Psychological Reports, 1990, 66, 899-904. @ Psychological Reports 1990
PERSONALITY VARIABLES AS ASSESSED BY T H E MMPI AND THEIR RELATIONSHIP TO MOTHER-INFANT INTERACTIONAL BEHAVIORS AT AGE THREE MONTHS ' MARK G. KAELLER AND KIKI V. ROE University of California, Lor Angekr
Summary.-The present study examined whether maternal personality variables, as assessed on the MMPI, relate to early infant and maternal variables. The subjects were 14 low-risk, preterm and 24 full-term 3-mo.-old infants and their mothers. A female experimenter visited each infant at home where the mother and the experimenter each took turns and talked to the baby for 3 min. Subsequently, each infant was given a differential vocal response score by taking the amount of time spent in nondistress vocalizing in response to the stranger from the time spent in similar vocalizing in response to the mother. Following the 3-min. interactions, the mother and infant were observed for 30 min. and several mother-infant behaviors were recorded. Finally, the mothers were administered the MMPI. Analysis showed that hgh scores on Scales 3, 6, and Ego Control were related to less favorable mother-infant interactions and high scores on the Femininity and Ego Control Scales were related to infants' low differential vocal response scores.
During the last 20 years a large amount of research has focused on infant development and related variables. Yet, very little of this research has concerned the relationship between mothers' personalities and behaviors towards the infants and/or the infants' behaviors. This holds in spite of the fact that several studies suggest that maternal personahty variables, as assessed on the MMPI, relate to maternal behaviors towards older children. For example, Paulson, Afifi, Thomason, and Chaleff (1974) showed that mothers who had a 4-9-6 MMPI profile were abusive towards their children. Recently, Diskin and Heinicke (1986) found that the Ego Control, Ego Strength and Warmth scales, as assessed by the MMPI, related to maternal behaviors towards the infant. Furthermore, Heinicke, Diskin, Ramsey-Klee, and Given (1983) suggest that the mother's MMPI index of Ego Strength may relate to the infant's soothability as well as the mother's affectionate responsiveness. O n the other hand, Block (1953) asserts that mothers scoring h g h on the Ego Control Scale of the MMPI may be over-controhng and rigid and show constriction of emotional expression, personality characteristics which are expected to interfere with optimal mothering. Finally, Finney (1965), who developed the MMPI Warmth scale, suggests that women scoring high on the Warmth Scale have higher interpersonal satisfaction and high sensitivity towards others, qualities that are essential for good mothering. The objective of the present study was to explore whether maternal per'Address correspondence to M. G. Kaeller, 22610 Liberty Bell Road, Woodland Hills, CA 91364-5718.
M. G. KAELLER & K.V. ROE
sonality variables, as measured on the MMPI, relate to both maternal behaviors towards the infant as well as infants' behaviors early in life. The measured maternal behaviors during natural sessions were maternal vocalizing to the infant, holding and cuddling of the infant, and mutual smiling and eye-contact between mother and infant at age 3 mo. The infant behavior chosen for the study was the infant's differential vocal response (DVR) to mother vs a stranger at 3 mo. This behavior has been (Roe, 1978; Roe, McClure, & Roe, 1982) related to children's later verbal-cognitive and academic functioning up to the age of 12 yr. as well as discriminated between groups of infants who are expected to show differences in their later cognitive development (Roe & Bronstein, 1988). T h s infant behavior is also believed to index the quality of the early mother-infant relationship (Roe, Drivas, Karagellis, & Roe, 1985; Roe, 1987). Specifically, it was expected that mothers who had T scores above 6 0 on any of the following MMPI scales would show less than optimal behaviors towards their infants: 3 (Hysteria), 4 (Psychopathy), 6 (Paranoia), 8 (Schizophrenia), and/or who had high 4-9 profiles and scored above 60 on the Ego Control Scale. In addition, the infants of such mothers were expected to show lower differential vocal response scores in comparison with infants of those mothers who did not have high scores on the above scales. The reverse was expected of mothers who scored high on the Warmth and Ego Strength Scales of the MMPI.
METHOD The study involved 38 English-speaking, white mothers and their first-born, 3-mo.-old infants. Fourteen of the infants were low-risk, preterm (8 boys and 6 girls) and 24 normal, full-term (11 boys and 13 girls). The gestational age of the preterm infants ranged from 32 to 36 wk., with a mean of 33.8 wk. and their birthweights ranged from 1,335 to 2,074 gm, with a mean of 1,866 gm. None of the preterm infants suffered from any serious medical complication. The full-term infants' birth weights ranged from 2,884 to 4,120 gm with a mean of 3,579 gm. They were all recruited from local hospital records. The mothers' ages ranged from 17 to 39 yr., with a mean of 26.7 yr., while their education ranged from high school diplomas to advanced professional degrees, with a mean of 15 yr. of education. The mothers answered the MMPI and their profiles were plotted using the K correction. Since the mothers were pulled from a nonpsychiatric normal population, their MMPI profiles were not elevated. For this reason we looked only at the 25% or more mothers who had T scores of at least 60 or higher. Because there were very few mothers who scored 60 or higher on Scales 1 (n = 2), 5 (n = I), 7 (n = 4) and 8 (n = 3), these scales were not used to make statistical analyses. However, since Scale 5, the Femininity/Masculinity scale, assesses some traditional feminine qualities which may be rele-
MATERNAL PERSONALITY, BEHAVIOR WITH INFANT
vant for mothering, it was decided to look at the 25% of mothers with the lowest scores on Scale 5, i.e., the most "feminine" mothers. I n a similar manner the following interpersonal style scales were evaluated: Ego Strength, Ego Control, and Warmth. When an infant was 3 mo. old (in the case of the preterm infants 3 mo. old from due date), a female experimenter visited the home. An effort was made to visit the homes in the morning and at a time when the baby had recently awakened, was fed, and was in a quiet, alert state. Upon arrival at the subject's home the baby was placed in an infant seat on top of a table and, after leaving the baby alone for 3 min. to get used to the situation, the mother talked to the infant for 3 min. from a distance of about 2 or 3 feet in a face-to-face fashon. The mother was instructed to talk to her baby in her natural manner without touching and trying to engage the infant in conversation." Subsequently the infant was left alone to rest for 2 min. and then the experimenter-stranger similarly interacted with the infant for another 3 min. If a baby cried during any of the interactional episodes, the session was stopped and the entire sequence was repeated at a later time or day when the baby was again in a quiet, alert state. All vocaLzations during these sessions were recorded and were coded a few months later by two independent coders. Both coders had received extensive training in the coding procedure to achieve high reliability before they were allowed to code actual subject recordings. Using a cumulative timer that recorded behaviors in rninutes and hundredths of a minute, the two coders Listened to the tapes and recorded the amount of time that the infants engaged in nondistress vocahzing in response to stimulation by the mother and the female stranger. Subsequently they assigned a differential vocal response (DVR) score to each subject by taking the time the infant spent in nondistress vocalizing in response to the mother minus the time spent in similar vocalizing in response to the stranger. The order of the mother and stranger presentation was not counterbalanced in this study since it has been shown (Roe, et al., 1985) that order of presentation of mother and stranger does not significantly affect 3-mo.-old infants' differential vocal response scores. After the 3-min. interactions with mother and stranger, the mother was asked to try to ignore the experimenter's presence and go about her business as usual. Subsequently the experimenter observed the mother and infant for 30 min. and logged the following mother-infant interactional behaviors each time they occurred during consecutive 10-sec. intervals: mother-infant eyecontact, mother-infant mutual smiling, and mother's holding/cuddling of the infant. Furthermore, all sounds during these naturalistic sessions were taperecorded and subsequently two independent trained coders listened to the tapes and recorded the cumulative amount of time that each mother spent I'
M. G. KAELLER
K. V. ROE
talking to her infant during the 30-min. sessions. Care was taken so the coder who coded the naturalistic behaviors was different from the one who had coded the infants' differential vocalization responses.
RESULTS Inter- and intracoder reliabhties for coding the amount of time that 3-mo.-old infants spend in nondistress vocalizing to mother and stranger have previously ranged from 0.83 to 0.93 (Roe, 1978, 1990) and for coding the various mother and mother-infant interactional behaviors during the naturahstic sessions currently ranged from .82 to .99. The 11 mothers who had the highest elevations on Scale 3 (T score M = 63.1, S D = 2.6 and range 60 to 68) spent significantly less time holding and cuddling their infants than other mothers. High Scale 3 mothers M holding = 57.1 and S D = 43.8 10-sec. intervals vs M = 89.0 and SD = 44.4 intervals for the rest of the mothers (t,, = 1.98, p < .O5). Similarly, the 10 mothers who scored high on Scale 6 (T score M = 63.8 and SD = 3.2, range from 60 to 70) spent significantly less time holding and cud&ng their infants than other mothers. High Scale 6 mothers held their babies for a M = 58.1 and SD = 36.0 10-sec. intervals vs a M = 89.4 and SD = 46.1 intervals for the rest of the mothers ( t = 2.02, p < .05). In addition, the 10 mothers who scored high on the Ego Control Scale (T score M = 64.4, SD = 4.5 and range 61 to 74) engaged in significantly less mutual s d n g with their babies than other mother-infant dyads (t = 1.74, p < .05; high Ego Control mothers had a M = 7.3, and SD = 7.2 of 10-sec. mutual smiling intervals with their infants vs a M = 13.4 and a S D = 10.4 for the rest of the mothers). It is interesting that the infants of these high Ego Control mothers showed significantly lower differential vocal response scores to their mothers vs the female stranger than other infants (t = 4.13, p < .05). Actually, these infants did not vocalize more to their mother than the stranger. This was also true (t = 2.49, p < .05, two-tailed) of the infants of the 10 most "ferninine" mothers, i.e., of the mothers who had the 25% lowest scores on Scale 5 (Masculinity/Femininity Scale, M = 34.0, S D = 6.8, range 20 to 40). Mothers who had above 60 T scores on Scale 4 (n = 13) and mothers who had a 4-9 profile (4 and 9 highest points, one of the two scales above T score 60, the other scale at least above T score of 55, n = 9) did not show significant differences from other mothers on the measured maternal behaviors, and their infants were not significantly different from other infants on their differential vocal response. Similarly, the 10 mothers who had the highest scores on the Warmth Scale (who endorsed at least 18 out of the 22 points of the scale) and the Ego Strength Scale (who endorsed at least 40 out of the 63 points of the scale) did not treat their infants differently from other mothers and their infants were not different from other infants on their differential vocal responses.
MATERNAL PERSONALITY, BEHAVIOR WITH INFANT
The purpose of the present study was twofold. First, we wanted to explore whether we could identify any maternal personality variables as assessed on the MMPI that could relate to mothering behaviors very early in life and, second, to examine whether such maternal personality variables relate to the young infants' vocal interaction with the mother vs a stranger. Analysis showed that mothers who scored above T score 60 on Scales 3 and 6 did not hold and cuddle their babies as much as other mothers and mothers who scored above T score 60 on the Ego Control scale did not engage in as much mutual smiling with their infants as other mothers. People with high scores on Scale 3 -and on the Ego Control scale are believed-to deny emotionality and to use excessive controls to cover up feelings. Similarly, people with high scores on Scale 6 tend to be hyperational and to avoid emotional involvement. All three scales associated with less desirable mothering in the present study appear to have in common a tendency to deny, cover up and/or control emotional involvement and expression. The fact that the infants of the high Ego Control mothers also did not vocalize more to the mother than the stranger further suggests that these mothers possibly had difficulty developing emotional bonds with their infants. The 3-mo.-old infant's ability to differentiate the mother from the female stranger and respond to her with heightened excitement and vocal output whde withholding such vocal output to the stranger is thought to reflect the quality of the infant-mother relationship and as such to be an indicator of the infantmother attachment. The finding that the infants of the (low Scale 5 ) very feminine scoring mothers also did not vocalize more to the mother than to the stranger was unexpected, and it is difficult to explain. Women with low Scale 5 scores are described as self-dissatisfied and self-distrusting, and this personality trait may have interfered with the development of an optimal mother-infant relationship. Another plausible explanation is that these very feminine women may be narcissistic and self-centered, which personality characteristics may have impeded the development of a strong emotional bond with the babies. We had expected that mothers who scored high on the Warmth and Ego Strength scales would show more affectionate behaviors towards their babies than other mothers. However, o m results did not support these expectations. We suspect the reason may be the fact that this group of normal, well-functioning mothers tended to have high scores on both of these scales. Such lack of heterogeneity probably precluded the finding of any significant relationships. I n summary, although the small sample calls for caution in the generalization of the results, the study is still interesting because it probes into an area of research that is still unexplored even though there is widespread
M. G . KAELLER & K. V. ROE
agreement that maternal personality variables are crucial for early infant development. REFERENCES BLOCK,J. (1953) The development of an MMPI based scale to measure ego control. (Mimeographed materials, Insticute of Penonality Assessment and Research, Univer. of California, Berkeley) DISKM, S. D., & HEMICKE,C. M. (1986) Maternal style of emotional expression: its description and relation to parent and infant behavior. Journal of Infant Behavior and Development, 7, 167-187. FINNEY,J. C. (1965) Development of a new set of MMPI scales. Psychological Reports, 17, 707-713. HEMCKE, C. M., DISKIN, S. D., RAMSEY-KLEE,D. M., & GWEN,K. (1983) Pre-birth parent characteristics and family development in the first year of life. ChiM Development, 54, 194-208. PAULSON,M., A m , A , , THOMASON, M., & CHALEFF,A. (1974) The MMPI: a descriptive measure of psychopathology in abusive parents. Journal of Clinical Psychology, 30, 387-390. ROE, K. V. (1978) Infants' mother-stranger discrimination at 3 months as a predictor of cognitive development at 3 and 5 years. Developmental Psychology, 14, 191-192. ROE, K. V, (1987) Planned vs. unplanned status of infant and vocal interaction with mother and stranger. Paper presented at the biennial meeting of the Society for Research i n Child Development, Baltimore, MD. ROE, K. V. (1990) Vocal interchange with mother and stranger as a function of infant age, sex and parental education. Early Childhood Research Quarterly, 5, 135-145. ROE, K. V., & BRONSTEIN,R. (1988) Maternal education and cognitive processing at three months as shown by the infants' response to mother vs. stranger. International Journal of Behavioral Development, 11, 389-393. ROE,K. V., DRIVAS,A,, KARAGELLIS, A,, & ROE, A. (1985) Sex differences in vocal interaction with mother and stranger in Greek infants: some cognitive implications. Devefopmental Psychology, 21, 372-377. ROE, K. V., MCCLURE,A., & ROE, A. (1782) Vocal interaction at 3 months and cognitive skills at 12 years. Developmental Psychology, 21, 372-377. ROE, K. V., ROE, A . , DRTVAS,A., & KARAGELLIS, A. (1788) Three month olds' vocal interaction with mother and stranger as a function of environment: some cognitive implications. In E. Hibbs (Ed.), Children and families. Madison, WI: International Univer. Press. Pp. 143-151.
Accepted April G, 1990.