The relationship between monogamous/polygamous family structure and the mental health of bedouin Arab adolescents
Abstract
Previous studies of polygamy and child mental health have primarily focused on younger children. The present studies are among the first to focus on adolescents. The first study involved 210 randomly selected Bedouin Arab adolescents (mean age 15.9), who were administered instruments assessing their family environment and mental health. The second study involved 182 Bedouin Arab adolescents in which the student participants completed a single instrument about themselves and in which the teachers of the students completed the Teachers’ Report Form of the Child Behaviour Checklist by Achenbach. The Bedouin Arab adolescents fell into two groups: (a) adolescents in monogamous family structures and (b) adolescents in polygamous family structures. The findings of the first study suggest that the two groups did not differ significantly in the majority of the assessed variables, even though there were significant differences obtained between groups for 4 of the 13 assessed variables. The two groups did not differ significantly in the second study. Results were discussed in terms of their cultural and developmental significance.
Introduction
Psychologists and other social scientists have extensively studied the influence of the family on children (e.g. Bronfenbrenner, 1986). However, most of these studies were conducted on monogamous family structures in Europe and North America (e.g. Lansford, Ceballo, Abbey, & Stewart, 2001; Lindner, Hagan, & Brown, 1992). The number of published studies on the effects of polygamous family structures (primarily in Africa, the Middle East, and Asia) on the indices of mental health of children is sparse (Elbedour, Bouchard, & Hur, 1997).
Polygamy is common in collective societies (especially in the Middle East and Africa) where the child mortality rate is high and children are valued for their labour, emotional support, and social and economic security that they offer to their parents. Hence, families are motivated to have as many children as possible, and polygamy becomes an attractive vehicle to achieve such an objective. In addition to their larger size, polygamous families are also characterized by lower socioeconomic status and higher rates of marital conflict, family violence, and unemployment than monogamous families (Elbedour, Bart, & Hektner, 2000, 2003; Al-Krenawi, 1988). For example, Eapen, Al-Gazali, Bin-Othman, and Abou-Saleh (1988) along with Kampambwe (1980) found that polygamous family life weakens the parent–child bond, thus reducing the child's emotional satisfaction and psychological security.
The lack of parental involvement was found to place children at risk for developmental disorders (Farber & Egeland, 1987), and the most damaging form of maltreatment is psychological unavailability (Egeland, Sroufe, & Erickson, 1983). The large sizes of polygamous families also tend to reduce the amounts of support and parental interaction received by the children, especially from the fathers. Absence of family support and warmth plays a risk factor in the development of childhood stress (Garmezy, 1983).
The association between polygamy and child abuse also appears to spill over to related indices of educational adjustment. For example, Brassard, Hart, and Hardy (1993) found that children of psychologically neglectful mothers in monogamous family structures were more developmentally delayed in school and preschool settings than their peers, who had attentive mothers. Eapen et al. (1988), Kampambwe (1980), and Owuamanam (1984) provided similar findings for polygamous families.
Cherian (1994) showed that school achievement test scores of children from polygamous families were significantly lower than those of children from monogamous families in South African Xhosa society. Additional data from polygamous families highlight the higher levels of conflict, hostility, and aggression within these families due to unbalanced treatment of wives by their husbands (Hassouneh-Phillips, 2001), financial anxiety and insecurity associated with supporting larger families (Elbedour, Onwuegbuzie, Caridine, & Abu-Saad, 2002), and other family challenges such as full and half sibling rivalry and jealousy (Al-Krenawi & Lightman, 2000; Oyefeso & Adegoke, 1992). Hence, it would appear that a polygamous home environment is less nurturing and more distressing to a child than a monogamous one.
However, in some cultures, polygamy is considered to be normal and consistent with cultural values and ideals, and perhaps may not have a deleterious effect on children. For example, polygamy is generally accepted and widely practiced in the Bedouin Arab community involved in this study. Larger families tend to have higher status than smaller families in this community. Polygamy can certainly contribute to family size.
Children often benefit from higher family status, because their levels of self-esteem often reflect the levels of status of their families. Larger families may also compensate for disadvantages of polygamy by allowing children to have more interpersonal relationships with adults and other children (Elbedour et al., 2000). In addition, the Bedouin Arab culture is collective (e.g. Elbedour, Shulman, & Kedem, 1997) and, as a result, makes available an extensive robust support network that could reduce the dependence of a child on the family (Belsky & Isabella, 1988).
A large family structure may benefit children by providing them with a wider range of adult role models and caregivers (e.g. older siblings). Given that children are highly valued in polygamous societies, polygamous families may generate more warmth and affection than their monogamous counterparts, thus providing greater psychological support to children. In other words, cultural norms may increase the efficacy of polygamous families to provide positive psychological support for children. The present studies are among the first designed to examine this issue. Their goal is to evaluate differences in psychopathology, self-esteem, family functioning, and violence between Bedouin Arab adolescents from monogamous and polygamous family structures.
To be more specific, three research questions are investigated in this paper. First, do adolescents from monogamous family structures differ from adolescents from polygamous family structures in their perceptions of family cohesion and violence in their lives? Second, do adolescents from monogamous family structures differ from adolescents from polygamous family structures in their reports of psychological well-being? Third, do adolescents from monogamous family structures differ from adolescents from polygamous family structures in the relationship of psychological well-being to perceptions of family cohesion and violence?
Study 1
Method
Participants
All of the participants for this study were Bedouin Arab youth living in the Negev region of Israel and attending one of six schools serving children enrolled in grades 7–12. This grade range (ages 13–18) was selected for two main reasons: (a) it was assumed that the participants have the ability to read, understand, and respond to the items on the instruments without difficulty, and (b) the participants have developed the cognitive maturity to reflect critically on the items presented. A list of the entire membership of Bedouin Arab children (grades 7–12) attending school was obtained from the principals, and 300 names (approximately one-tenth of the entire population) were selected randomly. The school administration and selected students granted their consent for participation in the study.
Upon request, the participants completed a demographic data sheet in which they noted whether they came from monogamous or polygamous families. The obtained information was checked with school officials for accuracy, and 90 participants were dropped from the sample, because the information they provided was either inaccurate or incomplete. The selected sample of 210 students consisted of 98 (46.7%) females and 112 (53.3%) males.
Nearly one-half (44.5%) of the participants were enrolled in the 10th grade, while the remaining participants were distributed among the five other grades. The mean age for the participants was 15.9 years (s.d.=1.3) and the majority (54.3%) of them came from polygamous households with 66.7% having fathers with two wives, 31.6% having fathers with three wives, and 1.7% having fathers with four wives. The children with fathers having 3 or 4 wives were collapsed into one group for analysis purposes. Furthermore, nearly one-half (48.1%) of the participants indicated that their parents were blood relatives with first cousins being the closest relation between spouses.
Instruments
Family structure and demographic information. On a questionnaire, participants reported the numbers of wives to whom their fathers were married. They also reported their age, grade level, gender, number of siblings, education level attained by their parents, and whether their parents were blood relatives.
Abbreviated form of the Moos Family Environment Scale (FES). The salient dimensions (cohesion, expressiveness, conflict, and control) of family interaction likely to be related to the mental health of children were selected from the Moos Family Environment Scale (Moos & Moos, 1986). Hence, an abbreviated form (four sub-scales, 59 items) of this 10-sub-scale 90-item instrument was used in this study. Unfortunately, internal consistencies (Cronbach's alphas) obtained for the current sample were unacceptably low on three of the scales, and thus only the cohesion scale was retained for the current study (alpha=.61). The cohesion scale includes items such as “Family members really help and support one another.” These alphas were based on the Arabic sample.
ACHIEVE: Assessment of Children's Exposure to Violent Events (Fox & Leavitt, 1994). This 60-item questionnaire consists of items with cartoon features. It measures the frequency with which a child experiences violent acts perpetrated against him/herself or others within three environmental settings: home, school, and community.
The items were divided evenly (one-third each) among the three settings (home, school, and community), and allocated approximately evenly between whether the child or others were victims of the violence. The items were graded in terms of severity from low (e.g. “An adult shouts at someone other than your self”) to extreme (e.g. “Someone shoots a gun at you”) forms of violent acts. Other forms of violent acts included in the questionnaire were pushing and being arrested by the police.
Each participant responded on a five-point scale ranging from zero (never) to four (often) to indicate level of subjection to violence. Three sub-scales representing each of the three settings were computed individually to reflect the mean of the items pertaining to the context (home, school, and community) being measured. An overall mean across all three contexts was also computed. Internal consistencies (Cronbach's alphas) for these scales ranged from .80 to .84.
Derogatis Symptom Checklist-Revised (SCL-90-R). This is a self-administered inventory of 90 items designed to differentiate measures of distress in nine symptom categories, plus an overall Global Severity Index (GSI). The GSI, anxiety, depression, and hostility scales were the only scales examined in the current study. Derogatis (1983) believes that the GSI “represents the best single indicator of the current level or depth of the disorder, and should be utilized in most instances where a single summary measure is required” (p. 4). This instrument was translated into Arabic and used in several research projects involving Arab-Palestinian populations (see, e.g. Elbedour, 1998). Internal consistencies (Cronbach's alphas) obtained for the current sample ranged from .72 to .97.
Arabic Version of The Coopersmith Self-Esteem Inventory. This 25-item instrument was translated into Arabic, and standardized on Arab children by Musa and Dassouki (1981). Participants indicated whether each statement was “Like me” or “Unlike me”. Examples of items included “There are lots of things about myself that I would change if I could” and “I am a lot of fun to be with.” Items were scored either one or zero with one being reflective of higher self-esteem and zero being reflective of a lower level of self-esteem. The sum of the scores on all of the items was computed for each participant to indicate the overall self-esteem score. The first author used this instrument in several research projects on Bedouin high school students (see, e.g. Elbedour, 1998). The internal consistency (Cronbach's alpha) obtained in the current sample was .69.
Cross-cultural validity and translation of the research instruments. The wording and translation of the research instruments followed the guidelines for cross-cultural research suggested by Brislin (1980, 1986) and van de Vijver and Leung (1997). The instruments (ACHIEVE, Arabic Version of the Coopersmith Self-Esteem Inventory, and the Abbreviated form of the Moos FES), originally constructed and written in English, were first translated into Arabic, and then back-translated by bilingual, bicultural Arab natives using standard back-translation techniques. Professional psychologists in the Negev checked the translated instrument for preservation of meaning and cultural appropriateness. Finally, pilot testing was completed to insure that the instruments were comprehensible and ethnographically valid.
Data analysis
To test whether family structure plays a role in self-esteem and psychopathology, t-tests and ANOVAs were used to compare groups of students from families with different structures. The first comparisons made were between students coming from monogamous and polygamous families as a whole. The monogamous group was then compared with two groups of polygamous families (two mothers versus three and four mothers). Among these initial comparisons were comparisons to a normative sample from the USA. Such comparisons to the normative sample are intended to assist readers in deciding whether group means indicate high or low values of a variable such as cohesiveness.
Next, correlations between environmental variables (family cohesion, violence, number of siblings, and parent education) and mental health variables (self-esteem, depression, hostility, anxiety, total symptoms) were computed separately for the monogamous and polygamous groups. Fisher's Z transformation was used to compare the strength of these correlations between the two groups.
Results
Preliminary analyses
Preliminary exploratory analyses focused on testing for gender differences. There were no differences in the proportion of males and females having fathers with one, two, or more than two wives, χ2
, 
. In addition, there were no significant differences in mean scores for males and females on the family cohesion, SCL-90-R, or self-esteem scales. Among the ACHIEVE scales, boys scored higher than girls on “Violence in school” and “Violence in neighbourhood”, 
, 
and 
, 
, respectively.
Perceptions of family cohesion and sources of violence
Table 1 provides the means and standard deviations for each variable within each family structure group (monogamous or polygamous). The polygamous group was bifurcated further according to the number of wives (two or three and four) reported. Although adolescents from monogamous families displayed significantly higher cohesion scores than their peers coming from polygamous families, the mean score for the group with 3–4 mothers was not significantly different from the normative mean of 6.36, 
, 
Indeed, across all participants as a whole, perceived family cohesion was found to be significantly higher than the mean published for the normative sample, 
, 
. This comparison indicated that the group means indicated generally above average levels of cohesiveness by American standards. The monogamous group reported more violence in their schools and overall than those who had two mothers.
| Variable | Mono | Poly | t Value | Types of polygamous | ||
|---|---|---|---|---|---|---|
| 2 Mothers | 3–4 Mothers | F valuea | ||||
| N | 85 | 114 | 76 | 38 | ||
| FES cohesion | 7.3 | 6.6 | 2.95⁎⁎ | 6.8 | 6.2 | 5.07⁎⁎,b |
| 1.5 | 2.0 | 1.8 | 2.2 | |||
| Violence | ||||||
| In home | 1.3 | 1.2 | 0.29 | 1.1 | 1.4 | 2.68 |
| 0.71 | 0.60 | 0.55 | 0.66 | |||
| In school | 1.5 | 1.3 | 3.14⁎⁎ | 1.2 | 1.4 | 5.85⁎⁎,c |
| 0.64 | 0.54 | 0.51 | 0.58 | |||
| In neighborhood | 1.4 | 1.2 | 1.85 | 1.2 | 1.3 | 2.38 |
| 0.75 | 0.62 | 0.58 | 0.70 | |||
| Total | 1.4 | 1.2 | 2.00⁎ | 1.2 | 1.4 | 3.75⁎,c |
| 0.62 | 0.49 | 0.45 | 0.55 | |||
| SCL-90 | ||||||
| Anxiety | 59.6 | 61.5 | 1.17 | 60.2 | 64.3 | 2.22 |
| 11.9 | 10.7 | 10.3 | 11.3 | |||
| Depression | 57.8 | 59.7 | 1.29 | 58.4 | 62.5 | 2.68 |
| 10.5 | 10.4 | 10.0 | 10.9 | |||
| Hostility | 54.4 | 54.7 | 0.18 | 53.1 | 58.2 | 2.70 |
| 10.6 | 11.1 | 10.6 | 11.4 | |||
| GSI | 59.1 | 61.2 | 1.26 | 59.6 | 64.6 | 3.09⁎,d |
| 11.1 | 11.3 | 10.5 | 12.5 | |||
| Self-esteem | 16.0 | 15.5 | 0.85 | 15.5 | 15.4 | 0.38 |
| 4.4 | 3.9 | 3.8 | 4.1 | |||
- a From ANOVA comparing monogamous vs. 2 mothers vs. 3 mothers.
- b Monogamous >3–4 mothers by Tukey post hoc.
- c Monogamous >2 mothers by Tukey post hoc.
- d Monogamous <3–4 mothers by Tukey post hoc.
- ⁎ p<.05;
- ⁎⁎ p<.01.
The effect sizes for the mean differences reported as significant in Table 1 were calculated with the use of Cohen's d, which is the standardized difference between means (i.e. the difference between means divided by the pooled standard deviation). To Cohen (1988), 
indicates a small effect, 
indicates a medium effect, and 
indicates a large effect. That convention of Cohen will be used in the interpretation of the effect sizes.
The effect size was in the medium to large range with 
for the comparison of the FES cohesion means of the monogamous group and the polygamous group with 3–4 parents. The effect size for violence in the schools for the comparison of the Monogamous group and the Polygamous group with 2 mothers was in the medium range, 
.
The remaining effect sizes for the mean differences reported as significant were small to approaching the medium range: (a) 
for the comparison of the FES cohesion means of the monogamous group and the polygamous group; (b) 
for the comparison of the violence in school means for the monogamous group and the polygamous group; (c) 
for the comparison of the total violence means for the monogamous group and the polygamous group, and (d) 
for the comparison of the total violence means for the monogamous group and the group with 2 mothers.
Psychological well-being
Table 1 also shows the means and standard deviations of the self-esteem and mental health variables. The SCL-90-R scores of the participants did not show significant differences between the mean scores for the groups in terms of anxiety, depression, or hostility levels. The groups did not differ significantly from each other in terms of self-esteem scores, and their mean scores (15.4–16.0) were found to be comparable to those obtained for Arab-Palestinian children (see Baker, 1992).
Students from monogamous families also did not differ from students coming from polygamous families on overall psychopathology (GSI), although the monogamous group scored significantly lower on the GSI index than their counterparts coming from homes in which the father is married to three or four wives. The effect size resulting from a comparison of the GSI means of the monogamous group and the polygamous group with 3–4 mothers was .48, a significant value that fell in the small to approaching the medium range.
Relations between environmental and mental health variables
Table 2 displays the zero-order correlation coefficients between environmental and mental health variables, separately for the monogamous and polygamous groups. For both groups, family cohesion was positively associated with self-esteem and negatively associated with depression. For the polygamous group only, family cohesion was also negatively correlated with anxiety, hostility, and the symptom index.
| Environment variable | Self-esteem | Anxiety | Depression | Hostility | GSI | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| M | P | M | P | M | P | M | P | M | P | |
| FES cohesion | .34⁎⁎ | .34⁎⁎ | −.08 | −.26⁎⁎ | −.23⁎ | −.21⁎ | −.11 | −.31⁎⁎ | −.20 | −.25⁎⁎ |
| ACHIEVE (Violence) | ||||||||||
| In home | −.14 | −.22⁎ | .19 | .28⁎⁎ | .17 | .34⁎⁎ | .22⁎ | .29⁎⁎ | .16 | .32⁎⁎ |
| In school | −.11 | −.06 | .27⁎ | .01 | .33⁎⁎ | .10 | .29⁎⁎ | .12 | .26⁎ | .08 |
| In neighborhood | −.15 | .02 | .25⁎ | −.04 | .29⁎⁎ | .04 | .32⁎⁎ | .06 | .29⁎ | .04 |
| Total | −.16 | −.10 | .27⁎ | .10 | .30⁎⁎ | .19⁎ | .31⁎⁎ | .19⁎ | .27⁎ | .18 |
| Number of siblings | −.09 | −.10 | .20 | .11 | .27⁎ | .00 | .25⁎ | .12 | .27⁎ | .04 |
| Father education | −.00 | −.16 | −.10 | .12 | −.07 | .14 | −.19 | .10 | −.17 | .15 |
| Mother education | .09 | −.02 | −.04 | .04 | −.06 | .14 | −.15 | .03 | −.16 | .09 |
- Note: M=Monogamous. P=Polygamous. Boldface indicates a pair of correlations that are significantly different from each other, following Fisher's Z transformation.
- ⁎ p<.05;
- ⁎⁎ p<.01.
Violence in the home was also negatively correlated with self-esteem and positively correlated with all psychopathological scales, but for the polygamous group only. For the monogamous group, violence in the home was significantly correlated only with hostility. Violence within the two other contexts (school and neighbourhood) was not significantly associated with self-esteem in either group, but was significantly correlated with anxiety, depression, hostility, and GSI scores for the monogamous group only. In both groups, total violence was correlated with depression and hostility; in the monogamous group, it was also correlated with anxiety and total symptoms. Parental educational level was not significantly correlated with self-esteem or reported psychopathological symptoms in either group; however, family size (number of siblings) was positively correlated with depression, hostility, and GSI scores in the monogamous group only.
Correlational comparisons
As Table 2 indicates, there were a number of instances in which the strength of the association between two variables differed significantly between the monogamous and polygamous groups. The correlations of violence in school with anxiety and depression, for example, were significantly stronger within the monogamous group than the polygamous group, where they did not significantly differ from zero. Similarly, the correlations of violence in the neighbourhood with all four SCL-90-R scales were significantly stronger and greater than zero within the monogamous group, but did not differ from zero in the polygamous group. The correlation between number of siblings and depression showed the same pattern—positive in the monogamous group but significantly weaker, and zero in the polygamous group. Finally, although none of the correlations involving father education were different from zero, the correlations of father education with hostility and GSI scores did differ significantly between monogamous and polygamous groups, being negative in the former and positive in the latter group.
Discussion
Results from Study 1 showed few differences between adolescents from monogamous and polygamous families in perceptions of their family and community environments and their psychological well-being. Exceptions to this generalization include a higher level of family cohesion and a lower level of psychopathological symptoms reported by adolescents from monogamous families than by adolescents from polygamous families. Adolescents from families with two mothers did not show any differences from those whose families had more than two mothers.
One of the limitations of Study 1 is the reliance on self-reports. To avoid problems associated with relying only on participants’ own self reports (Achenbach & McConaughy, 1997; Edelbrock, Costello, Dulcan, Kalas, & Conover, 1985), particularly with regards to reporting internalizing problems, we obtained data from their teachers in Study 2.
Study 2
Method
Participants
In Study 2, the sample consisted of 182 Bedouin Arab adolescents, age 13.5–18.5 years (
, s.d.=1.0), from the Negev region of Israel. Of these, 114 were from monogamous families and 68 were from polygamous families. In the polygamous group, 50 had fathers with two wives, and 18 had fathers with three or four wives. Approximately 47% of the sample was male.
The students were randomly selected from six Bedouin schools in villages such as Raha, Tel-Sheva, and Laqia. Using the Statistical Package for Social Sciences (SPSS; SPSS Inc., 2001) software (by the first author), 210 subjects were randomly chosen from a list of 300 subjects that was provided by seven school principals in Bedouin schools. The seven schools were also selected randomly from a list of the 17 schools that serve the Bedouin school population. The original list included students from polygamous (300) and non-polygamous (340) family structures. Furthermore, the original sample consisted of students in grades 7–12, but when administering the test measure, students in grades 7–9 were less represented as a result of a school outing. Therefore, this study includes more representation from students in grade 10.
The first author collected the data from this sample four months after collecting the data for Study 1. Participants completed a single instrument about themselves and the teachers of the participants completed the Teacher Report Form of the Child Behaviour Checklist (CBCL; Achenbach, 1991).
Instruments
TRF: Teachers’ Report Form of the Child Behaviour Checklist (Achenbach, 1991). This instrument was translated into Arabic, and standardized for use with the Negev Bedouin Arab by Auerbach, Goldstein, and Elbedour (2000). This questionnaire has questions about the academic performance and adaptive functioning of a child and provides a list of 118 specific emotional and behavioural problems. For each item, the teacher was to score on a 3-point scale: (a) 0—not true; (b) 1—sometimes or somewhat true; and (c) 2—often or very true. The TRF included the following scales: withdrawn, somatic complaints, anxious/depressed, social problems, thought problems, attention problems, and aggressive behaviour. Internal consistencies (Cronbach's alphas) obtained in the current sample ranged from .58 to .80.
WITF: What I Think and Feel (Reynolds & Richmond, 1978,1979). This self-report questionnaire has 28 anxiety items and has a total anxiety scale. Reynolds and Richmond (1978) generated this test as a revision of the Children's Manifest Anxiety Scale (CMAS) (Castenada, McCandless, & Palermo, 1956) and renamed their revision What I Think and Feel. Reynolds and Richmond (1979) analysed WITF item data and concluded that the WITF has construct validity. Back-translation procedures similar to those used for the ACHIEVE and FES scales were used in generating an Arabic version of this questionnaire. The internal consistency (alpha) obtained in the current sample was .91.
Data analysis
In parallel with Study 1, t-tests and ANOVAs were used to compare groups of students from families with different structures. The first comparisons made were between students coming from monogamous and polygamous families as a whole. The monogamous group was then compared with two groups of polygamous families (two mothers versus three and four mothers).
Results
As in Study 1, there were no differences in the proportion of males and females having fathers with one, two, or more than two wives, χ2
, 
. In comparisons of mean scores, girls had higher levels than boys of anxiety/depression, 
, 
; social problems, 
, 
; somatic complaints, 
, 
; withdrawal, 
, 
; and WITF anxiety, 
, 
.
To test whether adolescents from monogamous and polygamous families differed in scale means, t-tests and ANOVAs were performed. On none of the 8 scales (7 TRF scales and 1 WITF scale) were there significant differences between those from monogamous and polygamous families. When groups with fathers with one, two, or three or more wives were compared, there were also no significant differences among the three groups. Means and standard deviations for each group on each scale are provided in Table 3.
| Variable | Monogamous | Polygamous | Types of polygamous | |
|---|---|---|---|---|
| 2 Mothers | 3–4 Mothers | |||
| N | 114 | 68 | 50 | 18 |
| CBCL-TRF | ||||
| Aggressive | 14.9 | 14.6 | 14.7 | 14.2 |
| 5.9 | 6.3 | 6.2 | 6.8 | |
| Anxious/depressed | 13.6 | 13.9 | 13.2 | 15.9 |
| 6.3 | 7.0 | 6.7 | 7.3 | |
| Attention problems | 15.2 | 15.0 | 14.8 | 15.6 |
| 5.0 | 5.7 | 5.8 | 5.6 | |
| Social problems | 7.0 | 7.6 | 7.3 | 8.2 |
| 4.3 | 4.7 | 4.5 | 5.3 | |
| Somatic complaints | 4.5 | 4.6 | 4.6 | 4.6 |
| 3.1 | 3.6 | 3.7 | 3.2 | |
| Thought problems | 5.1 | 4.9 | 5.0 | 4.8 |
| 3.2 | 3.7 | 3.3 | 4.8 | |
| Withdrawn | 7.9 | 8.5 | 8.1 | 9.4 |
| 3.0 | 3.1 | 3.1 | 2.9 | |
| WITF anxiety | 2.5 | 2.5 | 2.4 | 2.7 |
| 0.72 | 0.77 | 0.70 | 0.92 | |
- Note. There were no significant differences at p<.05, by t-test for monogamous–polygamous comparisons, or by Tukey post hoc comparison following ANOVA for Number of mothers comparisons. CBCL-TRF=Child Behavior Checklist—Teacher Report Form. WITF=What I Think and Feel.
General discussion
The primary goal of the reported studies was to compare the adolescents from monogamous and polygamous households on a number of mental health variables (e.g. level of distress and self esteem) to determine if the adolescents coming from monogamous families differed significantly from their cohorts coming from polygamous families. The data were collected from Bedouin Arab adolescents living in the southern region of Israel (Negev). The studies provide the opportunity to examine the impact of family structure on child outcomes that is not possible in the West. Their family milieus were assessed according to levels of conflict, expressiveness, and control. Furthermore, the levels of violence to which the participants were subjected in the home, school, and community were also measured. There were two studies examining the relationships between family structure and the mental health of Bedouin Arab adolescents.
The findings from Study 1 demonstrate that the two groups did not differ significantly on most of the assessed variables, although there were differences obtained between groups on some of the measures. Specifically, no differences were found between adolescents coming from polygamous or monogamous families in terms of anxiety, depression, hostility, or GSI, although children coming from homes where the father is married to three or four wives appear to display more psychopathological symptoms (GSI) than their monogamous counterparts. Furthermore, adolescents from monogamous families reported more family cohesion than their peers from polygamous families, but on the whole the sample did not deviate from available norms in terms of family cohesion. The levels of stress (GSI scores) reported for the sample also did not deviate considerably from that shown in previous research conducted on this population (see, e.g. Elbedour, 1998; Elbedour, Van Slyck, & Stern, 1988).
There were also some interesting findings from the correlational and other statistical analyses done by analysing the monogamous and polygamous groups separately. Family cohesion has more positive effects on mental health for adolescents from polygamous families than those from monogamous families. Also, violence in the school and neighbourhood has more of a negative effect on the mental health of the monogamous group than on the polygamous group, where it has no effect. In addition, having more siblings has a negative effect on mental health for the monogamous group but not for the polygamous group. This may be due to more potential caregivers being available in polygamous families than in monogamous families to provide psychological support for children in the families.
The findings from Study 2 demonstrate that the two groups did not differ significantly on teacher-perceived student mental health problems and on a measure of anxiety. The findings from Study 2 tended to support the findings from Study 1. Monogamous and polygamous families in the Bedouin Arab community were far more similar than they were different. In addition, among polygamous families, adolescents in families with two mothers tended to be quite psychologically similar to adolescents in families with more than two mothers. The only differences between the two groups were that adolescents from monogamous families reported, on the average, a higher level of family cohesion and a lower level of psychopathological symptoms than adolescents from polygamous families.
The sampled population comes from an ethnic minority (Bedouin/Arab) in Israel that does not enjoy economic, social, or political equity (see, e.g. Auerbach et al., 2000), and participants’ elevated GSI scores may be a reflection of this inequity. However, adolescents coming from large polygamous families (3–4 wives) could be more vulnerable to this stress. It may very well be that large family constellations are more susceptible to daily stress than families where the father is married to one or two wives. The overall picture of only a few significant differences observed in this study between children from monogamous and polygamous families warrants some explanation, given that the results obtained here do not corroborate previous research (Al-Krenawi and Lightman, 2000; Cherian, 1993, 1990), which reported, for example, higher levels of aggression, conflict, and hostility among polygamous families when compared to monogamous families and which both involved younger participants than the studies being presently reported.
It is likely that developmental factors are one explanation for the difference between the results in these studies and the findings of previous research. Given that the participants in these studies were of adolescent age at the time the studies were conducted (mean age 15.9 and 15.8), their developed perception, internalization, and processing of their family environments could have accounted for the incongruent results.
Thus it becomes essential to consider the potential role of developmental factors in the course of adaptation of children and adolescents to the polygamous marital structure. After all, adolescents are better equipped with internal and external resources to cope with family disputes, conflicts, and stress than children. In addition, the study by Al-Krenawi and Lightman (2000) of children's problems relied heavily on only one method of assessment. Their data source drawn from one method engendered a limited agreement among parents, children, and their teachers in the assessment of child problems and competencies and was more vulnerable to design problems and misinformation (Young, O’Brien, Gutterman, & Cohen, 1987). The present studies use data gathered from two classes of informants: adolescents and their teachers. This strategy provides readers and researchers with an opportunity to compare adolescents and their parental support for the same problems and symptoms or complaints particularly if the assessment instruments are equivalent.
Also, the Al-Krenawi and Lightman (2000) study on the same Bedouin community relied almost on younger children, who may still have been strongly impacted by their familial rearing environment. As indicated, earlier developmental determinants may be an important moderator of family environmental influence on children, with studies of younger samples showing higher health and emotional problems. It is clear that child reactions to families are shaped by both shifts in cognitive abilities and the contextualized realities of these children (Elbedour, Bart et al., 2003). Compared to younger children, older children are less likely to experience the adverse impact of family environments, because the older children know how to separate parental family problems from parent–child problems. In addition, because of the collectivity of the Bedouin community, older children are more likely to form multiple attachment relationships outside the home and perceive the extended family as an extension of their families and that care for children is distributed among all members of the extended clan. Community responsiveness (actual or perceived) helps older children to cope better with family issues and plays a major role in the expression of the development and resiliency of older children.
The relative influences of home environment on the development of children are not only influenced by the age variable, but also by a variety of other factors may be involved, especially social support systems that sometimes offset or complement care and protection received in the home (Elbedour, Onwuegbuzie, & Alatamin, 2003). As children develop, they are less confined to the household and more likely to spend time with peers and other relatives in the Bedouin extended clan and less time with their parents; thus, they are less likely to experience or witness family conflict. For some, forming relationships with peers and other adults in this collective community, who provide care and protection for children, can compensate for the less than optimal home environment. Advances in the child's level of cognitive development plays an important role as well. As adolescence approaches, children are more likely to possess a more sophisticated repertoire of coping media to manage family stressors (Elbedour et al., 2000). Taken together, developmental factors related to the kinds of relationships that children develop within the variety of social networks outside the family home need to be considered in order to fully understand the development of children in polygamous families.
A second major explanatory factor for the incongruent results is social context. If family structure (polygamous or monogamous) is stable and follows a pattern that is valued by society at large, then the adverse effects of conflict within it may be mitigated. In the studied Bedouin community, polygamy is associated with high social status; hence, children of polygamous families are not perceived as endangered by this accepted and valued practice.
Third, a third explanatory factor for the incongruent results is social group structure. In Bedouin Arab communities, the clan features a high level of group cohesiveness. Much importance is placed on the group cohesiveness within Bedouin Arab communities where group values determine psychological outcomes. The extent to which a family, clan, or tribe can protect and consolidate itself socially and economically is predicated on the size of its “izweh” (number of males it has). Consequently, polygamy is valued rather than stigmatized in Bedouin communities. It is perceived as a means to propagate a large number of children. In Bedouin Arab communities, polygamous families may be as socially acceptable as monogamous families. A lack of differences between monogamous and polygamous families in Bedouin Arab communities may be a reflection of such social acceptability.
As illustrated by Elbedour and his colleagues (e.g. Elbedour et al., 1997, 2002), values placed on polygamy play a substantial role in the course of children's development. These core values may vary from culture to culture and the discrepancy in findings in previous results and the present studies may, in part, be explained by the variation in core values placed on polygamy among different cultural groups. In the Bedouin Arab culture, polygamy is a natural expansion of other family structures and carries no social stigma. In fact, social status is highly linked with larger families, and it is clear that polygamy is one vehicle to achieve higher social position.
In addition, child-rearing practices have a collective orientation in the Palestinian Arab community in general (see Dwairy, 1997, 1999) and in the Bedouin community in particular (Elbedour, Shulman, & Kedem, 1997). Also, as in collective communities in Asia, Africa, and South America (Markus & Kitayama, 1991), emphasis is placed on values such as cooperation, social expectations, similarities, group attachment, and/or identification and independence with other in-group members (Hofstede, 1980, Triandis, 2001, 2002, Triandis, Bontempo, Villareal, Asai, & Lucca, 1988). Individuals in these communities are concerned about their communities, families, and group identity. Members suppress their personal goals and their behaviour is guided by their respect, commitment, and conformity to the way of life of the tribe and the extended family and ideals that their traditional culture or religion enforce.
According to Brown et al. (1992), “It is for collectivists that the group—and the group's outcomes—really matter.” (p. 329). Such collective orientation may affect not only behaviour but also the course and nature of development as well. In individualistic communities, there is less orientation towards group concerns and more emphasis on separateness and personal goals and needs (Singelis, 1994).
The Bedouin community emphasizes the centrality of the clan (i.e. extended family) and the social identity of the clan. It is the expectation of the clan that the members are motivated to expand and include the others in their lives. School-aged children and adolescents are aware of the social status and ranking of each clan and relate to each other in school settings according to these memberships. Thus, self-evaluation reflects the knowledge that the individual has a certain clan membership and shares a common group identity with the clan. Simply put, self-esteem is derived from the status of the clan, its resources, its size, and its achievements and reflects the comparison of the status of a particular clan and those of other clans.
Furthermore, the concept of “family” within the Bedouin community reaches far beyond the nuclear grouping. The extended family is a salient feature of such communities, and there are no clear-cut demarcation lines separating the family (nuclear and extended) from the larger community. The community prizes its collective nature where conformity, consensus, and collectivity are highly valued, and similarity and control are key components of relationships among kin and between friends (Elbedour et al., 1997).
Bedouin children from polygamous and monogamous families are members of the same culture: they share the same leisure and non-leisure activities in that they attend the same tribal school and mosque, and play the same games (Elbedour et al., 2000; Elbedour, Bart et al., 2003). They are often related to each other, and parents of both groups share the same culture and child-rearing patterns. They view the world in similar ways, and their attitudes and outlook on life are alike. Children go to their homes primarily for the purposes of eating and sleeping, and mothers are not surprised or perturbed to learn that their children have slept in another home.
In such a collective community, the boundaries between different families are less salient, and there is little distinction between families within a given clan. Hence, the children in this study may be less vulnerable to the negative effects of marital conflict, because they are less confined to their homes than their Western counterparts (Harris, 1995). Furthermore, the sense of group connectedness and “collective” self-esteem appear to provide support to all members within the clan. Although the children in this study live in different family structures, they share the same environment, and this overall similarity may be one reason for the similar level of psychological adjustment found in the groups. Consequently, resilience among subjects from polygamous family structures may reside more in the cultural and social tribal system than in the subject alone. The monogamous/polygamous family structure distinction may be of little value in interpreting adolescent behaviour in some collective societies.
In addition, there is growing evidence to suggest that psychological concepts formulated on Euro-American societies may not be applicable to collective societies in Africa and Asia (Al-Issa, 1995; Dwairy, 1998; Ibrahim & Ibrahim, 1993). The obtained results in this study may be “incongruent” only because they were interpreted within the wrong context. Compared to individualistic Western communities, less differentiation exists within the studied population between individuals and their collective group in terms of values and interests.
In a collective society such as that of the Bedouin Arab community in the Negev region, polygamous family structures seem to offer some advantages over monogamous family structures in supporting adolescent mental health. Although violence in the school and neighbourhood and the number of siblings had no negative effects on adolescent mental health in polygamous families, those variables did have negative effects on adolescent mental health in monogamous families. Also, family cohesion has positive effects on adolescent mental health in polygamous families, but family cohesion has no positive effects on adolescent mental health in monogamous families. Such findings would less likely in individualistic societies.
Two problems remain in these studies. The failure to find many overall differences could plausibly be attributed to the insensitivity of the instruments rather than to the lack of differences in the mental health and self-esteem of the children from the two groups. Second, failure to obtain differences could be due, in part, to lack of statistical power. In this case, the effect sizes of our significant findings were in the moderate range. There was not enough statistical power to detect differences that may have been smaller than that. Third, there may be a tendency in such studies to report significant differences and but not null results. As a result, monogamous and polygamous groups may be viewed as more discrepant than they actually are. Readers and researchers should take these essential concerns into consideration.
Acknowledgements
The authors wish to express their appreciation for the cooperation of all participants in this inquiry and for the support provided for this inquiry by school authorities in the Negev region of Israel, the Ben-Gurion University of the Negev, Howard University, the University of Minnesota, and North Dakota State University. The authors also wish to express their appreciation to the editor, Dr. Alan Waterman, and two reviewers for their constructive and incisive comments that contributed to improvements in this research report.
