the influence of parenting stress and child behavior problems on parental estimates of expressive...

Journal of Abnormal Child Psychology, Vol. 19, No. 1, 1991

The Influence of Parenting Stress and Child Behavior Problems on Parental Estimates of Expressive and Receptive Language Development

C. A. Chaffee, 1 Charles E. Cunningham, 1,4 Margaret Seeord-Giibert, 3 Heather Elbard, 2 and Joanne Richards 3

This study examined the relative contributions o f parental reports o f child- related and parent-related stress to their estimates o f expressive and receptive language skills on the Minnesota ChiM Development Inventory in 152 children referred to a communicative disorders clinic. Correlational analysis showed moderately high agreement between parental estimates and objective measures of expressive and receptive language ability. Multiple-regression analyses revealed that independent measures o f language acquisition accounted for the greatest amount o f variance in parental estimates of expressive and receptive language skills. Stressful child behavior accounted for a small but statistically significant proportion o f the variance in estimates o f expressive and receptive language skills. Parental stress, in contrast, did not contribute significantly to language estimates. These findings suggest that parents are able to formulate estimates o f language development which are relative- ly independent o f personal or child-related stressors.

Although parents are widely used as informants regarding their children's behavior and development (Mash & Terdal, 1988), studies regarding their

Manuscript received in final form July 18, 1990 We gratefully acknowledge the assistance of Mary Ann Evans for the valuable review of this manuscript.

IChedoke-McMaster Hospitals and McMaster University, Hamilton, Ontario, Canada L8N 3Z5. 2Chedoke-McMaster Hospitals, Hamilton, Ontario, Canada L8N 3Z5. 3Department of Pediatrics, Chedoke-McMaster Hospitals and McMaster University, Hamil- ton, Ontario, Canada L8N 3Z5.

4Address all correspondence to Charles E. Cunningham, Ph.D., Department of Psychology, Chedoke Division, Chedoke-McMaster Hospital, Box 2000 A, Hamilton, Ontario, Canada L8N 3Z5.

65

0091-0627/91/0200-0065506.50/0 �9 1991 Plenum Publishing Corporation

66 Chaffee, Cunningham, Secord-Gilbert, Eibard, and Richards

objectivity have produced conflicting results. While some studies find that parent reports on standardized behavior problem checklists correspond to more objective assessments (Lachar, Kline, & Gdowski, 1987; Richters & Pellegrini, 1989; Schaughency & Lahey, 1985), others suggest parent ratings may be negatively biased by stresses such as marital conflict, depression, or psychopathology (Christensen, Phillips, Glascow, & Johnson, 1983; Estroff et al., 1984; Fergusson, Horwood, Gretton, & Shannon, 1985; Forehand, Wells, McMahon, Griest, & Rogers, 1982; Griest, Wells, & Forehand, 1979; Rickard, Forehand, Wells, Griest, & McMahon, 1981; Schaughency & La- hey, 1985; Webster-Stratton, 1988). While a considerable number of studies have examined the effects of personal stressors on parental perceptions of child behavioral difficulties, little is known regarding the influence of social variables on parental evaluation of their child's cognitive, language, motor, or adaptive skills. Nonetheless, parental reports regarding children's development on standardized developmental checklists such as the Minnesota Child Development Inventory (MCDI; Ireton & Thwing, 1972) are widely used to assess children with a variety of developmental difficulties (Byrne, Backman, & Smith, 1986; Chaffee, Cunningtlam, Secord-Gilbert, Elbard, & Richards, in press; Colligan, 1977; Ireton, Thwig, & Currier, 1977; Saylor & Brandt, 1986). While most studies reveal moderate correlations between parental reports and more objective measures of development, there is little informa- tion regarding variables which may influence the accuracy of parental evaluations.

Among children with language disorders, for example, parent reports correlate moderately with independent measures of receptive and expressive language acquisition (Chaffee et al., in press). Nonetheless, a considerable proportion of the variance in maternal estimates appears to be influenced by factors unrelated to the child's language skills. Several factors suggest that the markers of social adversity which appear to distort parental perceptions of behavioral difficulty may also influence estimates of language development.

Parents of language-delayed children experience higher levels of psychosocial distress and report an increase in the prevalence of behavioral difficulties which are, themselves, a source of considerable parenting stress (Beitchman, Nair, Clegg, Ferguson, & Patel, 1986). The prevalence of attention deficit hyperactivity disorder, a group whose parents report higher levels of stress (Breen & Barkley, 1988), depression (Befera & Barkley, 1985; Cunningham, Benness, & Siegel, 1988), and marital discord (Befera & Bark- ley, 1985), for example, is up to six times the level observed in the population at large (Baker & Cantwell, 1982; Beitchman, Nair, Clegg, Ferguson, and Patel, 1986; Cantwell, Baker, & Mattison, 1979; Szatmari, Offord, & Boyle, 1989).

Parenting Stress and Language Estimates 67

Stressors experienced by families of language-disordered chgildren (Beitchman, Nair, Clegg, Ferguson, and Patel, 1986) are similar to those which appear to reduce the accuracy of parental reports of behavioral problems. If these variables bias reports regarding child development, the interpretation and utility of widely used questionnaires may need to be recon- sidered. This study, therefore, examined the relative influence of child-related stress, parent-related stress, and sociodemographic variables on parental estimates of expressive and receptive language skills in a sample of children referred to a communicative disorders clinic.

METHOD

Subjects

A sample of 152 children (115 boys, 37 girls) ranging in age from 24 to 87 months (M = 49 months, SD = 12.6) referred consecutively to the Communicative Disorders Clinic in the Children's Hospital at Chedoke- McMaster Hospitals participated in this study. The Communicative Disord- ers Clinic is a regional program providing diagnostic services for a metropoli- tan area with a population of approximately 429,500. The majority of subjects were white (98%), 1% were black, and 1% were Native American. Most (84%) parents were married or living with a partner. On the Blishen socioeconomic scale (SES) (Blishen & McRoberts, 1976), 3% of the subjects were in the highest class, 15% in the second, 11% in the third, 12% in the fourth, 16% in the fifth, and 43% in the lowest socioeconomic category. Using a cutoff score 1.5 standard deviations below the mean for age (Beitchman, Nair, Clegg, & Patel, 1986) on the Reynell Developmental Language Scales (Re- vised) (Reynell, 1977), 37% of the referral sample demonstrated no significant delay in language, 27% demonstrated delays in expressive language alone, 6% showed delays in language comprehension alone, and 30% showed delays in both comprehension and expression.

The means and standard deviations for each demographic measure are presented in Table I.

Dependent Measures

Language. The Reynell Developmental Language Scales (Revised) (Rey- nell, 1977) is a widely used standardized measure of the expressive and receptive language of children ranging in age from 6 months to 6 years. The Verbal Comprehension scale requires the child to respond nonverbally to a series

68 Chaffee, Cunningham, Secord-Gilbert, Elbard, and Richards

Table I. Mean and Standard Deviation Scores for Nonverbal Age, Language, and Stress Scores ~

M SD Chronological age (months) 49.0 12.6 Leiter nonverbal IQ score 93.5 20.6 MCDI expressive language age 32.3 11.7 MCDI comprehension-conceptual age 37.8 14.1 Reynell Expressive Language age 35.6 16.3 Reynell Language Comprehension age 38.9 13.2 PSI Parent Domain z score -0.1 1.0 PSI Child Domain z score 1.1 1.2

aMCDI = Minnesota Child Development Inventory; PSI = Parenting Stress Index.

of increasingly complex questions and instructions. The Expressive Language scale measures presymbolic skills, naming, the ability to describe word mean- ings, and the use of language to express consecutive ideas. Measures f rom both scales are given in terms of z scores.

Parent Evaluation of Language Development. On the MCDI (Ireton & Thwing, 1972), parents respond yes or no to 320 descriptions of developmental milestones accomplished between the ages of 6 and 75 months. The MCDI yields age equivalent scores for general development, gross motor , fine motor , expressive language, comprehension-conceptual, situation comprehension, self-help, and personal-social skills. This study examined only the scores for the expressive language and comprehension-conceptual scales.

Parenting Stress. The Parenting Stress Index (PSI) (Abidin, 1983) is a 150-item self-report questionnaire intended for use with parents of children under 10 years of age. The Child Domain Stress subscale of the PSI includes measures of the child's adaptabili ty/plasticity, acceptability to parent, demandingness/degree of bother, mood, distractibility/activity, and ability to reinforce the parent. The Parent Domain Stress subscale of the PSI includes measures of depression, a t tachment to child, restrictions imposed by parental role, sense of competence, social isolation, relationship with spouse, and parental health. This study is based on z scores for the Parent Domain Stress and Child Domain Stress subscales.

R E S U L T S

Predictors of MCDI Expressive Language

Hierarchical multiple-regression analyses (Pedhazur, 1982) were calcu- lated to determine the relative contribution of independent measures of ex-


pressive and receptive language ability, stressful child behavior, parental stress, and sociodemographic variables to parent reports of expressive and receptive language development on the MCDI. Measures of gender, SES, child stress, and parental stress were entered first followed by Reynell expressive and receptive language scores. Age scores were entered last to determine the residual contribution of this variable. The coefficients of multiple determination (r 2) are presented in Table II.

While the child's gender and family SES did not contribute significantly to parental estimates, PSI Child Domain Stress accounted for 4% of the variance in expressive language estimates. Parents whose children were judged more stressful provided lower estimates of the child's expressive language skills. Reynell Expressive Language score accounted for an additional 45% of the variance in parental estimates of expressive language ability on the MCDI. The Reynell Comprehension score contributed an additional 5% to the variance in parental estimates. Finally, age accounted for a residual 2% of the variance. When Reynell Expressive Language and Comprehension scores preceded the entry of PSI Stress measures in a second equation (Ped- hazur, 1982), Child and Parent Domain Stress did not contribute to the variance in expressive language estimates.

Table I1. Hierarchical Regression Equation Predicting MCDI Expressive Language Age Score

Multiple

Predictor variable r r 2 r z Change

MCDI language expression

Gender - .067 .005 .005 Socioeconomic status - . 086 .012 ,008 PSI Child Domain Stress - . 209 .053 ,041 b PSI Parent Domain Stress - . 0 9 9 ,053 .000 Reynell expression age .702 .511 .449 ~ Reynell comprehension age .705 .557 .046 c Chronological age .608 .573 ,020 b

MCDI language comprehension

Gender - .049 .002 .002 Socioeconomic status - , 077 ,009 ,006 PSI Child Domain Stress - . 239 .062 .054 d PSI Parent Domain Stress - . 172 .067 .005 Reynell comprehension age .758 .573 .506 c Reynell expression age .745 .634 .063 c Chronological age .686 .669 .035 c

"MCDI = Minnesota Child Development Inventory; PSI = Parenting Stress Index.

~z, < .05. cp < .05. dp < .01.


Predictors of MCDI Comprehension-Conceptual !

The coefficients of multiple determination (r 2) for the predictors of MCDI language comprehension scores are presented in the lower half of Table II. Again, while child gender and family SES did not contribute significantly, PSI Child Domain Stress accounted for 5~ of the variance in MCDI lan- gauge comprehension scores. Reynell Language Comprehension scores accounted for an additional 51070 of the variance, with Reynell Expressive Language scores contributing an additional 6070 to the variance in MCDI comprehension scores. Finally, the child's age accounted for a residual 3.5070 of the variance. When language scores preceded the entry of PSI Stress scores in a second regression equation, Stress scores did not contribute significantly to language comprehension estimates.

While stressful child behavior was linked to lower estimates of language expression and comprehension skills, the individual Child Domain subscales revealed no clear pattern: demandingness (SD = 1.2), chilld's acceptability (SD = 1.1), moodiness (SD = 0.8), distractibility/hyperactivity (SD = 0.8), adaptability (SD = 0.7), reinforces parent (SD = 0.4). Moreover, while the composite Child Domain score accounted for approximately 4070 of the variance in expressive and receptive language estimates, the individual scores did not contribute significantly.

DISCUSSION

Although Child Domain Stress scores were associated with a very small, though significant, reduction in parental estimates of expressive language skills, measures of Parent Domain Stress did not contribute to estimates of expressive and receptive language ability. This finding is consistent with previous reports of consistently strong correlations between the MCDI and objective measures of developmental status (Colligan, 1977; Dean & Steffen, 1984; Eisert, Spector, Shankaran, Faigenbaum, & Szego, 1980; Ireton et al., 1977; Kenny, Hebel, Sexton, & Fox, 1987; Saylor & Brandt, 1986). The MCDI, consequently, appears to have good predictive validity (Colligan, 1976, 1981; Gottfried, Guerin, Spencer, & Meyer, 1984; Guerin & Gottfried, 1987) and classification ability (Byrene et al., 1986; Chaffee et al., in press; Gott- fried et al., 1984; Kenny et al., 1987; Sturner, Funk, Thomas, & Green, 1982).

The results of this study are consistent with recent reports failing to support the depression-distortion hypothesis (Richters & Pellegrini, 1989). Moreover, they suggest that the negative bias psychosocial stress exerts on parental evaluations of their children's behavior (Panaccione & Wahler, 1986; Webster-Stratton, 1988) may not extend to their evaluations of language de-


velopment. Several variables may account for this discrepancy. First, parents of the language-delayed children studied here may not have experienced levels of stress which were significant enough to influence estimates of language skills. Indeed, the Parent Domain subscale, a measure reflecting depression, attachment to child, restrictions imposed by parental role, parents' sense of competence, social isolation, relationship with spouse, and parental health, was well within the normal range. The populations studied by investigators examining child behavior problems, in contrast, appear to have been exposed to substantially higher levels of psychosocial distress (Panaccione & Wah- ler, 1986; Webster-Stratton, 1988). Although this sample of language-delayed children was considered to be more stressful than 86~ of the normative population, stress stemming from social isolation, marital conflict, or depression might bias parental evaluations more than stressful child behavior. Alterna- tively, problematic behavior, which is often perceived as intentional and un- controllable (Sobol, Ashbourne, Earn, & Cunningham, 1989), may elicit an affective response which may distort parental perceptions more than delays in language acquisition. Finally, the item content of the MCDI and Reynell language scales may be more similar than parental ratings of behavior and the measures collected by independent home observers. Schaughency and La- hey (1985), for example, reported that parental ratings of child behavior correlate more highly with teacher ratings than with the reports of observers in the home.

While the elevated Child Domain Stress scores observed here are consistent with studies reported an increased prevalence of psychiatric disorders among language-delayed children (Baker & Cantwell, 1982, 1987; Beitchman, Nair, Clegg, Ferguson, & Patel, 1986; Cantwell & Baker, 1980; Cantwell et al., 1979; Silva, Williams, & McGee, 1987; Stevenson & Richman, 1978), Parent Domain Stress scores were within normal limits. Although these findings appear inconsistent with studies reporting elevated psychosocial stress in families of speech- and language-impaired children (Beitchman, Nair, (Clegg, & Patel, 1986), the different sampling frames adopted in the present vs. previous studies limit direct comparison. Beitchman, Nair, Clegg, Ferguson, and Patel's (1986) 5 �89 sample was, for example, 17 months older than the group studied here. Since normal language skills emerge by age 5 �89 in a significant number of language-delayed preschoolers (Bishop & Edmun- son, 1987), older samples represent a more severely impaired population. Second, since this study was designed to examine variables influencing the validity of the MCDI in clinical contexts, we included all referrals to a preschool Communicative Disorders Clinic. Although language delays were suspected in all referrals, 37% evidenced expressive and receptive language scores within normal limits. Beitchman's epidemiologically selected sample, in contrast, included only 5 Vz year olds with demonstrated language delays.

72 Chaffee, Cunningham, Secord-Gilberl, Elbard, and Richards

These findings require replication. Nonetheless, while limited to preschool referrals to a communicative disorders clinic, they suggest that parental evaluations of expressive and receptive language skills are relative- ly independent of the types of personal stressors which appear to bias evaluations of child behavior.

REFERENCES

Abidin, R. R. (1983). Parenting Stress Index. Charlottesville: Pediatric Psychology Press. Baker, L., & Cantwell, D. P. (1982). Developmental, social and behavioral characteristics of

speech and language disordered children. ChiM Psychiatry and Human Development, 12, 195-206.

Baker, L., & Cantwell, D. P. (1987). Comparisons of well, emotionally disordered, and be- haviorally disordered children with linguistic problems. Journal of the American Academy of Child Psychiatry, 26, 193-196.

Befera, M. S., & Barkley, R. A. (1985). Hyperactive and normal girls and boys: Mother-child interaction, parent psychiatric status and child psychopathology. Journal of Child Psy- chiology and Psychiatry, 26, 439-452.

Beitchman, J. H., Nair, R., Clegg, M., & Patel, P. G. (1986). Prevalence of speech and language disorders in 5-year-old kindergarten children in the Ottawa-Carleton region. Journal of Speech and Hearing Disorders, 51, 98-110.

Beitchman, J., Nair, R. Clegg, M., Ferguson, B., & Patel, P. G. (1986). Prevalence of psychiatric disorder in children with speech and language disorders. Journal of the A meri- can Academy of Child Psychiatry, 25, 528-535.

Bishop, D. V. M., & Edmundson, A. (1987). Language-impaired 4 year olds: Distinguishing transient from persistent impairment. Journal of Speech and Hearing Disorders, 52, 156-173.

Blishen, B. R., & McRoberts, H. A. (1976). A revised socioeconomic index for occupations in Canada. Canadian Review of Sociology and Anthropology, 13, 71-79.

Breen, M. J., & Barkley, R. A. (1988). Psychopathology and parenting stress in girls and boys having attention deficit disorder with hyperactivity. Journal of Pediatric Psychology, 13, 265-280.

Byrne, J. M., Backman, J. E., & Smith, I. M. (1986). Developmental assessment: The clinical use and validity of parental report. Journal of Pediatric Psychololgy, 11, 549-559.

Cantwell, D. P., & Baker, L. (1980). Psychiatric and behavioral characteristics of children with communicative disorders. Journal of Pediatric Psychology, 5, 161-178.

Cantwell, D. P., Baker, L., & Mattison, R. E. (1979). The prevalence of psychiatric disorder in children with speech and language disorder: An epidemiologic study. Journal of the American Academy of Child Psychiatry, 18, 450-461.

Chaffee, C. A., Cunningham, C. E., Secord-Gilbert, M., Elbard, H., & Richards, J. (in press). Screening effectiveness of the Minnesota Child Development Inventory expressive and receptive language scales: Sensitivity, specificity, and predictive value. Psychological As- sessment: A Journal of Consulting and Clinical Psychology.

Christensen, A., Phillips, S., Glascow, R. E., & Johnson, S. M. (1983). Parental characteristics and interactional dysfunction in families with child behavior problems: A prelimi- nary investigation. Journal of Abnormal Child Psychology, 11, 153-166.

Colligan, R. C. (1976). Prediction of kindergarten reading success from preschool report of parents. Psychology in the Schools, 13, 304-308.

Colligan, R. C. (1977). The Minnesota Child Development Inventory as an aid in the assessment of developmental disability. Journal of Clinical Psychology, 33, 162-163.

Colligan, R. C. (1981). Prediction of reading difficulty from parental preschool report. Learn- ing Disability Quarterly, 4, 31-37.


Cunningham, C. E., Benness, B. B., & Siegel, L. S. (1988). Family functioning, time alloca- tion, and parental depression in the families of normal and ADDH children. Journal of Clinical Child Psychology, 17, 169-177.

Dean, R. S., & Steffen, J. E. (1984). Direct and indirect pediatric screening measures. Journal of Pediatric Psychology, 9, 65-75.

Eisert, D. C., Spector, S., Shankaran, S., Faigenbaurn, D., & Szego, E. (1980). Mother's report of their low birth weight infants' subsequent development of the Minnesota Child De- velopment Inventory. Journal of Pediatric Psychology, 5, 353-364.

Estroff, T. W., Herrera, C., Gaines, R., Shaffer, D., Gould, M., & Green, A. H. (1984). Maternal psychology and perception of child behavior in psychiatrically differed and child maltreat- ment families. Journal of the American Academy of Child Psychiatry, 23, 649-652.

Fergusson, D. M., Horwood, L. J., Gretton, M. E., & Shannon, F. T. (1985). Family life events, maternal depression, and maternal and teacher descriptions of child behavior. Pediatrics, 75, 30-35.

Forehand, R., Wells, K. C., McMahon, R. J., Griest, D., & Rogers, T. (1982). Maternal perceptions of maladjustment in clinic-referred children: An extension of earlier research. Journal of Behavioral Assessment, 4, 145-151.

Gottfried, A. W., Guerin, D., Spencer, J. E., & Meyer, C. (1984). Validity of Minnesota Child Development Inventory in screening young children's developmental status. Journal of Pediatric Psychology, 9, 219-229.

Griest, D., Wells, K. C., & Forehand, R. (1979). An examination of predictors of material perceptions of maladjustment in clinic-referred children. Journal of Abnormal Psycholo- gy, 88, 277-281.

Guerin, D., & Gottfried, A. W. (1987). Minnesota Child Development Inventory: Predictors of intelligence achievement and adaptability. Journal of Pediatric Psychology, 12, 595-609.

Ireton, H. R., & Thwing, E. J. (1972). Minnesota Child Development Inventory. Minneapolis: Behavior Science Systems.

Ireton, H., Thwing, E., & Currier, S. (1977). Minnesota Child Development Inventory: Iden- tification of children with developmental disorders. Journal of Pediatric Psychology, 2, 18-22.

Kenny, T. J., Hebel, J. R., Sexton, M. J., & Fox, N. L. (1987). Developmental screening using parent report. Developmental and Behavioral Pediatrics, 8, 8-11.

Lachar, D., Kline, R. B., & Gdowski, C. L. (1987). Respondent psychopathology and interpre- tive accuracy of the Personality Inventory for Children: The evaluation of a "most reasona- ble" assumption. Journal of Personality Assessment, 51, 165-177.

Mash, E. J., & Terdal, L. G. (1988). Behavioral assessment of child and family disturbance. In E. J. Mash & L. G. Terdal (Eds.), Behavioral assessment of childhood disorders (pp. 3-65). Behavioral assessment of childhood disorders (pp. 3-65). New York: Guilford Press.

Panaccione, V. F., & Wahler, R. G. (1986). Child behavior, maternal depression, and social coercion as factors in the quality of child care. Journal of Abnormal Child Psychology, 14, 263-278.

Pedhazur, E. J. (1982). Multiple regression in behavioral research: Explanation and predietion (2nd ed.). Toronto: CBS College Publishing.

Reynell, J. (1977). ReynellDevelopmental Language Scales (Revised). Windsor, Ontario: NFER Publishing.

Richters, J., & Pellegrini, D. (1989). Depressed mothers' judgments about their children: An examination of the depression-distortion hypothesis. Child Development, 60, 1068-1075.

Rickard, K. M., Forehand, R., Wells, K. C., Griest, D. L., & McMahon, R. J. (1981). Factors in the referral of children for behavioral treatment: A comparison of mothers of clinic- referred deviant, clinic-referred non-deviant and non-clinic children. Behavior Research and Therapy, 19, 201-205.

Saylor, C. F., & Brandt, B. J. (1986). The Minnesota Child Development Inventory: A valid maternal-report form for assessing development in infancy. Developmental and Behavioral Pediatrics, 7, 308-311.

Schaughency, E. A., & Lahey, B. B. (1985). Mothers' and fathers' perceptions of child deviance: Roles of child behavior, parental depression, and marital satisfaction. Journal of Counsulting and Clinical Psychology, 53, 718-723.


Silva, P. A., Williams, S., & McGee, R. (1987). A longitudinal study of children with developmental language delay at age three: Later intelligence, reading, and behaviour problems. Developmental Medicine and Child Neurology, 29, 630-640.

Sobol, M. P., Ashbourne, D. T., Earn, B. M., & Cunningham, C. E. (1989). Parents' attribu- tions for achieving compliance from attention-deficit-disordered children. Journal of Ab- normal Child Psychology, 17, 359-369.

Stevenson, J., & Richman, N. (1978). Behavior, language, and development in three-year-old children. Journal of Autism and Childhood Schizophrenia, 83, 299-313.

Sturner, R. A., Funk, S. G., Thomas, P. D., Green, J. A. (1982). An adaptation of the Min- nesota Child Development, J. A. (1982). An adaptation of the Minnesota Child Develop- ment Inventory for preschool developmental screening. Journal of Pediatric Psychology, 7, 295-306.

Szatmari, P., Offord, D. R., & Boyle, M. H. (1989). Ontario Child Health Study: Prevalence of attention deficit disorder with hyperactivity. Journal of ChiM Psychology and Psy- chiatry, 30, 219-230.

Webster-Stratton, C. (1988). Mothers' and fathers' perceptions of child deviance: Roles of parent and child behaviors and parent adjustment. Journal of Consulting and Clinical Psychology, 56, 909-915.

the influence of parenting stress and child behavior problems on parental estimates of expressive...

Documents