro/a; 4 – Bastante característico; 5 – Totalmente característico.
This work has been possible thanks to the participation of people who have taken part in it and/or have encouraged others to participate, as well as the fine work of our two translators: Montse and James. It was financially supported in part by the American Family Foundation (AFF) under the project title, “Spanish Adaptation of the GPA Scale”. The authors appreciate the comments and suggestions made by Peter Malinoski, Ph.D. and Michael Langone, Ph.D. The paper is a slightly edited translation of the original Spanish version of the paper, "Adaptacion Psicometrica de la Version Espanola de la Group Psychological Abuse Scale Para la Medida de Abuso Psicologico en Contextos Grupales, which was published in Psicothema, Vol. 15, No. 4 (2003), pp. 132-138, and is translated with that journal's permission.
Cultic Studies Review Vol. 2, No. 3, 2003, Page
 For the purposes of this paper, the term "cult" ("secte," "secta," "setta," and "sekta") will be defined according to Zablocki (1997): "an ideological organization held together by charismatic relationships and demanding total commitment."
Carmen Almendros is a doctoral student in the Clinical and Health Psychology Program at the Universidad Autónoma de Madrid. She has received a public research grant and teaches a postgraduate course in clinical psychology at the same university.
José Antonio Carrobles, Ph.D., is Full Professor of Personality, Assessment and Treatment and Director of the Biological and Health Psychology Department of the Universidad Autónoma de Madrid. His work focuses in the areas of Psychopathology and Clinical and Health Psychology. He is President of the European Association for Behavioural & Cognitive Therapies (EABCT). He has directed numerous Doctoral Theses and is author of an important number and variety of articles and books in his areas of specialization. He has organized and participated in numerous national and international psychology congresses, among which stands out his participation as President of the Scientific Committee at the "23rd International Congress of Applied Psychology" held in Madrid in 1994. He is member of the Editorial Boards of several national and international journals.
Josep Maria Jansà, M.D. Since interning at AFF in 1985, Dr. Jansà, a medical doctor, has worked with AIS (Assessment and Information on Cults—"sectas" in Spanish) since 1984, where he has assisted families, group members, and former group members. The AIS clinic has dealt with more than 1000 patients since January 1986. Dr. Jansà has participated in research initiatives and issued various publications on this topic. He also works as an epidemiologist for the City Council of Barcelona and is responsible for international health and health and migrations.
Álvaro Rodríguez-Carballeira, Ph.D. is professor of Social Psychology, Social Movements, and Legal Psychology at the University of Barcelona (Spain). Since 1999 he has been Director of the Social Psychology Department. During the 1980s, before and after a 1985 internship at AFF, he worked with families and victims affected by cult membership. He then worked as a professor at the University of Barcelona, where he completed a doctoral dissertation in 1991 on psychology of coercive persuasion. During recent years he has extended this line of research, linking it to other contexts (e.g., domestic, work, school) where manipulation and psychological violence may occur. His publications include the book, El Lavado de Cerebro: Psicología de la Persuasión Coercitiva. (Brainwashing: Psychology of Coercive Persuasion).