Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Aug 24.
Published in final edited form as: Dev Psychobiol. 2020 May 29;63(1):65–73. doi: 10.1002/dev.21985

Multi-Group Multi-Time Point Confirmatory Factor Analysis of the Triadic Structure of Temperament: A Nonhuman Primate Model

Elizabeth K Wood a, J Dee Higley a, Maribeth Champoux b, Michael Marsiske c, Joseph A Olsen d, Stephen J Suomi e, Daniel B Kay a
PMCID: PMC9398891  NIHMSID: NIHMS1597623  PMID: 32469093

Abstract

Attempts to describe the latent structure of human infant temperament have led some to suggest the existence of three major dimensions. An earlier exploratory factor analysis (EFA) supported a triadic structure of temperament in week-old rhesus monkey infants, paralleling the structure in human infants. This study sought to confirm the latent triadic structure of temperament across the first month of life in a larger sample of rhesus monkey infants (N=668), reared by their mothers or in a neonatal nursery. A weekly behavioral assessment was obtained during the first month of life using a subset of items from the widely utilized Infant Behavioral Assessment Scale (IBAS), an instrument designed to measure temperament in infant monkeys. Using the latent constructs proposed by the earlier EFA (Orienting/Regulation, Surgency/Extraversion, Negative Affectivity), multi-group, multi-time point confirmatory factor analyses were conducted to confirm the latent temperament structure across rearing groups at each time point (weeks 1–4). Results confirm and extend those of the earlier EFA: latent Orienting/Regulation, Surgency/Extraversion, and Negative Affectivity constructs were present across the rearing groups at each time point, with the IBAS items consistently loading onto the latent factors to a similar degree across rearing groups at each time point. These findings suggest foundational evolutionary roots for the triadic structure of human infant temperament, but that its behavioral manifestations vary across maturation and rearing condition. Similarities in latent temperament structure in humans and a representative nonhuman primate highlights the potential for utilizing translational nonhuman primate models to increase understanding of human temperament.

Keywords: Early development, Infant temperament, Multi-group longitudinal confirmatory factor analysis, Multi-group multi-time confirmatory factor analysis, Neonatal behavioral assessment scale, Rhesus monkey, Temperament, Triadic structure of temperament

Temperament is a cluster of constitutional traits that underlie stable individual differences in behavior. It is one of the first quantifiable predictors of developmental outcomes (Goldsmith et al., 1987), predicting both ongoing and future mental and physical health (e.g., Boyce & Ellis, 2005; Caspi et al., 1997; Kern & Friedman, 2008; Rettew & McKee, 2005). Temperament is widely believed to have a latent structure involving multiple dimensions that are most obvious in early development and which uniquely influence behavior throughout the lifespan (Rothbart, 2007; Rothbart, Derryberry, & Posner, 1994). An evolutionary basis for temperament has long been assumed and there is increasing interest in determining whether there is cross-species agreement in the latent structure of temperament, particularly between human and nonhuman primates. Preliminary evidence suggests strong cross-species parallels in human and nonhuman primate temperament and personality (Chamove, Eysenck, & Harlow, 1972; Kay, Marsiske, Suomi, & Higley, 2010; King & Figueredo, 1997; King, Weiss, & Farmer, 2005; Uher & Asendorpf, 2008; Weinstein & Capitanio, 2008). Establishing the stability of temperament structure in these animals and meaningful agreement between the underlying latent structure and behavioral manifestations of temperament in human and rhesus monkey (Macaca mulatta) infants may improve the usefulness and translational power of temperament research.

Advances in theory, measurement, and statistical techniques have been crucial to establishing the latent structure of human temperament (Goldsmith et al., 1987). Theoretical advances in temperament burgeoned after Thomas, Chess, and Birch (1968) proposed three temperament types based on infant ease of adapting to the social and physical environment. Buss and Plomin (1975) later refined temperament assessment by establishing criteria for behavioral markers of temperament, specifying that the observable trait must be heritable, stable across the lifespan, and have clear evolutionary origins. Based on these criteria, they proposed three temperament dimensions in humans: emotionality, activity, and sociability. Later, advances in factor analytic techniques allowed the latent structure of temperament to take shape, with studies converging on a three-factor structure of human infant and child temperament (Ahadi, Rothbart, & Ye, 1993; Gartstein, Knyazev, & Slobodskaya, 2005; Gartstein & Rothbart, 2003; Rothbart, Ahadi, Hershey, & Fisher, 2001). Proposing that temperament represents underlying biological processes that account for individual differences in emotional and motivational behavior, Gartstein and Rothbart developed a psychobiological model of temperament and designed questionnaires to capture the broad dimensions of emotionality, activity, and attention that are manifest in observable behavior (Derryberry & Rothbart, 1988; Gartstein & Rothbart, 2003; Rothbart, 1981; Rothbart, Ahadi, & Evans, 2000; Rothbart & Bates, 2006). Factor analysis of these measures has converged on three major factors: Orienting/Regulation, Surgency/Extraversion, and Negative Affectivity. Of the various models of human infant temperament proposed, Gartstein and Rothbart’s (2003) psychobiological model is used as the framework in the present study, due to its wide-acceptance, being supported by factor analyses and cross-cultural validation in large samples (Gartstein et al., 2005).

Paralleling work done by human temperament researchers, translational researchers have sought to apply temperament theory, measurement, and statistical techniques to identify comparable temperament structures in nonhuman primates (Buirski, Kellerman, Plutchik, Weininger, & Buirski, 1973; Chamove et al., 1972; King & Figueredo, 1997; Locke, Locke, Morgan, & Zimmerman, 1964; Schneider & Suomi, 1992). Studies have identified three major factors of temperament in week-old (Kay et al., 2010), 3–5 month old (Weinstein & Capitanio, 2008), and 9–12 month old (Chamove et al., 1972) rhesus monkeys. Because rhesus monkeys mature approximately 3–4 times faster than humans (Roth et al., 2004), these studies provide preliminary support for a stable triadic structure of temperament during a maturational timeframe comparable to newborn-to-3-year-old humans. However, due to maturational and environmental influences on behavior, longitudinal studies are needed to determine the extent to which the specific behaviors representing specific temperament constructs vary over development and environmental conditions.

This study sought to further elucidate the latent structure of temperament of rhesus monkeys during early development, when temperament traits are thought to manifest most noticeably in dispositional behavior. The Infant Behavioral Assessment Scale (IBAS) is a widely used and standardized test of such behaviors in rhesus monkeys. Schneider and colleagues (1992) specifically designed the IBAS to have cross-species relevance by translating behavioral measures of human temperament and maturation into homologous measures of nonhuman primate behavior. Using only temperament-related items on the IBAS with clear conceptual comparison to human temperament, we previously showed that temperament in week-old rhesus monkeys involves three major factors (Kay et al., 2010), Orienting/Regulation, Surgency/Extraversion, and Negative Affectivity, that have meaningful agreement with the triadic model of temperament developed by Gartstein and Rothbart (Table 1). It is important to note that there are some differences in the behaviors thought to represent infant human and rhesus monkey temperament. For example, for human infants, the Surgency/Extraversion factor includes extraversion-related items, as well as activity-related items. For rhesus monkeys, which possess precocial motor abilities at birth, the Surgency/Extraversion factor includes only activity-related items, which are nevertheless thought to be related to temperament profiles of surgency and extraversion, as has been proposed in studies of personality in other nonhuman primates (King & Figueredo, 1997; King et al., 2005; Uher & Asendorpf, 2008). The present study sought to confirm the triadic structure of rhesus monkey infant temperament and to test whether the IBAS items consistently represent the temperament constructs during the first month of life in infants with differential early life experiences. Using multi-group multi-time confirmatory factor analyses (mmCFAs—also known as multi-group longitudinal CFAs), we hypothesized that the triadic latent structure of temperament would have good fit and adequate measurement invariance across four measurement occasions in first four weeks of life in infant monkeys with differential early rearing experiences.

Table 1.

Summary of the latent factors from EFAs of the triadic structure of human infant temperament using the IBQ-R (Gartstein & Rothbart, 2003) and of the triadic structure of rhesus monkey infant temperament using the IBAS (Kay et al., 2010).

Measure Latent Factor Items Description
Theoretical Domain: Attention
IBQ-R (Human) Orienting/ Regulation low-intensity pleasure Amount of pleasure or enjoyment related to low stimulus intensity, rate, complexity, novelty, and incongruity.
cuddliness Expression of enjoyment and molding of the body to being held by a caregiver.
duration of orienting Attention to and/or interaction with a single object for extended periods of time.
soothability Reduction of fussing, crying, or distress when soothing techniques are used by the caregiver.
IBAS (Monkey) Orienting/ Regulation visual orienting Eyes oriented toward toy (plastic Mickey Mouse face) held in 4 positions in infant’s periphery
attention Examiner rating of attention
visual following Eyes following moving toy in both horizontal and vertical directions
duration looking Examiner rating of length of looking at orienting items
Theoretical Domain: Activity
IBQ-R (Human) Surgency/ Extraversion approach Rapid approach, excitement, and positive anticipation of pleasurable activities.
vocal reactivity Amount of vocalization exhibited by the infant in daily activities.
high-intensity pleasure Pleasure or enjoyment related to high stimulation intensity, rate, complexity, novelty, and incongruity.
smiling and laughter Smiling or laughter during general caretaking and play.
activity level Gross motor activity, including movement of arms and legs, squirming and locomotor activity.
perceptual sensitivity Detection of slight, low-intensity stimuli from the external environment.
IBAS (Monkey) Surgency/ Extraversion motor activity (activeness) Observation of motor activity during the 5-minute observation period
spontaneous locomotion (activity level) Quality of locomotion
passivity (sensation seeking) Duration of time spent inactive
coordination Quality of movement
Theoretical Domain: Emotion
IBQ-R (Human) Negative Affectivity sadness Lowered mood and activity related to personal suffering, physical state, object loss, inability to perform a desired action; general low mood.
distress to limitation Fussing, crying, or showing distress while (a) in a confined place or position, (b) in caretaking activities, or (c) unable to perform a desired action.
fear Startle or distress to sudden changes in stimulation, novel physical objects or social stimuli; inhibited approach to novelty.
falling reactivity/rate of recovery from distress Rate of recovery from peak distress excitement, or general arousal; ease of falling asleep.
IBAS (Monkey) Negative Affectivity calm self (distress) Infant’s behavior when placed in an enclosed area (closed human infant incubator) for 5 minutes
response intensity (fear or hostility) Scored based on the quality of vocalizations during testing
irritability Amount of distress noted during the entire examination
predominant state (agitation) State of infant during examination
low tremulousness (nervousness) Examiner rating of tremulousness
vocalizations Number of vocalizations in a 1-minute time period
soothability (upset) The frequency in which soothing attempts of examiner were necessary to calm infant during testing
Open in a new tab

Note. For interpretation purposes, several items were relabeled to better highlight their similarity to human behaviors or to clarify their relation with respective factors by noting the appropriate antonym. These comparatively-adjusted labels are included in parenthesis in the table.

Methods

Subjects were 668 rhesus monkey infants (358 males, 310 females) housed at the Laboratory of Comparative Ethology, National Institute of Child Health and Human Development in Poolesville, Maryland, USA. As part of a larger research program aimed at assessing individual differences in maturation and temperament across time in rhesus monkeys reared in normative and impoverished early conditions, infants were randomly assigned to one of two rearing conditions at birth: mother-reared infants (MR; n = 307, 46% of the sample)—raised in conditions that approximated the natural social composition of the rhesus macaque (multiple females, infants, and adult rhesus monkey males; and nursery-reared infants (NR: n = 361, 54% of the sample)—separated from their mothers at birth and hand-raised in a neonatal nursery (Schneider, Moore, Suomi, & Champoux, 1991; Schneider & Suomi, 1992). All procedures were conducted in compliance with the Animal Care and Use Committee of the National Institutes of Health and in compliance with the US National Research Council’s Guide for the Care and Use of Laboratory Animals, the US Public Health Service’s Policy on Humane Care and Use of Laboratory Animals, and the Guide for the Care and Use of Laboratory Animals.

The IBAS was administered to each subject, once per week for the first month of life, with weeks 1, 2, 3 and 4 of life approximately comparable to human infant age at 3 weeks, 1.5 months, 2 months, and 3 months, respectively (Roth et al., 2004). It should be noted that rhesus monkey infants possess some developmental features that are more advanced than human infant abilities, even when the more rapid aging process is considered, for example, rhesus monkeys possess advanced motor abilities at birth, when compared to human infants. Infants were separated from their mothers (MR subjects) or removed from the neonatal nursery (NR subjects) for IBAS testing, which took place in a quiet room between 1000 and 1200 hours, lasting 20–30 minutes. Details of the IBAS have been previously described (Schneider, Moore, Suomi, & Champoux, 1991; Schneider & Suomi, 1992). Test battery items were administered and recorded by six technicians, who were trained by a senior scientist who oversaw the nursery (r >.90 inter-rater reliability). Yearly reliability checks were performed to confirm that this degree of inter-rater reliability was maintained. The items were administered in the same order for all infants during each administration. All items were rated on a 3-point Likert-type scale (range 0–2; half-scores allowed), with the exception of one item that was recorded on a continuous scale (i.e. number of distress vocalizations).

Data Analysis

Items from the IBAS were used to build the mmCFAs reported in this study. Included items were previously determined to be theoretically- and statistically-related to temperament in this study (Table 1). Prior to model analyses, indices of univariate normality were examined for each of the items. With the exception of one item (distress vocalizations), the IBAS items had a restricted range that rendered traditional normalizing techniques ineffective. In some cases, the items were relabeled to reflect their meaningful agreement and valence with the human literature. For example, the original IBAS item “soothability” represents the frequency in which soothing attempts of the examiner were necessary to calm the infant during testing. In contrast, “soothability” on the Infant Behavior Questionnaire-Revised (IBQ-R; Gartstein & Rothbart, 2003) represents a reduction of fussing, crying, or distress when soothing techniques are used by the caregiver. The former indicates a need for soothing (i.e., the infant was more upset); the latter indicates a tendency to be soothed (i.e., the infant was soothable). The previous analysis by Kay et al. (2010) found that the IBAS item “soothability” loaded onto the Negative Affectivity factor, supporting that this item represents the infant’s tendency for getting “upset”. See Table 1 for a full description of items with taxonomical modifications used in the mmCFA.

To confirm whether the triadic structure of temperament is present and stable in rhesus monkeys across the first four weeks of life and across two rearing conditions (MR and NR), measurement invariance was assessed for a series of multi-group and multi-time point CFA models (mmCFAs—also known as multi-group longitudinal CFAs) for each of the latent constructs. The data analysis plan consisted of three steps for assessing measurement invariance (Putnick & Bornstein, 2016; Widaman, Ferrer, & Conger, 2010): First, configural invariance was assessed to test whether the basic pattern of the latent factors is supported across all time points and rearing groups. In these unconstrained models, all factor loadings, intercepts, and the unique variability of each item were allowed to freely vary at all time points and rearing groups. Second, weak invariance (also known as metric invariance) was assessed to test whether the variances and covariances of the latent factors can be compared across all time points and rearing groups. In these models, each of the factor loadings were constrained to be equivalent at all time points and rearing groups, while the intercepts and the unique variability of each item were allowed to freely vary. The weak invariance models were compared to the configural invariance models to determine whether their fit was significantly different than the configural invariance model. Third, strong invariance (also known as scalar invariance) was assessed to test whether the means, variances, and covariances of the latent factors can be compared across all time points and rearing groups. In these models, the item intercepts are constrained to be equivalent across all time points and rearing groups and the unique variability of each item was allowed to freely vary. The strong invariance models were compared to the weak invariance models to determine whether their fit was significantly different than the weak invariance model. Strong invariance is meant to establish that the latent factors are similarly constituted across all time points and rearing groups, which is considered acceptable for comparing factor scores across all time points and rearing (Gregorich, 2006).

Maximum likelihood with robust standard errors (MLR) estimates were used to test the fit of each model. Models were evaluated on the basis of standard multivariate normality fit indices including Root Mean Square Error of Approximation (RMSEA), the Comparative Fit Index (CFI), and Standard Root Mean Square Residual (SRMR). RMSEA values ≤ 0.08 were used as the threshold of acceptable fit for determining whether or not to reject the model (Browne & Cudeck, 1993). Due to the complexity of the model, CFI values ≥ 0.90 were regarded as indicative of good fit (Hu & Bentler, 1999). SRMR values < 0.08 were considered a good fit (Hu & Bentler, 1999).

When comparing models, the change in the Satorra-Bentler chi-square (ΔSB χ2) values, a model fit statistic that is commonly used for nested models (Bryant & Satorra, 2012; Satorra & Bentler, 1988) and is considered the most reliable test statistic for evaluating nonnormal distributions (Byrne & Stewart, 2006), was evaluated. The change in CFI (ΔCFI), a model fit statistic that, unlike ΔSB χ2, is not sensitive to sample size and number of model constraints (Cheung & Rensvold, 2000), was also evaluated. ΔSB χ2 values with p > .05 were considered to indicate nonsignificant difference between the two models. ΔCFI ≤ 0.01 were considered to indicate a nonsignificant difference between the two models (Cheung & Rensvold, 2000). The Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) were also used to determine the best fitting model. AIC and BIC are used to evaluate the fit of models that do not meet the criteria of standard multivariate normality and are designed to penalize for non-parsimony, with an AIC penalty for each parameter estimated and a BIC penalty as sample size increases. When model comparisons are made, lower AIC and BIC values are typically regarded as characteristics of the best fitting model (Kenny, 2015). All reported results represent standardized solutions. All mmCFAs were conducted in Mplus, version 8.

Results

LCFAs

Table 2 shows model fit indices for each mmCFA model. Configural, weak, and strong models for the each of the factors were recursive and over-identified. The configural models for the Orienting and Negative Affect factors had RMSEA values ≤ 0.08, CFI values ≥ 0.90, and SRMR values < 0.08, suggesting that these models had good fit (Table 2). The configural model for the Surgency/Extraversion factor had an RMSEA value > 0.08, an SRMR > 0.08, and a CFI value < 0.90, suggesting that this model had poor fit. For all models, AIC values suggested that the configural invariance models were the best fitting models, while BIC values suggested that the weak invariance models had the best fit.

Table 2.

Summary of model fit indices.

Orienting/Regulation Factor (N = 668)
Model χ2 df p CFI RMSEA [90% CI] SRMR AIC BIC
Configural 215.95 148 < .0002 0.99 0.04 [0.03, 0.05] 0.04 11381.85 12084.52
Weak 270.64 169 < .00001 0.98 0.04 [0.03, 0.05] 0.06 11393.13 12001.21
Strong 711.01 190 < .00001 0.91 0.09 [0.08, 0.10] 0.10 11794.80 12308.28
Negative Affectivity Factor (N = 668)
Model χ2 df p CFI RMSEA [90% CI] SRMR AIC BIC
Configural 1482.59 604 < .00001 0.89 0.07 [0.06, 0.07] 0.08 38350.70 39539.84
Weak 1677.17 646 < .00001 0.88 0.07 [0.07, 0.07] 0.10 38484.31 39484.26
Strong 3000.78 691 < .00001 0.73 0.10 [0.10, 0.10] 0.22 39822.59 40619.85
Surgency/Extraversion Factor (N = 668)
Model χ2 df p CFI RMSEA [90% CI] SRMR AIC BIC
Configural 846.08 148 < .00001 0.82 0.12 [0.11, 0.13] 0.11 13727.19 14429.62
Weak 959.09 169 < .00001 0.81 0.12 [0.11, 0.13] 0.13 13820.70 14428.58
Strong 2759.14 190 < .00001 0.37 0.20 [0.20, 0.21] 0.37 15763.60 16276.91
Open in a new tab

Note. For each set of models, CFI values ≥ 0.90, RMSEA values ≤ 0.08, SRMR values < 0.08, and the lowest AIC and BIC values are bolded.

Table 3 shows the indices of the model comparisons used to investigate measurement invariance. As expected, ΔSBχ2 was significant across models, likely due its sensitivity to sample size and model constraints. Therefore, significant ΔSBχ2 values were disregarded as determiners of significant model differences. Comparisons of ΔCFI between each configural invariance and weak invariance mmCFA model resulted in ΔCFI ≤ 0.01 for each of the factors, suggesting evidence for weak invariance (Supporting Figures 1–3 and Supporting Tables 1–3). Comparisons of ΔCFI between each weak invariance and strong invariance mmCFA models resulted in ΔCFI > 0.01, indicating that strong invariance was not tenable and no further testing was conducted.

Table 3.

Summary of comparison indices of compared models.

Orienting/Regulation Factor (N = 668)
Model Comparison SBχ2diff df p ΔCFI
Configural to Weak 56.09 21 < .0001 0.01
Weak to Strong 437.15 21 < .0001 0.07
Negative Affectivity Factor (N = 668)
Model Comparison SBχ2 df p ΔCFI
Configural to Weak 184.86 42 < .0001 0.01
Weak to Strong 1153.00 45 < .0001 0.15
Surgency/Extraversion Factor (N = 668)
Model Comparison SBχ2 df p ΔCFI
Configural to Weak 113.70 21 < .0001 0.01
Weak to Strong 1604.35 21 < .0001 0.44
Open in a new tab

Note. ΔCFI ≤ 0.01 are bolded.

Discussion

Consistent with the hypothesis, the latent temperament factors thought to represent the triadic structure of temperament (i.e., Orienting/Regulation, Negative Affectivity, and Surgency/Extraversion) were confirmed for each time point and rearing group in rhesus monkey infants. The weak invariance findings indicate that, for each time point and rearing group, the IBAS items contribute to their respective latent factors to a similar degree and that the variances and covariances of the latent temperament factors can be compared. The strong noninvariance findings suggest that at least one of the item intercepts differs over the time points or between rearing groups and that the mean factor scores may not be directly comparable for each time point and rearing group. One explaination for this finding could be that items manifest to a greater or lesser degree for one of the time points or rearing groups without being related to actual differecnes in the latent factor but to maturational or environmental influences. For example, MR subjects could show greater visual orienting at week 2, but that increase in visual orienting may not be related to changes in the Orienting/Regulation factor. Overall, the triadic structure of temperament confirmed in this study is consistent with human infant temperament structure, suggesting an evolutionarily-old temperament pattern is present in rhesus monkey infants across the first four weeks of life, regardless of rearing context, further strengthening the translational value of the primate model to human temperament research.

This is the first major attempt to use mmCFA to assess the fit of the triadic structure of human infant temperament across time and in a diverse sample of very young rhesus monkey infants. This work is supported by CFA studies investigating the structure of temperament or personality in older rhesus monkeys (Capitanio, 1999; Capitanio & Widaman, 2005; Chamove et al., 1972; Weinstein & Capitanio, 2008; Weiss, Adams, Widdig, & Gerald, 2011) and by studies identifying comparable personality traits in humans and nonhuman primates (Weiss & King, 2015; King and Figueredo, 1997; Uher & Asendorpf, 2007). The confirmed triadic factor structure in a translational model may improve efforts to explore unanswered questions of temperament in the human literature, including its relationship to heritability, evolution, and individual differences. The results of this study also lend empirical justification for utilizing a translational rhesus monkey model to conduct longitudinal studies concerning infant temperament and later outcomes. The longitudinal relationship between temperament and other developmental behaviors (e.g., sleep) or biological substrates (e.g., cortisol) may be better understood through the lens of a comparative model of temperament with both a similar structure and meaningful agreement. Furthermore, the behavioral outcomes of these underlying latent traits may be better understood through the use of a translational model where environmental influences are held constant and known and where the infant develops at a more rapid rate and temperament outcomes can be assessed in a shorter time frame than in humans. The results of this study also provide a basis for studying temperament as a foundational developmental pathway for normative individual differences, as well as the risk for adverse developmental outcomes such as future anxiety and alcohol abuse. In line with this, future studies in human infants should assess for longitudinal invariance of the triadic structure of infant temperament (Garstein & Rothbart, 2003). Testing whether this structure captures temperament consistently across development would help to inform and clarify how the structure of temperament changes and remains stable throughout development. Given the literature on sex differences in human infant temperament development (for an example, see Miller et al., 2019), future studies should assess the whether these sex differences bear out in a rhesus monkey sample, and whether sex differences change or remain stable across development. Such analyses were beyond the scope of the present paper due to the lack of strong invariance Most importantly, the results provide evidence for an ancient, evolutionary-based triadic structure for human and nonhuman primate temperament.

There are some important limitations to this study that should be noted. First, results of this study did not support evidence of strong invariance, precluding the comparisons of the latent factor scores for each time point and rearing group. This limitation may be due to differences in the behavioral manifestations of that temperament factors as the infants age and between rearing groups. For example, we have observed infant rhesus monkeys exhibit repeated distress vocalizations while being tested in the Human Intruder Paradigm, an ecologically-meaningful test designed to elicit stress and anxiety in nonhuman primates regardless of subject age (Kalin & Shelton, 1989). Yet, adolescent monkeys tend to remain silent during the test. suggesting that behavioral manifestations of stress and anxiety may not be consistent across age groups. Another limitation is differences in the specific behavioral manifestations of temperament inhuman and rhesus monkey infant. For example, Gartstein and Rothbart (2003) conceive of the Surgency/Extraversion factor as having positively-valanced items (Table 1), in addition to the motor activity items. The precocial motor skills of the rhesus monkey constitute a useful behavioral manifestation of temperament but may not capture the full range of positively-valanced temperament items associated with Surgency/Extraversion assessed in human infants, such as smiling, cuddling with caregiver, and high intensity pleasure. Nevertheless, motoric abilities and expression may indeed be indicative of temperament profiles of surgency and extraversion, consistent with what has been proposed in studies of personality in great apes (King & Figueredo, 1997; King et al., 2005; Uher & Asendorpf, 2008). Consistent with this, other work evaluating temperament in infant rhesus monkeys shows that infants rated as behaving slowly had fewer social relationships later in life (Weinstein & Capitanio, 2008), suggesting that they were low on the Extraversion/Surgency factor, an interpretation that is consistent with human studies showing that activity levels in childhood are correlated with later extraversion (Hagekull & Bohlin, 1998). This interpretation is corroborated by research considering passivity an indication of anxiety and fear in the rhesus monkey (Spinelli et al., 2007), with infants showing increased passivity during a stressful social separation. It is possible that rhesus monkeys that do not exhibit passivity or that engage in low rates of passivity are exhibiting aspects of surgency and extraversion for a rhesus monkey. Attempts should be made to refine rhesus monkey temperament assessment tools so that behavioral items more fully capture the full range of comparative behaviors associated with temperament seen in humans (such as cuddliness with mother (MR monkeys) or with a blanket (NR monkeys)) and also show stability in their contribution to their respective latent temperament factor across time and in diverse samples.

Studies show that there are important long-term consequences for each of these temperament dimensions. For example, negative affect in human infancy is negatively predictive of effortful control in toddlers and positively predictive of aggression (Streit, Carlo, Ispa, & Palermo, 2017) and behavioral problems and symptoms of internalizing disorders in school-age children. Studies suggest that infant scores on the Orienting/Reactivity factor are positively predictive of effortful control in toddlerhood (Erickson, Gartstein, & Beauchaine, 2017). Others show the link between infant surgency and extraversion to feeding habits and body growth, which are thought to have implications for impulsivity (Burton et al., 2011; Vollrath, Tonstad, Rothbart, & Hampson, 2011). There are difficulties associated with conducting longitudinal studies assessing infant temperament and later life outcomes in humans. The similarities in human infant and rhesus monkey temperament structure, coupled with the experimental control that can be maintained in a rhesus monkey sample and their more rapid rate of aging, suggests that rhesus monkeys have high utility as translational longitudinal models.

In summary, the current study shows that the rhesus monkey infants and human infants may share a similar underlying temperament structure. These findings suggest foundational evolutionary roots for the triadic structure of human infant temperament in early development, but that its behavioral manifestations may vary across maturation and rearing condition. A similar latent temperament structure in humans and a representative nonhuman primate highlights the potential for increased understanding of human temperament via the use of translational nonhuman primate models.

Supplementary Material

Supp info

Acknowledgements:

The authors wish to thank Courtney Shannon and Angela Ruggiero for their contributions to data collection and data summary. This research was conducted with intramural funds from the National Institutes of Health and with small mentoring grants from Brigham Young University.

Footnotes

Conflict of Interest: None of the authors have a conflict of interest to declare.

Data Availability Statement:

The data that support the findings of this study are available from the corresponding author upon reasonable request and with permissions from the National Institutes of Health.

References

  1. Ahadi SA, Rothbart MK, & Ye R. (1993). Children’s temperament in the US and China: Similarities and differences. European Journal of Personality, 7, 359–377. [Google Scholar]
  2. Boyce WT, & Ellis BJ (2005). Biological sensitivity to context: I. An evolutionary-developmental theory of the origins and functions of stress reactivity. Development and Psychopathology, 17(2), 271–301. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/16761546 [DOI] [PubMed] [Google Scholar]
  3. Browne MW, & Cudeck R. (1993). Alternative ways of assessing model fit. Sage Focus Editions, 154, 136–136. [Google Scholar]
  4. Bryant FB, & Satorra A. (2012). Principles and practice of scaled difference chi-square testing. Structural Equation Modeling: A Multidisciplinary Journal, 19(3), 372–398. [Google Scholar]
  5. Buirski P, Kellerman H, Plutchik R, Weininger R, & Buirski N. (1973). A field study of emotions, dominance, and social behavior in a group of baboons (Papio anubis). Primates, 14(1), 67–78. [Google Scholar]
  6. Burton P, Wells J, Kennedy K, Nicholl R, Khakoo A, & Fewtrell M. (2011). Association between infant correlates of impulsivity-surgency (extraversion)-and early infant growth. Appetite, 57(2), 504–509. [DOI] [PubMed] [Google Scholar]
  7. Buss AH, & Plomin R. (1975). A temperament theory of personality development. New York: John Wiley & Sons. [Google Scholar]
  8. Byrne BM, & Stewart SM (2006). Teacher’s corner: The MACS approach to testing for multigroup invariance of a second-order structure: A walk through the process. Structural Equation Modeling, 13(2), 287–321. [Google Scholar]
  9. Capitanio JP (1999). Personality dimensions in adult male rhesus macaques: Prediction of behaviors across time and situation. American Journal of Primatology, 47, 299–320. [DOI] [PubMed] [Google Scholar]
  10. Capitanio JP, & Widaman KF (2005). Confirmatory factor analysis of personality structure in adult male rhesus monkeys (Macaca mulatta). American Journal of Primatology, 65(3), 289–294. doi: 10.1002/ajp.20116 [DOI] [PubMed] [Google Scholar]
  11. Caspi A, Begg D, Dickson N, Harrington H, Langley J, Moffitt TE, & Silva PA (1997). Personality differences predict health-risk behaviors in young adulthood: evidence from a longitudinal study. Journal of Personality and Social Psychology, 73(5), 1052–1063. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/9364760 [DOI] [PubMed] [Google Scholar]
  12. Chamove AS, Eysenck HJ, & Harlow HF (1972). Personality in monkeys: Factor analyses of rhesus social behaviour. The Quarterly Journal of Experimental Psychology, 24(4), 496–504. doi: 10.1080/14640747208400309 [DOI] [PubMed] [Google Scholar]
  13. Cheung GW, & Rensvold RB (2000). Assessing extreme and acquiescence response sets in cross-cultural research using structural equations modeling. Journal of Cross-Cultural Psychology, 31(2), 187–212. [Google Scholar]
  14. Derryberry D, & Rothbart MK (1988). Arousal, affect, and attention as components of temperament. Journal of Personality and Social Psychology, 55(6), 958. [DOI] [PubMed] [Google Scholar]
  15. Erickson NL, Gartstein MA, & Beauchaine TP (2017). Infant predictors of toddler effortful control: A multi-method developmentally sensitive approach. Infant and Child Development, 26(2), e1971. [Google Scholar]
  16. Gartstein MA, Knyazev GG, & Slobodskaya HR (2005). Cross-cultural differences in the structure of infant temperament: United States of America (U.S.) and Russia. Infant Behavior & Development, 28(1), 54–61. [Google Scholar]
  17. Gartstein MA, & Rothbart MK (2003). Studying infant temperament via the Revised Infant Behavior Questionnaire. In (Vol. 26, pp. 64–86). Infant Behavior and Development. [Google Scholar]
  18. Goldsmith HH, Buss AH, Plomin R, Rothbart MK, Thomas A, Chess S, . . . McCall RB (1987). Roundtable: What is temperament? Four approaches. Child Development, 58(2), 505–529. doi: 10.2307/1130527 [DOI] [PubMed] [Google Scholar]
  19. Gregorich SE (2006). Do self-report instruments allow meaningful comparisons across diverse population groups? Testing measurement invariance using the confirmatory factor analysis framework. Medical Care, 44(11 Suppl 3), S78. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hagekull B, & Bohlin G. (1998). Preschool temperament and environmental factors related to the five-factor model of personality in middle childhood. Merrill-Palmer Quarterly, 44(2), 194–215. [Google Scholar]
  21. Hu LT, & Bentler PM (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. [Google Scholar]
  22. Kalin NH, & Shelton SE (1989). Defensive behaviors in infant rhesus monkeys: Environmental cues and neurochemical regulation. Science, 243(4899), 1718–1721. [DOI] [PubMed] [Google Scholar]
  23. Kay DB, Marsiske M, Suomi SJ, & Higley JD (2010). Exploratory factor analysis of human infant temperament in the rhesus monkey. Infant Behavior & Development, 33(1), 111–114. doi: 10.1016/j.infbeh.2009.11.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kenny DA (2015, November 24). Measuring model fit. http://www.davidakenny.net/cm/fit.htm
  25. Kern ML, & Friedman HS (2008). Do conscientious individuals live longer? A quantitative review. Health Psychology, 27(5), 505–512. doi: 10.1037/0278-6133.27.5.505 [DOI] [PubMed] [Google Scholar]
  26. King JE, & Figueredo A. (1997). The five-factor model plus dominance in chimpanzee personality. Journal of Research in Personality, 31(2), 257–271. [Google Scholar]
  27. King JE, Weiss A, & Farmer KH (2005). A chimpanzee (Pan troglodytes) analogue of cross-national generalization of personality structure: Zoological parks and an African sanctuary. Journal of Personality, 73(2), 389–410. doi: 10.1111/j.1467-6494.2005.00313.x [DOI] [PubMed] [Google Scholar]
  28. Locke KD, Locke EA, Morgan GA, & Zimmerman RR (1964). Dimensions of social interactions among infant rhesus monkeys. Psychological Reports, 15(2), 339–349. [Google Scholar]
  29. Miller NV, Degnan KA, Hane AA, Fox NA, & Chronis-Tuscano A. (2019). Infant temperament reactivity and early maternal caregiving: Independent and interactive links to later childhood attention-deficit/hyperactivity disorder symptoms. Journal of Child Psychology and Psychiatry, 60(1), 43–53. doi: 10.1111/jcpp.12934 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Putnick DL, & Bornstein MH (2016). Measurement invariance conventions and reporting: The state of the art and future directions for psychological research. Developmental Review, 41, 71–90. 10.1016/j.dr.2016.06.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Rettew DC, & McKee L. (2005). Temperament and its role in developmental psychopathology. Harvard Review of Psychiatry, 13(1), 14–27. doi: 10.1080/10673220590923146 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Roth GS, Mattison JA, Ottinger MA, Chachich ME, Lane MA, & Ingram DK (2004). Aging in rhesus monkeys: Relevance to human health interventions. Science, 305(5689), 1423–1426. doi: 10.1126/science.1102541 [DOI] [PubMed] [Google Scholar]
  33. Rothbart MK (1981). Measurement of temperament in infancy. Child Development, 52, 569–578. [Google Scholar]
  34. Rothbart MK, Ahadi SA, & Evans DE (2000). Temperament and personality: Origins and outcomes. Journal of Personality and Social Psychology, 78(1), 122–135. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/10653510 [DOI] [PubMed] [Google Scholar]
  35. Rothbart MK, Ahadi SA, Hershey KL, & Fisher P. (2001). Investigations of temperament at three to seven years: the Children’s Behavior Questionnaire. Child Development, 72(5), 1394–1408. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/11699677 [DOI] [PubMed] [Google Scholar]
  36. Rothbart MK & Bates JE (2006). Temperament. In Damon W. & Lerner RM (Eds.), Handbook of child psychology, social, emotionally, and personality development (6th ed., Vol. 3). John Wiley & Sons. [Google Scholar]
  37. Rothbart MK (2007). Temperament, development, and personality. Current Directions in Psychological Science, 16(4), 207–212. [Google Scholar]
  38. Rothbart MK, Derryberry D, & Posner MI (1994). A psychobiological approach to the development of temperament. In Bates JE & Wachs TD (Eds.), APA science Vols. Temperament: Individual differences at the interface of biology and behavior (p. 83–116). American Psychological Association. 10.1037/10149-003 [DOI] [Google Scholar]
  39. Satorra A, & Bentler P. (1988). Scaling corrections for statistics in covariance structure analysis. ASA 1988 Proceedings of the Business and Economic Statistics Section, 308–313. Alexandria, VA: American Statistical Association. [Google Scholar]
  40. Schneider ML, Moore CF, Suomi SJ, & Champoux M. (1991). Laboratory assessment of temperament and environmental enrichment in rhesus monkey infants (Macaca mulatta). American Journal of Primatology, 25, 137–155. doi: 10.1002/ajp.1350250302 [DOI] [PubMed] [Google Scholar]
  41. Schneider ML, & Suomi SJ (1992). Neurobehavioral assessment in rhesus monkey neonates (Macaca mulatta): Developmental changes, behavioral stability, and early experiences. Infant Behavior and Developmental, 15, 155–177. doi: 10.1016/0163-6383(92)80021-L [DOI] [Google Scholar]
  42. Spinelli S, Schwandt ML, Lindell SG, Newman TK, Heilig M, Suomi SJ, Higley JD, Goldman D, & Barr CS (2007). Association between the recombinant human serotonin transporter linked promoter region polymorphism and behavior in rhesus macaques during a separation paradigm. Developmental Psychopathology, 19, 977–987. doi:S095457940700048X [pii] 10.1017/S095457940700048X [DOI] [PubMed] [Google Scholar]
  43. Streit C, Carlo G, Ispa JM, & Palermo F. (2017). Negative emotionality and discipline as long-term predictors of behavioral outcomes in African American and European American children. Developmental Psychology, 53(6), 1013. [DOI] [PubMed] [Google Scholar]
  44. Thomas A, Chess S, & Birch HG (1968). Temperament and Behavior Disorders in Children. New York: New York University Press. [Google Scholar]
  45. Uher J, & Asendorpf JB (2008). Personality assessment in the Great Apes: Comparing ecologically valid behavior measures, behavior ratings, and adjective ratings. Journal of Research in Personality, 42(4), 821–838. [Google Scholar]
  46. Vollrath ME, Tonstad S, Rothbart MK, & Hampson SE (2011). Infant temperament is associated with potentially obesogenic diet at 18 months. International Journal of Pediatric Obesity, 6(sup3), e408–414. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Weinstein TA, & Capitanio JP (2008). Individual differences in infant temperament predict social relationships of yearling rhesus monkeys (Macaca mulatta). Animal Behavior, 76(2), 455–465. doi: 10.1016/j.anbehav.2008.01.024 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Weiss A, Adams MJ, Widdig A, & Gerald MS (2011). Rhesus macaques (Macaca mulatta) as living fossils of hominoid personality and subjective well-being. Journal of Comparative Psychology, 125(1), 72–83. doi: 10.1037/a0021187 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Widaman KF, Ferrer E, & Conger RD (2010). Factorial invariance within longitudinal structural equation models: Measuring the same construct across time. Child Development Perspectives, 4(1), 10–18. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp info
NIHMS1597623-supplement-Supp_info.docx (780.2KB, docx)

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request and with permissions from the National Institutes of Health.

RESOURCES