Stability of temperament in infancy: parent-rated and observed measures

Abstract

Temperament refers to individual differences in reactivity and regulation and is a relatively stable and biologically based characteristic (Goldsmith et al., 1987). Infants also show variability in their responses to environmental factors. One child can be susceptible to reading the environmental signals, reactive to the environmental stimuli, and not easily be soothed in stressful situations. However, another child may not be sensitive to environmental cues, does not react to the environment in a loud fashion, and is easily soothed (Rothbart, 2007). Those innate characteristics may show relative stability through the infancy period. Different measurement techniques can be used to understand temperamental characteristics, such as parent-report measures and observations. Using different measurement technics may contribute to understanding both temperament itself and also stability of it. Therefore, the current study examines children's temperament characteristics over a year and compares parent-reported outcomes with the observed temperament. 96 three-24 months old infants were followed with a 4 months’ interval (Meanage = 11.93, SD = 6.07; nmale = 37, nfemale = 59). Their temperament was measured by the IBQ (Parent-rated: Gartstein & Rothbart, 2003) and gentle arm restrained task (LAB-TAB; Goldsmith & Rothbart,1996). Mothers filled the perceptual sensitivity, soothability, and falling reactivity/rate of recovery from distress sub-scales of IBQ, and two experimenters observed the child during the gentle arm restraint task. A composite score was created from the observed task. Bivariate correlational analysis showed that parent-reported and observed temperament score was not correlated with each other. However, all temperament dimensions were correlated with their subsequent scores at three-time points (see Table 1). Besides, multilevel modeling was run with HLM to see whether there is a change in time in all temperament scores. Temperament scores were nested to time. The time variable was included in the model as a Level 1 variable, while T1 age and gender of the children were included as a Level 2 variable (gender: 0=male, 1=male). Results showed that there was significant growth in perceptual sensitivity score of children (β = 0.39, p < 0.001) and anger/frustration score coming from gentle arm restrained task (B = 2.02, p< 0.05). The slope of the soothability and falling reactivity/recovery 0.20, p< .05), but it is steeper for males. Simple slopes of soothability were insignificant for males and females (see Figure 1). Overall, results showed continuity of temperament over a year, similar to recent findings (Planalp et al.,2017). However, parent-reported and observed temperament was not correlated with each other, opposite of recent findings (Moran et al.,2013; Planalp et al.,2017). The results will be discussed in light of the literature.