- Email: [email protected]
Child abuse potential and work satisfaction in day-care employees
ChrldAhrr\c A f+c/c
Copyright
0145.2134/x7 s3.w * .(X) c 19x7 Pergamon Journals Ltd
CHILD ABUSE POTENTIAL AND WORK SATISFACTION IN DAY-CARE EMPLOYEES DONALD
W. ATTEN,
M. ED. AND JOEL
S. MILNER,
PH.D.
Western Carolina University, Cullowhee. NC Abstract-The Child Abuse Potential (CAP) Inventory and a Work Satisfaction Questionnaire were administered to 228 day-care employees at 40 centers to investigate the relationship between potential for child abuse and degree of job satisfaction. In addition. the ability of the CAP abuse scale to distinguish between groups of day-care employees thought to differ in the quality of child care provided was studied. Small. albeit significant, inverse correlations were observed between abuse scores and two job satisfaction items. As clarity of job expectations and as overall job satisfaction increased. abuse scores decreased. Although abuse scores for all day-care employees were not significantly different from norm scores. significantly lower scores were found for employees from “superior” day-care centers and for employees rated “high” by center directors. Abuse scores for employees rated “high” were also significantly below scores for employees rated “low.” While these data indicate only limited relationships between child abuse and job satisfaction. the findings suggest the CAP abuse scale can distinguish some subgroups of day-care employees thought to differ in quality of child care provided. Resume-Linventaire de potentiel de s&ices a I’egard d’enfants (CAP ou Child Abuse Potential Inventory) ainsi qu’un questionnaire portant sur la satisfaction au travail ont ete soumis a 228 employ& de creches dans 40 centres. Le but Ctait de voir quelles relations on pouvait etablir entre le potentiel de s&ices a I’egard d’enfants et le degre de satisfaction apporte par I’emploi. En outre la capacite de I’echelle CAP de distinguer entre differents groupes d’employes et de creches a ete etudie. Ces groupes etaient form& de personnes qui semblaient se differencier entre elles par la qualite des soins qu’elles apportaient aux enfants. Des differences significatives bien que petites ont ete observees et une relation inverse est apparue entre le potentiel de s&ices et la satisfaction apportee par I’emploi. En d’autres termes. lorsque la gratification que le personnel ressentait de ce genre d’emploi apparaissait clairement et que consequemment, la satisfaction augmentait, les scores potentiels de s&ices diminuaient. On a trouve des scores de potentiel de s&ices plus bas lorsqu’on avait affaire a des employes de creches qualifies de bons employes par les directeurs des centres. Par contre lorsque les employes etaient moins apprecies des directeurs de centres les scores de s&ices potentiels etaient plus eleves. Les donnees recoltees indiquent done une relation limitte mais certaine entre la satisfaction apportte par I’emploi et le risque de s&ices; les observations suggerent surtout que I’inventaire CAP permet de distinguer des sousgroupes a I’interieur d’une cohorte d’employes de creches qui sont soupconnes de donner des soins de qualite inferieure aux enfants.
INTRODUCTION EMPLOYEE MALTREATMENT of children in day-care centers has recently received increased attention in the general media. While there are articles on the stresses and subsequent burnout of employees in the field of child care [l-4], a literature review [5] reveals few empirical studies on the putative indicators of child abuse in out-of-home child care facilities. One study [6] which attempted to investigate variables related to employee child abuse used the Child Abuse Potential (CAP) Inventory, work satisfaction items, and demographic variables to distinguish between known institutional child abusers and matched employee controls. Overall, the CAP abuse score, three work satisfaction items, and several demographic variables were found to correctly classify 93% of
Reprint requests to Joel S. Milner, Ph.D.. Department of Psychology, Western Carolina University, Cullowhee. NC 28723 II7
118
Donald W. Atten and Joel S. Milner
the abusive and nonabusive employees. This study, however, has not been replicated and no similar study has been conducted in day-care settings. Given the concetn of the general public and professionals in the field of day care and the paucity of empirical data regarding child abuse in day-care settings, the need to investigate variables related to employee child abuse remains a pressing issue. The present study was designed to investigate the relationship between the CAP Inventory and work satisfaction items in day-care employees. In order to provide preliminary data on the ability of the CAP Inventory to distinguish between levels of quality of child care. daycare employees were divided into subgroups including Level 1 day-care center employees, Level 2 day-care employees, “superior” day-care employees, and groups of day-care employees rated “high” and “low” in terms of relative quality of child care provided.
METHOD Subjects A total of 228 Level 1 and Level 2 day-care employees from 40 participating day-care sites in North Carolina volunteered to take part in the study. Eight workers from the 40 participating sites declined to take part in the study. Further, three directors/coordinators of additional sites declined to grant permission for their day-care staff to be contacted. After testing, 9 subjects withdrew from the study leaving 219 subjects for the study sample. All subjects were employed at Level 1 (n = 44) or Level 2 (n = 175)‘state subsidized daycare centers. To qualify for subsidy funds, a day-care center must meet minimum requirements in 16 standards areas defined by the North Carolina State Social Services Commission. Level 2 day-care centers must meet all minimum standards set for Level 1 centers, plus additional requirements in 8 of the 16 standards areas. Standards for Level 2 centers are stricter in the following areas: operational policies, personnel policies, square footage per child, number of toilets per child, staff-child ratios, qualifications of the director, number of age-appropriate children’s activities, and number of different activity areas. Requirements for child care staff and staff training are the same for both Level 1 and Level 2 centers. Still, the assumption remains that higher standards and increased subsidy for Level 2 centers means a better quality of care is provided at these sites. In addition to the identification of Level 1 and Level 2 day-care employees, a subgroup of Level 2 sites were identified as “superior” day-care sites by the North Carolina State Office of Day Care. All employees from the “superior” sites were included in a single “superior” subgroup for comparison purposes. Further, all day-care center directors identified employees they believed comprised the top (rated high) and bottom’ (rated low) 10% of their staff in terms of overall quality of child care. Demographic characteristics for each of the day-care employee groups and subgroups are presented in Table I.
Test instruments The CAP Inventory is a brief, client-administered 160-item screening device which is answered in a force-choice, agree-disagree format [7]. Classification data indicate overall classification rates for physical abusers and matched comparison parents are generally in the low 90s [8]. Cross-validation data on over 3.000 subjects from a variety of maltreating.
mean (SD)
number
of children
trith children)
74 62 46
53 36 57
% Married
Groups
(see percent
88 58 58
60 s7 61
% White
of employees
99 IO0 I 00
99 IO0 99
Q Female
for the Day Care Employee
for the subgroup
13.3 (2.3) 13.8 (2.1) 13.8 (2.0)
13.2 (2.0) 12.7 (1.7) 13.4 (2.01
x (SD) Education
Characteristics
for each group was determined
38.3 (11.8) 39.6 (10.7) 32.4 (I 1.8)
96 26 24
a The
34.8 (I2.0) 3O.6 112.0) 35.7 (Il.81
219 44 I75
All day care employees Level I day care employees Level 2 day care employees “Superior” day care center employees Rated “high” day care employees Rated “low” day care employees
Age
II
x (SIN
Demographic
Group
Table I.
reporting
71 69 54
59 48 62
% With Children
children.
I .5 (0.6)
2.1 (1.1) I .7 to.71
I.9rI.l) I.5 10.7) ?.O(l.l)
x (.SUl Children”
120
Donald W. Atten and Joel S. Milner
general population, and nurturing parent groups are available in a recent manual [7]. A longitudinal, predictive validity study [9] indicates that elevated abuse scores are significantly (p < .OOOl) related to later physical child abuse, marginally (p < .05) related to subsequent neglect, and not related 0, > .05) to later failure-to-thrive cases. Factor analytic studies [7] indicate that six dimensions appear to be descriptive of the CAP abuse scale. These factors include distress, rigidity, unhappiness, problems with child and self, problems with family, and problems from others. Numerous construct validity studies provide additional data on the characteristics measured by the abuse scale. For example, elevated abuse scores are related to a childhood history of abuse [lo-121; low self-esteem [lo]; poor ego strength [13]; external locus of control [14, 151; stress [16-191; isolation [17, 201; anxiety [19-211; immaturity, poor impulse control, aggression, assaultative behavior [19, 20, 22, 231; depression [19, 20, 241; neuroticism [21, 231; perception of a child with problems [15, 221; physiological liability to child stimuli [2.5]; and family conflict [lo, 261. Other research indicates the CAP abuse scale is sensitive to changes produced by intervention/treatment programs [27, 281. Reliability data on the internal consistency and temporal stability of the CAP inventory are available [7]. Internal consistency estimates have been published for control, at-risk, neglect, and abuse groups as a function of location, gender, age, educational level, and ethnic background. Temporal stability estimates are available for control subjects across four time intervals (i.e., l-day, l-week, l-month, and 3-month) and as a function of demographic variables. The abuse scale has internal consistency reliabilities of .92 to .96 for control subjects and .95 to .98 for abusive subjects. No demographic effects were reported. Temporal consistency estimates for the abuse scale are also adequate with .91, .90, .83, and .75 for l-day, l-week, l-month, and 3-month intervals, respectively. Again no demographic effects were found. A brief, 8-item Work Satisfaction Questionnaire was developed based on putative indicators of child abuse found in a comprehensive review of institutional abuse literature [5]. The eight items cover the following content areas: Item 1, quality of supervision; Item 2, relevance of training for job: Item 3, satisfaction with pay: Item 4, opportunity for career advancement; Item 5, ability to handle difficult children; Item 6. ability to handle job-related stress; Item 7, clear job performance expectations; and Item 8, overall job satisfaction. All items were answered on a .5-point Likert Scale. Procedure
The CAP Inventory and the Work Satisfaction Questionnaire were administered on a voluntary basis to day-care employees at 40 participating centers. Prior to completing the test materials, each subject was given an information sheet which included the name of the principal investigator, a toll-free phone number that could be called for additional information, an informed consent, and a self-addressed, prepaid, number-coded postcard which, if mailed at a later date, would remove the employee from the study. RESULTS Two hundred and twenty-eight day-care employees volunteered to take the CAP lnventory and Job Satisfaction Questionnaire. Nine postcards requesting withdrawal from the study were received and the corresponding test materials were destroyed. Of the 219 remaining employees, 6 had incomplete (i.e., more than 10% blanks) CAP Inventories and 61 had elevated lie scale scores. Since validity data show elevated lie scores are indicative of faking good behavior [7, 291 and demonstrate use of the lie scale reduces the number of
Child abuse and work satisfaction
121
abuse scale misclassi~cations [8], responses with elevated lie scores were considered questionable and removed. After exclusions, data from 152 subjects were available for analysis. The abuse scale internal consistency estimate (KR-20) for this group of day-care employees was .90. Only two of the eight work satisfaction items were found to have significant inverse correlations with abuse: clear job performance expectations and CAP abuse scores, r = .26, & = 150, p < .OOl; and overall job satisfaction and CAP abuse scores, r = - .14, df = 149, p c .Ol. The mean CAP abuse scores for each of the day-care groups are presented in Table 2. For comparison purposes a norm group of subjects (parents) are also presented in Table 2. An analysis of variance revealed no significant difference, F = .lS, df = l/986, p > .05, between the day-care employees and the norm group. Several additional analyses were performed on the data presented in *able 2, WhiIe higher than the norm group, the mean abuse score for Level 1 employees was not significantly elevated, F = I. 10, df = 11859, p > .05, above the norm group mean. While lower, the mean abuse score for Level 2 employees was not significantly below, F = .98, df = l/961, p > .0.5, the norm group mean. Further, a direct comparison of Level 1 and Level 2 group mean abuse scores did not reveal any significant difference, F = 1.19, df = l/150, p > .05.
In contrast, the “superior” center employees’ mean abuse score was significantly below, F = 4.92, df = l/906, p < .05, the mean norm abuse score. The rated “high” employees were also significantly below, F = 7.29, df = 11848, p < .OI, the mean norm abuse score. While the mean abuse score for the rated “low” employees was not significantly above the mean norm score (F = 1.30, df = 11852, p > .0.5), the mean norm abuse score for the rated “low” employees was signi~cantly above the mean abuse score for rated “high” employees (F = 4.42, df = 1130, p < .OS>. Inspection of the individual CAP abuse scores revealed that none of the rated “high” employees had scores above the abuse scale cutoff [7], while three of the rated “low” employees had abuse scores above the cutoff for physical abuse. For the work satisfaction items, education was positively correlated with relevance of training for the job, r = .27, df = 143, p -=c ,001; and employee’s age was positively correlated with satisfaction with pay, r = 23, @’ = 14.5,p < .Ol; and with ability to handle job related stress, Y = .22, df = 145, p < .Ol. The only demographic variable correlated with abuse was education, Y = - .23, df = 146, p < .OI. DISCUSSION Although the institutional abuse literature [5] is replete with references that indicate job-related factors are important in the occurrence of employee child abuse, the present Table 2.
Mean (SD) Abuse Scores for Norm Subjects and Day Care Employee Groups Group
Norm subjects* All day care employees Level 1 day care employees Level 2 day care employees “Superior” day care employees Rated “high” day care employees Rated “Low“ day care emplovees
n
B
SD
836 152 2s 127 72 14 18
91.0 88.6 106.8 as.2 75.6 67.1 113.9
75.0 62.0 74.0 58.8 54.7 31.7 84.6
* Data for norm subjects were taken from Milner (7).
122
Donald
W. Atten
and Joel S. Mimer
data revealed only two of eight job satisfaction items- clear job performance expectawere modestly related to a measure of child abuse tions and overall job satisfactionpotential in day-care employees. These data support the general findings of Haddock and McQueen [6] who found that only 3 of 20 job satisfaction items discriminated known institutional child abusers from matched comparison employees. Institutional child abusers were dissatisfied with their chances to work independently, with their opportunity for advancement, and with their feelings of accomplishment. While conclusions based on these two correlational studies must be viewed with caution, the findings do not support the view that job satisfaction variables are strongly related to employee child abuse. Overall, no differences were found between day-care employees’ CAP abuse scores and norm parent abuse scores. These data suggest that, as a group, the North Carolina day-care employees tested were not different in abuse potential from nonscreened, general population parents. While differences had been expected between Level 1 and Level 2 employee abuse scores, no differences were found. Differences had been expected because additional requirements are maintained in 8 of the 16 standards areas for Level 2 centers. However, Level 1 and Level 2 centers have the same standards for child care staff and staff training. Whether additional standards are unrelated to quality of child care or whether the CAP Inventory is insensitive to differences that do exist is not known. As expected, day-care employees from superior centers had abuse scores significantly below norm parent abuse scores. These centers had been recommended by the North Carolina State Offtce of Day Care as outstanding in quality of child care, in supervision of staff, and in stability of staff. Further, the day-care employees rated high by center directors had the lowest abuse scores and the smallest variance of any subgroup tested. The rated high employees were significantly below the rated low employees and the norm parent group. These data indicate coticordance between the CAP abuse scores and ratings by center directors based on their observational knowledge of the quality of child care provided by their employees. As previously noted, the present study is correlational in nature and the findings must be viewed as only suggestive. Another study limitation was that the sample of day-care subjects were considered normal and, as such, had a limited range of abuse scores. While this makes observed subgroup differences in CAP abuse scores more impressive, it limits the degree of correlations possible between the CAP abuse scale and the work satisfaction measures. Likewise, use of a normal sample provides no information on the ability of the CAP Inventory to distinguish employees who are abusive from those who are not. Nonetheless, these data do indicate additional research is warranted on the use of the CAP Inventory in the screening of day-care employees.
REFERENCES 1. FREUDENBERGER. H. J. Burn-out: Occupational hazard of the child care worker. Child 6:9O-99 (1977). 2. HYSON, M. C. Playing with children all day: Job stresses in early childhood education. 37:25-32 (1982). 3. MASLACH. C. and PINES, A. The burn-out syndrome in the day-care setting. Child Carr II3 (1977). 4. MATTINGLY. M. A. Sources of stress and burn-out in professional child care work. Child
CUI.P Q~~urtrr/~ Yolrn~ Cltildren
QIUI~IC~/.V6: IOOCure QrwrIer/y
6:127-137 (1977). 5. MILNER. Document.s
J. S. and ATTEN, D. W. Institutional child abuse and neglect: A bibliography. P.sw/to/o,qicrrl 15:22 (1985). 6. HADDOCK, M. D. and MCQUEEN. W. h4. Assessing employee potentials for abuse. Jorrrnd c;fC/inictr/ Ps~rhol0~y 39: I02 I - I029 t I983 1. 7. MILNER, J. S. 771~ Child Ahrrse Potatrtirrl /rt~~,r~rcwv: Mtrrrrd. (2nd ed.1. Psytec. Weh\ter. NC (1986).
Child abuse
and work
satisfaction
123
8. MILNER. J. S.. GOLD. K. G. and WIMBERLEY. R. C. Cross-validation of the Child Abuse Potential Inventory. Jorrrricrl c!f C‘r~ri.\rt//i~i~ trrid C/ipfi(.0/ Ps,vcho/o,~,v. (in press). 9. MILNER. J. S.. GOLD. R. G.. AYOUB. C. and JACEWITZ. M. M. Predictive validity of the Child Abuse Potential Inventory. Jorrrxcll 0/ Cofrsrrlrin~ rr/id C/inic,u/ P.r.vc~ho/oR~ 52:X79-884 ( 1984). IO. CHAN. D. A. and PERRY. M. A. Child abuse: Discriminating factors toward a positive outcome. Paper presented at the biennial meeting of the Society for Research in Child Development, .Boston (1981 I. I I. CALISO. J. A. A psychological study of mothers who do not physically abuse their children despite histories of physical abuse in their childhood. Unpublished doctoral dissertation. Seton Hall University, South Orange. NJ (1986). I?. ROBERTSON. K. R., MILNER. J. S. and ROGERS. D. L. History of childhood abuse and later abuse potential. Paper presented at the meeting of the Southwestern Psychological Association, Ft. Worth (1986). 13. ROBERTSON, K. R. and MILNER. J. S. Construct validity of the Child Abuse Potential Inventory. Jorrrnol of Clinic~ul Pswho/o~~ 39:426-429 ( 1983). 14. ELLIS. R. H. and MILNER. J. S. Child abuse and locus of control. Pswhdogicul Reporrs 48:507-510 (1981). 15. STRINGER. S. A. and LA GRECA. A. M. Child abuse potential. Jortrnul qf Ahnormul Child Psvcholog? 13:217-226 (1985). 16. COURON. B. L. Assessing parental potentials for child abuse in contrast to nurturing. Dissertution Ahsrrucr.7 Infernurionui 43~34128 ( 1982). 17. BURGE, E. B. Child abusive attitudes and life changes in an overseas military environment. Dis.certurion Ahsrrucrs Inrernuriontrl 43:562A t 1982). 18. MEE. J. The relationship between stress and the potential for child abuse. Unpublished thesis, Macquarie University, Australia (1983). 19. GOLD. R. G.. MILNER. J. S.. GOLD, S. R. and ROBERTSON, K. R. Mental health and child abuse potential. Paper presented at the Congress of the World Federation for Mental Health. Brighton. England (1985). 20. MATTHEWS, R. D. Screening and identification of child abusing parents through self-report inventories. Unpublished doctoral dissertation, Florida Institute of Technology, Melbourne. FL (19841. 21. ROBERTSON, K. R. and MILNER. J. S. Convergent and discriminant validity of the Child Abuse Potential Inventory. Jorrrnul of Personulity As.sr.ssmen/ 49:86-88 ( 1985). 22. ARAGONA. J. A. Physical child abuse: An interactional analysis. Dissertation Abstracts Inttwzufionul 44:1225B (1983). 23. PRUITT, D. L. A predictive model of child abuse: A preliminary investigation. Disserration Ahsrrucfs Inrernutionu/44:3206B (1983). 24. ROBITAILLE. J., JONES. E.. GOLD. R. G., ROBERTSON. K. R. and MILNER. J. S. Child abuse potential and authoritarianism. Jortrnul of Clinicul Ps.vcho/ogy 41:839-843 (1985). 25. PRUITT. D. L. and ERICKSON. M. T. The Child Abuse Potential Inventory: A study of concurrent validity. Journal of C/irtic,cr/ P.s~c~ho/og~ 41: 104-l I I (1985). 26. LAMPHEAR. V. S.. STETS. J. P., WHITAKER. P. and ROSS. A. 0. Maladjustment in at-risk for physical abuse and behavior problem children: Differences in family environment and marital discord. Paper presented at the meeting of the American Psychological Association. Los Angeles (19851. 27. THOMASSON. E.. BERKOVITZ. T.. MINOR. S.. CASSLE. G.. MCCORD. D. and MILNER. J. S. Evaluation of a family life education program for rural “high risk” families. Journal ofCommnnir.v Psychology 9:246-249 (1981). 28. D’AGOSTINO. P. A.. CHAPIN. E and MOORE. J. B. Rainbow Family Learning Center: Help for parent. haven for children. Paper presented at the Fifth International Congress on Child Abuse and Neglect. Montreal (1984). 29. ROBERTSON. W. K. T, and MILNER. J. S. Detection of conscious deception using the Child Abuse Potential Inventory lie scale. Jonmu/ ofPersonulify As.sessmenr 49:541-W (1985).
Copyright
0145.2134/x7 s3.w * .(X) c 19x7 Pergamon Journals Ltd
CHILD ABUSE POTENTIAL AND WORK SATISFACTION IN DAY-CARE EMPLOYEES DONALD
W. ATTEN,
M. ED. AND JOEL
S. MILNER,
PH.D.
Western Carolina University, Cullowhee. NC Abstract-The Child Abuse Potential (CAP) Inventory and a Work Satisfaction Questionnaire were administered to 228 day-care employees at 40 centers to investigate the relationship between potential for child abuse and degree of job satisfaction. In addition. the ability of the CAP abuse scale to distinguish between groups of day-care employees thought to differ in the quality of child care provided was studied. Small. albeit significant, inverse correlations were observed between abuse scores and two job satisfaction items. As clarity of job expectations and as overall job satisfaction increased. abuse scores decreased. Although abuse scores for all day-care employees were not significantly different from norm scores. significantly lower scores were found for employees from “superior” day-care centers and for employees rated “high” by center directors. Abuse scores for employees rated “high” were also significantly below scores for employees rated “low.” While these data indicate only limited relationships between child abuse and job satisfaction. the findings suggest the CAP abuse scale can distinguish some subgroups of day-care employees thought to differ in quality of child care provided. Resume-Linventaire de potentiel de s&ices a I’egard d’enfants (CAP ou Child Abuse Potential Inventory) ainsi qu’un questionnaire portant sur la satisfaction au travail ont ete soumis a 228 employ& de creches dans 40 centres. Le but Ctait de voir quelles relations on pouvait etablir entre le potentiel de s&ices a I’egard d’enfants et le degre de satisfaction apporte par I’emploi. En outre la capacite de I’echelle CAP de distinguer entre differents groupes d’employes et de creches a ete etudie. Ces groupes etaient form& de personnes qui semblaient se differencier entre elles par la qualite des soins qu’elles apportaient aux enfants. Des differences significatives bien que petites ont ete observees et une relation inverse est apparue entre le potentiel de s&ices et la satisfaction apportee par I’emploi. En d’autres termes. lorsque la gratification que le personnel ressentait de ce genre d’emploi apparaissait clairement et que consequemment, la satisfaction augmentait, les scores potentiels de s&ices diminuaient. On a trouve des scores de potentiel de s&ices plus bas lorsqu’on avait affaire a des employes de creches qualifies de bons employes par les directeurs des centres. Par contre lorsque les employes etaient moins apprecies des directeurs de centres les scores de s&ices potentiels etaient plus eleves. Les donnees recoltees indiquent done une relation limitte mais certaine entre la satisfaction apportte par I’emploi et le risque de s&ices; les observations suggerent surtout que I’inventaire CAP permet de distinguer des sousgroupes a I’interieur d’une cohorte d’employes de creches qui sont soupconnes de donner des soins de qualite inferieure aux enfants.
INTRODUCTION EMPLOYEE MALTREATMENT of children in day-care centers has recently received increased attention in the general media. While there are articles on the stresses and subsequent burnout of employees in the field of child care [l-4], a literature review [5] reveals few empirical studies on the putative indicators of child abuse in out-of-home child care facilities. One study [6] which attempted to investigate variables related to employee child abuse used the Child Abuse Potential (CAP) Inventory, work satisfaction items, and demographic variables to distinguish between known institutional child abusers and matched employee controls. Overall, the CAP abuse score, three work satisfaction items, and several demographic variables were found to correctly classify 93% of
Reprint requests to Joel S. Milner, Ph.D.. Department of Psychology, Western Carolina University, Cullowhee. NC 28723 II7
118
Donald W. Atten and Joel S. Milner
the abusive and nonabusive employees. This study, however, has not been replicated and no similar study has been conducted in day-care settings. Given the concetn of the general public and professionals in the field of day care and the paucity of empirical data regarding child abuse in day-care settings, the need to investigate variables related to employee child abuse remains a pressing issue. The present study was designed to investigate the relationship between the CAP Inventory and work satisfaction items in day-care employees. In order to provide preliminary data on the ability of the CAP Inventory to distinguish between levels of quality of child care. daycare employees were divided into subgroups including Level 1 day-care center employees, Level 2 day-care employees, “superior” day-care employees, and groups of day-care employees rated “high” and “low” in terms of relative quality of child care provided.
METHOD Subjects A total of 228 Level 1 and Level 2 day-care employees from 40 participating day-care sites in North Carolina volunteered to take part in the study. Eight workers from the 40 participating sites declined to take part in the study. Further, three directors/coordinators of additional sites declined to grant permission for their day-care staff to be contacted. After testing, 9 subjects withdrew from the study leaving 219 subjects for the study sample. All subjects were employed at Level 1 (n = 44) or Level 2 (n = 175)‘state subsidized daycare centers. To qualify for subsidy funds, a day-care center must meet minimum requirements in 16 standards areas defined by the North Carolina State Social Services Commission. Level 2 day-care centers must meet all minimum standards set for Level 1 centers, plus additional requirements in 8 of the 16 standards areas. Standards for Level 2 centers are stricter in the following areas: operational policies, personnel policies, square footage per child, number of toilets per child, staff-child ratios, qualifications of the director, number of age-appropriate children’s activities, and number of different activity areas. Requirements for child care staff and staff training are the same for both Level 1 and Level 2 centers. Still, the assumption remains that higher standards and increased subsidy for Level 2 centers means a better quality of care is provided at these sites. In addition to the identification of Level 1 and Level 2 day-care employees, a subgroup of Level 2 sites were identified as “superior” day-care sites by the North Carolina State Office of Day Care. All employees from the “superior” sites were included in a single “superior” subgroup for comparison purposes. Further, all day-care center directors identified employees they believed comprised the top (rated high) and bottom’ (rated low) 10% of their staff in terms of overall quality of child care. Demographic characteristics for each of the day-care employee groups and subgroups are presented in Table I.
Test instruments The CAP Inventory is a brief, client-administered 160-item screening device which is answered in a force-choice, agree-disagree format [7]. Classification data indicate overall classification rates for physical abusers and matched comparison parents are generally in the low 90s [8]. Cross-validation data on over 3.000 subjects from a variety of maltreating.
mean (SD)
number
of children
trith children)
74 62 46
53 36 57
% Married
Groups
(see percent
88 58 58
60 s7 61
% White
of employees
99 IO0 I 00
99 IO0 99
Q Female
for the Day Care Employee
for the subgroup
13.3 (2.3) 13.8 (2.1) 13.8 (2.0)
13.2 (2.0) 12.7 (1.7) 13.4 (2.01
x (SD) Education
Characteristics
for each group was determined
38.3 (11.8) 39.6 (10.7) 32.4 (I 1.8)
96 26 24
a The
34.8 (I2.0) 3O.6 112.0) 35.7 (Il.81
219 44 I75
All day care employees Level I day care employees Level 2 day care employees “Superior” day care center employees Rated “high” day care employees Rated “low” day care employees
Age
II
x (SIN
Demographic
Group
Table I.
reporting
71 69 54
59 48 62
% With Children
children.
I .5 (0.6)
2.1 (1.1) I .7 to.71
I.9rI.l) I.5 10.7) ?.O(l.l)
x (.SUl Children”
120
Donald W. Atten and Joel S. Milner
general population, and nurturing parent groups are available in a recent manual [7]. A longitudinal, predictive validity study [9] indicates that elevated abuse scores are significantly (p < .OOOl) related to later physical child abuse, marginally (p < .05) related to subsequent neglect, and not related 0, > .05) to later failure-to-thrive cases. Factor analytic studies [7] indicate that six dimensions appear to be descriptive of the CAP abuse scale. These factors include distress, rigidity, unhappiness, problems with child and self, problems with family, and problems from others. Numerous construct validity studies provide additional data on the characteristics measured by the abuse scale. For example, elevated abuse scores are related to a childhood history of abuse [lo-121; low self-esteem [lo]; poor ego strength [13]; external locus of control [14, 151; stress [16-191; isolation [17, 201; anxiety [19-211; immaturity, poor impulse control, aggression, assaultative behavior [19, 20, 22, 231; depression [19, 20, 241; neuroticism [21, 231; perception of a child with problems [15, 221; physiological liability to child stimuli [2.5]; and family conflict [lo, 261. Other research indicates the CAP abuse scale is sensitive to changes produced by intervention/treatment programs [27, 281. Reliability data on the internal consistency and temporal stability of the CAP inventory are available [7]. Internal consistency estimates have been published for control, at-risk, neglect, and abuse groups as a function of location, gender, age, educational level, and ethnic background. Temporal stability estimates are available for control subjects across four time intervals (i.e., l-day, l-week, l-month, and 3-month) and as a function of demographic variables. The abuse scale has internal consistency reliabilities of .92 to .96 for control subjects and .95 to .98 for abusive subjects. No demographic effects were reported. Temporal consistency estimates for the abuse scale are also adequate with .91, .90, .83, and .75 for l-day, l-week, l-month, and 3-month intervals, respectively. Again no demographic effects were found. A brief, 8-item Work Satisfaction Questionnaire was developed based on putative indicators of child abuse found in a comprehensive review of institutional abuse literature [5]. The eight items cover the following content areas: Item 1, quality of supervision; Item 2, relevance of training for job: Item 3, satisfaction with pay: Item 4, opportunity for career advancement; Item 5, ability to handle difficult children; Item 6. ability to handle job-related stress; Item 7, clear job performance expectations; and Item 8, overall job satisfaction. All items were answered on a .5-point Likert Scale. Procedure
The CAP Inventory and the Work Satisfaction Questionnaire were administered on a voluntary basis to day-care employees at 40 participating centers. Prior to completing the test materials, each subject was given an information sheet which included the name of the principal investigator, a toll-free phone number that could be called for additional information, an informed consent, and a self-addressed, prepaid, number-coded postcard which, if mailed at a later date, would remove the employee from the study. RESULTS Two hundred and twenty-eight day-care employees volunteered to take the CAP lnventory and Job Satisfaction Questionnaire. Nine postcards requesting withdrawal from the study were received and the corresponding test materials were destroyed. Of the 219 remaining employees, 6 had incomplete (i.e., more than 10% blanks) CAP Inventories and 61 had elevated lie scale scores. Since validity data show elevated lie scores are indicative of faking good behavior [7, 291 and demonstrate use of the lie scale reduces the number of
Child abuse and work satisfaction
121
abuse scale misclassi~cations [8], responses with elevated lie scores were considered questionable and removed. After exclusions, data from 152 subjects were available for analysis. The abuse scale internal consistency estimate (KR-20) for this group of day-care employees was .90. Only two of the eight work satisfaction items were found to have significant inverse correlations with abuse: clear job performance expectations and CAP abuse scores, r = .26, & = 150, p < .OOl; and overall job satisfaction and CAP abuse scores, r = - .14, df = 149, p c .Ol. The mean CAP abuse scores for each of the day-care groups are presented in Table 2. For comparison purposes a norm group of subjects (parents) are also presented in Table 2. An analysis of variance revealed no significant difference, F = .lS, df = l/986, p > .05, between the day-care employees and the norm group. Several additional analyses were performed on the data presented in *able 2, WhiIe higher than the norm group, the mean abuse score for Level 1 employees was not significantly elevated, F = I. 10, df = 11859, p > .05, above the norm group mean. While lower, the mean abuse score for Level 2 employees was not significantly below, F = .98, df = l/961, p > .0.5, the norm group mean. Further, a direct comparison of Level 1 and Level 2 group mean abuse scores did not reveal any significant difference, F = 1.19, df = l/150, p > .05.
In contrast, the “superior” center employees’ mean abuse score was significantly below, F = 4.92, df = l/906, p < .05, the mean norm abuse score. The rated “high” employees were also significantly below, F = 7.29, df = 11848, p < .OI, the mean norm abuse score. While the mean abuse score for the rated “low” employees was not significantly above the mean norm score (F = 1.30, df = 11852, p > .0.5), the mean norm abuse score for the rated “low” employees was signi~cantly above the mean abuse score for rated “high” employees (F = 4.42, df = 1130, p < .OS>. Inspection of the individual CAP abuse scores revealed that none of the rated “high” employees had scores above the abuse scale cutoff [7], while three of the rated “low” employees had abuse scores above the cutoff for physical abuse. For the work satisfaction items, education was positively correlated with relevance of training for the job, r = .27, df = 143, p -=c ,001; and employee’s age was positively correlated with satisfaction with pay, r = 23, @’ = 14.5,p < .Ol; and with ability to handle job related stress, Y = .22, df = 145, p < .Ol. The only demographic variable correlated with abuse was education, Y = - .23, df = 146, p < .OI. DISCUSSION Although the institutional abuse literature [5] is replete with references that indicate job-related factors are important in the occurrence of employee child abuse, the present Table 2.
Mean (SD) Abuse Scores for Norm Subjects and Day Care Employee Groups Group
Norm subjects* All day care employees Level 1 day care employees Level 2 day care employees “Superior” day care employees Rated “high” day care employees Rated “Low“ day care emplovees
n
B
SD
836 152 2s 127 72 14 18
91.0 88.6 106.8 as.2 75.6 67.1 113.9
75.0 62.0 74.0 58.8 54.7 31.7 84.6
* Data for norm subjects were taken from Milner (7).
122
Donald
W. Atten
and Joel S. Mimer
data revealed only two of eight job satisfaction items- clear job performance expectawere modestly related to a measure of child abuse tions and overall job satisfactionpotential in day-care employees. These data support the general findings of Haddock and McQueen [6] who found that only 3 of 20 job satisfaction items discriminated known institutional child abusers from matched comparison employees. Institutional child abusers were dissatisfied with their chances to work independently, with their opportunity for advancement, and with their feelings of accomplishment. While conclusions based on these two correlational studies must be viewed with caution, the findings do not support the view that job satisfaction variables are strongly related to employee child abuse. Overall, no differences were found between day-care employees’ CAP abuse scores and norm parent abuse scores. These data suggest that, as a group, the North Carolina day-care employees tested were not different in abuse potential from nonscreened, general population parents. While differences had been expected between Level 1 and Level 2 employee abuse scores, no differences were found. Differences had been expected because additional requirements are maintained in 8 of the 16 standards areas for Level 2 centers. However, Level 1 and Level 2 centers have the same standards for child care staff and staff training. Whether additional standards are unrelated to quality of child care or whether the CAP Inventory is insensitive to differences that do exist is not known. As expected, day-care employees from superior centers had abuse scores significantly below norm parent abuse scores. These centers had been recommended by the North Carolina State Offtce of Day Care as outstanding in quality of child care, in supervision of staff, and in stability of staff. Further, the day-care employees rated high by center directors had the lowest abuse scores and the smallest variance of any subgroup tested. The rated high employees were significantly below the rated low employees and the norm parent group. These data indicate coticordance between the CAP abuse scores and ratings by center directors based on their observational knowledge of the quality of child care provided by their employees. As previously noted, the present study is correlational in nature and the findings must be viewed as only suggestive. Another study limitation was that the sample of day-care subjects were considered normal and, as such, had a limited range of abuse scores. While this makes observed subgroup differences in CAP abuse scores more impressive, it limits the degree of correlations possible between the CAP abuse scale and the work satisfaction measures. Likewise, use of a normal sample provides no information on the ability of the CAP Inventory to distinguish employees who are abusive from those who are not. Nonetheless, these data do indicate additional research is warranted on the use of the CAP Inventory in the screening of day-care employees.
REFERENCES 1. FREUDENBERGER. H. J. Burn-out: Occupational hazard of the child care worker. Child 6:9O-99 (1977). 2. HYSON, M. C. Playing with children all day: Job stresses in early childhood education. 37:25-32 (1982). 3. MASLACH. C. and PINES, A. The burn-out syndrome in the day-care setting. Child Carr II3 (1977). 4. MATTINGLY. M. A. Sources of stress and burn-out in professional child care work. Child
CUI.P Q~~urtrr/~ Yolrn~ Cltildren
QIUI~IC~/.V6: IOOCure QrwrIer/y
6:127-137 (1977). 5. MILNER. Document.s
J. S. and ATTEN, D. W. Institutional child abuse and neglect: A bibliography. P.sw/to/o,qicrrl 15:22 (1985). 6. HADDOCK, M. D. and MCQUEEN. W. h4. Assessing employee potentials for abuse. Jorrrnd c;fC/inictr/ Ps~rhol0~y 39: I02 I - I029 t I983 1. 7. MILNER, J. S. 771~ Child Ahrrse Potatrtirrl /rt~~,r~rcwv: Mtrrrrd. (2nd ed.1. Psytec. Weh\ter. NC (1986).
Child abuse
and work
satisfaction
123
8. MILNER. J. S.. GOLD. K. G. and WIMBERLEY. R. C. Cross-validation of the Child Abuse Potential Inventory. Jorrrricrl c!f C‘r~ri.\rt//i~i~ trrid C/ipfi(.0/ Ps,vcho/o,~,v. (in press). 9. MILNER. J. S.. GOLD. R. G.. AYOUB. C. and JACEWITZ. M. M. Predictive validity of the Child Abuse Potential Inventory. Jorrrxcll 0/ Cofrsrrlrin~ rr/id C/inic,u/ P.r.vc~ho/oR~ 52:X79-884 ( 1984). IO. CHAN. D. A. and PERRY. M. A. Child abuse: Discriminating factors toward a positive outcome. Paper presented at the biennial meeting of the Society for Research in Child Development, .Boston (1981 I. I I. CALISO. J. A. A psychological study of mothers who do not physically abuse their children despite histories of physical abuse in their childhood. Unpublished doctoral dissertation. Seton Hall University, South Orange. NJ (1986). I?. ROBERTSON. K. R., MILNER. J. S. and ROGERS. D. L. History of childhood abuse and later abuse potential. Paper presented at the meeting of the Southwestern Psychological Association, Ft. Worth (1986). 13. ROBERTSON, K. R. and MILNER. J. S. Construct validity of the Child Abuse Potential Inventory. Jorrrnol of Clinic~ul Pswho/o~~ 39:426-429 ( 1983). 14. ELLIS. R. H. and MILNER. J. S. Child abuse and locus of control. Pswhdogicul Reporrs 48:507-510 (1981). 15. STRINGER. S. A. and LA GRECA. A. M. Child abuse potential. Jortrnul qf Ahnormul Child Psvcholog? 13:217-226 (1985). 16. COURON. B. L. Assessing parental potentials for child abuse in contrast to nurturing. Dissertution Ahsrrucr.7 Infernurionui 43~34128 ( 1982). 17. BURGE, E. B. Child abusive attitudes and life changes in an overseas military environment. Dis.certurion Ahsrrucrs Inrernuriontrl 43:562A t 1982). 18. MEE. J. The relationship between stress and the potential for child abuse. Unpublished thesis, Macquarie University, Australia (1983). 19. GOLD. R. G.. MILNER. J. S.. GOLD, S. R. and ROBERTSON, K. R. Mental health and child abuse potential. Paper presented at the Congress of the World Federation for Mental Health. Brighton. England (1985). 20. MATTHEWS, R. D. Screening and identification of child abusing parents through self-report inventories. Unpublished doctoral dissertation, Florida Institute of Technology, Melbourne. FL (19841. 21. ROBERTSON, K. R. and MILNER. J. S. Convergent and discriminant validity of the Child Abuse Potential Inventory. Jorrrnul of Personulity As.sr.ssmen/ 49:86-88 ( 1985). 22. ARAGONA. J. A. Physical child abuse: An interactional analysis. Dissertation Abstracts Inttwzufionul 44:1225B (1983). 23. PRUITT, D. L. A predictive model of child abuse: A preliminary investigation. Disserration Ahsrrucfs Inrernutionu/44:3206B (1983). 24. ROBITAILLE. J., JONES. E.. GOLD. R. G., ROBERTSON. K. R. and MILNER. J. S. Child abuse potential and authoritarianism. Jortrnul of Clinicul Ps.vcho/ogy 41:839-843 (1985). 25. PRUITT. D. L. and ERICKSON. M. T. The Child Abuse Potential Inventory: A study of concurrent validity. Journal of C/irtic,cr/ P.s~c~ho/og~ 41: 104-l I I (1985). 26. LAMPHEAR. V. S.. STETS. J. P., WHITAKER. P. and ROSS. A. 0. Maladjustment in at-risk for physical abuse and behavior problem children: Differences in family environment and marital discord. Paper presented at the meeting of the American Psychological Association. Los Angeles (19851. 27. THOMASSON. E.. BERKOVITZ. T.. MINOR. S.. CASSLE. G.. MCCORD. D. and MILNER. J. S. Evaluation of a family life education program for rural “high risk” families. Journal ofCommnnir.v Psychology 9:246-249 (1981). 28. D’AGOSTINO. P. A.. CHAPIN. E and MOORE. J. B. Rainbow Family Learning Center: Help for parent. haven for children. Paper presented at the Fifth International Congress on Child Abuse and Neglect. Montreal (1984). 29. ROBERTSON. W. K. T, and MILNER. J. S. Detection of conscious deception using the Child Abuse Potential Inventory lie scale. Jonmu/ ofPersonulify As.sessmenr 49:541-W (1985).