Review of CABS-PA-J reliability and validity
In study 1, we carried out a back translation when developing the scales to verify the validity of the contents. We also obtained the permission of the author of CAPS-PA, Petretic-Jackson, to modify scales that were adjusted to the Japanese population by adding new items, taking cultural differences into account.
Exploratory factor analysis produced four factors theoretically based on the criterion that the factor loading was above .35. Although these four factors differ slightly from the items in the original CABS-PA, their contents were essentially the same. However, the differences found from previous studies are believed to have been produced by differences in cultural background. In 2002, the number of calls made to child consultation centers around Japan asking for child abuse counseling increased to 24,000, a two-fold increase from 1999, and by 2004 had risen to 33,408, up 25.7% from 2003 [
17,
18] . It has only been over the last couple of years that Japanese society has actually started to demonstrate concerns over child abuse, which may explain the relative immature awareness of the Japanese towards abuse. Unlike the U.S., where countless studies on the subject have been conducted in the more than 100 years since the report of the 1874 Mary Ellen incident [
19] and the study by Kempe [
20] "Battered-Child Syndrome," the problem of child abuse is not adequately recognized in this country. Studies of child abuse awareness in Japan among professionals and general office employees have pointed out that adult persons with general office jobs tend to have less awareness of psychological abuse compared to those working in the areas of education and childcare, that awareness of abuse in the country is diverse, and that raising common awareness and understanding involves numerous difficulties [
21‐
24]. Next, results of reviewing alpha reliability coefficient and test-retest reliability of the scale confirmed that CABS-PA-J is generally reliable. The child victim and societal factors were also found to serve as independent scales due to their high reliability.
In confirmatory factor analysis, the oblique factor model was found to be the most suited to CABS-PA-J of the three models. In addition, GFI, AGFI, and RMSEA, the goodness-of-fit indices of the oblique factor model, also indicated sufficient compatibility with CABS-PA-J. In the present studies, GFI was .89, indicating sufficient compatibility. RMSEA below .050 indicates models with sufficient goodness of fit, and RMSEA below .080 was also taken to be suitable, therefore, .050 was considered sufficient [
25]. Consequently, the results of the analysis of confirmatory factors in the present studies verified that goodness-of-fit indices were excellent, and that structural concept validity was also sufficiently appropriate.
The results of criterion-related validity examination showed a significant negative correlation comparing the CABS-PA-J subscale child victim with social support by spouses and by others. In addition, the subscale was found to have a significant trend with the physical abuse score, significant positive correlation with psychological abuse score, as well as significant correlation with the child abuse behavior checklist total score. The total CABS-PA-J score, on the other hand, had a significant negative correlation with social support by spouse, positive correlation with the abuse behavior checklist subscale "psychological abuse," and a significant positive correlation with the child abuse behavior checklist total score. The four CABS-PA-J subscales were weak in relation to other scales in existence, because the concepts of existing scales were for other parameters.
Consequently, the criterion-related validity of CABS-PA-J has been demonstrated to a certain extent. These results also suggest, as pointed out by the belief theory proposed by Wright
et al. [
26] and theories of cognitive behavioral therapy [
27], that cognitive variables play an important role in factors influencing interactions between people, which may be evident theoretically. The results of the present studies show that CABS-PA-J is an appropriate scale for measuring the mother's awareness and belief of abuse, and suggest that, as a result of that awareness, a cognitive variable may be involved in abusive acts towards their children.
However, such awareness (belief) can change according to the mother's situation [
28]. Aramaki demonstrated that if social support by the husband is sought but not acquired by the mother, her positive feelings towards their child are the weakest [
29]. The results of the present studies also confirmed a significant negative correlation between the CABS-PA-J subscale and high social support, as well as between the subscale and the child victim factor. This suggests the presence of a process where the stressful situation of not receiving social support interacts with awareness of abuse, resulting in abusive acts.
Although the original scale targets awareness towards physical abuse, the CABS-PA-J showed a higher association with the psychological than the physical abuse scores. These results should be interpreted taking into consideration the following points: the present studies were conducted of mothers who took their children to our hospital; and, since there were no mothers who were actually abusing their children physically, CABS-PA-J may have demonstrated an association with psychological abuse. However, the premise of this study involved a model whereby a one-dimensional correlation exists for abusive parents as well as those that do not abuse their children. In addition, as this was primary research aimed at developing a scale, the subjects weren't limited to those parents who committed child abuse. Consequently, we consider the results of this study to be valid.
The results of a review of the criterion-related validity found that the CABS-PA-J subscale is associated with social support and abusive acts, and that the CABS-PA-J total score shows the same tendency. Therefore criterion-related validity is thought to have been demonstrated to a certain extent.
Characteristics found by CABS-PA-J: Mothers of low-birth weight children up to age 3 (clinical group) vs. Mothers of standard-birth weight children (standard Group)
The purpose of Study 2 was to investigate whether or not CABS-PA- J can be clinically applied. Mothers of low birth weight children up to age 3 (clinical group) and those of standard birth weight children (standard group) were compared using the CABS-PA-J subscales. The results revealed that mothers of children over one year had higher preserver-perpetrator scores than those of children less than one year. The preserver-perpetrator is thought to be a factor indicating guilty feelings by the parent based on the view that the parent is responsible for abuse, with the score being higher at about age one [
30], when attachment to infants becomes stable. In general parent-child relationships, therefore, if child abuse occurs when the parent's attachment to the child is established, then the mother usually recognizes that she is wrong.
ANOVA of the two factors, age (< 2 years and ≧ 2 years) and birth weight, found the main effects of age in term of the CABS-PA-J child victim. The child victim score increased in mothers of infants aged over two years. When a child reaches age two, his/her range of actions broadens, and the mother experiences more stress and anxiety in bringing up her child, which may be why the child victim score increases [
31,
32].
Essentially, ANOVA of the two factors, age (< 3 years and ≧ 3 years) and birth weight found the only a trend in terms of the CABS-PA-J Preserver-perpetrator.
In addition, correlation analysis comparing CABS-PA-J subscales with respective social support, family income, mother's age, and mother's education level demonstrated a difference in the characteristics of the two groups: mothers of low-birth weight children up to age three (clinical group) and those of standard-birth weight children (standard group).
Mothers in the clinical group showed a significant negative correlation between the CABS-PA-J preserver-perpetrator subscale and support by others and a significant positive correlation between the subscale and mother's age. This indicates that the higher the mother's age, the higher the preserver-perpetrator tendency.
On the other hand, mothers in the standard group showed a significant negative correlation between the child victim subscale score and any social support by spouse or support by others. In other words, social support to mothers in the standard group showed a significant negative correlation with child victims. This consequently suggests that increasing social support may help reduce abusive acts in the standard group. There was also a significant positive correlation between the situational factor and the mother's education level. Situational factors, attributing child abuse to the situation, consist of items including various environmental stressors to the parents [
6]. Subsequently, it shows the possibility that even mothers in the standard group, who have no concerns about the physical development of their child, may feel environmental stress more easily if they have a higher educational background. Moreover, mothers in the standard group showed significant correlations with the preserver- perpetrator and mother's education level; specifically, mothers with high educational background are more likely to think that they are responsible for child abuse, which serves as a strong control over child abuse.
In this way, the results of calculating the correlation between CABS-PA-J subscales and respective social support, family incomes, mother's age, and mother's education level between the clinical and standard groups indicated characteristic differences in the correlation pattern by parameter. Mothers from the clinical group showed a strong tendency to have guilty feelings [
33,
34]. Since it has been pointed out that giving birth during middle age poses as an environmental factor of low-birth weight children [
35], mothers will tend to feel a greater sense of guilt towards their child, represented by the preserver-perpetrator scale, and as age increases. However, since the subscale showed a significant negative correlation with social support by others, increased social support by others may help reduce preserver-perpetrator scores in such cases.
As a result of a review of the differences in correlation patterns between the two groups, awareness of child abuse was found to have similar progress as child development between the two groups. On the other hand, the significance of their awareness may be dissimilar. In other words, even if no substantial difference is seen in development by age, the mother's feeling of guilt will increase and she may develop depression as a result if her age is high, as in the present studies, and the child is born at a low birth weight. Breslau and Davis [
36] reported that 30% of middle-aged mothers who gave birth to babies with low birth weight suffer serious depression. In addition, their daily stressor was associated with a lack of social resources. These can be predictors of both the severe state of depression in mothers as well as their potential for child abuse [
37]. A number of studies have shown that most abusive mothers lack social support as well as social resources and suffer from severe depression [
37‐
40]. Therefore, the results of the present studies suggest that mothers of low-birth weight children have higher risks of developing depression, and thus higher risks of abusing their children.
Singer
et al. [
41] reported that mothers of extremely-low birth weight children who do not have high risks experience the same psychological pain as mothers of term infants at the point their children become two years old. Even though the parenting stress of mothers of high-risk, low-birth weight children continues to increase, several psychological adaptation scales show the same scores as mothers of term infants when their children become three years of age. On the other hand, mothers of high-risk, low-birth weight children show positive adaptation on the whole at three years, yet they are still reported to suffer from five times more serious psychological pain than mothers of term infants; almost one-third of mothers of ultra-low birth weight children are reported to suffer depression and anxiety at the clinical level. The results of this study also suggest that the older the mothers of low birth weight children, the greater their guilt feeling will be towards their children, resulting in psychological pain.
Since social support by others correlated negatively with the preserver-perpetrator phenomenon, it also suggests that complete support may be important to such mothers, indicating the importance of support by professionals to such children and their families.
Limitations of this study and the clinical use of CABS-PA-J in the future
The results of this study showed that CABS-PA-J has moderate validity and reliability as a scale for measuring awareness towards child abuse. It also successfully clarified several cognitive characteristics of mothers of the two study groups. Because the subjects of the present studies were mothers and children who visited hospitals, however, the results may include sampling bias instead of actual abuse. The potential of a difference in awareness towards abuse between the two groups of mothers was also hinted at. In the future, to enable use of this scale in the actual clinical setting, there will be a need to continue studies of mothers who are actually abusing their children, as well as studies of association with other psychological processes.
On the other hand, the results suggested an association between awareness of child abuse and actual acts. The combination of the CABS-PA-J scale and other scales for measuring environmental stressors of parents may help predict the risks of child abuse without having to actually ask mothers directly if they are abusing their children, which may help prevent child abuse from an early stage. Moreover, there is the potential to create educational programs for parents who commit child abuse, using CABS-PA-J to specify the educational material needed for those parents and to create individual training programs for them.
Furthermore, the use of the scale for mothers receiving therapy for child abuse may help determine changes in attitude and assess therapeutic effects. In order to enhance clinical efficacy in the future, further studies using the CABS-PA-J should be carried out.