Abstract
The present longitudinal study repeatedly tested motor development in 345 infants at ages 3, 6, and 9 months in two eco-cultural contexts (German middle-class and Cameroonian Nso). To analyze differential motor developments and its co-determinants (parental education and growth in weight), person-centered analyses (hierarchical cluster analyses) were applied. Results indicate that cluster analyses of fine motor development scores led to two culture-mixed cluster groups. Four cluster groups were extracted in the gross motor domain showing differential growth curves. Differential growth curves were partly explained by education of mother and weight trajectories. Besides considering methodological aspects, inter-individual differences of intra-individual change within and across eco-cultural contexts are discussed.
Developmental processes are culture-specific manifestations of universal behavioral predispositions and developmental pathways constitute culture-specific solutions to universal developmental tasks (for a comprehensive review, see Keller, 2007; Keller & Kärtner, 2013). It is therefore crucially important to take the specific eco-cultural contexts which are associated with specific orientations toward the two basic human needs of autonomy and relatedness into account. Western middle-class and traditional farming communities can be regarded as two prototypically different eco-cultural contexts. Western middle-class communities typically live anonymous, large-scale social lives within money-based, highly educated, post-industrialized information societies. Individuals are characterized by personal traits such as uniqueness, independence, assertiveness, self-expression, and self-actualization. Psychological autonomy centrally organizes development. Reproduction starts late with only few children. Socialization strategies, best described as distal parenting, emphasize face-to-face interaction, vocal interaction, and object play (Keller, 2007) aiming at developing self-contained individuals, expressing their preferences and emotions, achieving their individual goals, and making their own decisions. In contrast, traditional farming communities are characterized by a subsistence-based economy, dense social network, many children, early reproduction, and a low level of formal school education. Relational adaptation is the fundamental organizer of development. Traditional farming families socialize their children to find their place within the social hierarchy, become respectful and obedient, and serve the well-being of the community. A proximal strategy of parenting with extensive body contact and body stimulation is associated with this cultural prototype (Keller, 2007). Whereas body contact promotes closeness, relatedness, and emotional warmth, body stimulation is functionally related to motor development. Many sub-Saharan African ethnotheories (e.g., the Yoruba concept of mobility and the Nso concept of stimulation) appreciate accelerated motor development, because children can start responsibility training (sending on errands, making their contribution to the viability of the family) once they are able to walk. Fostering gross motor development is therefore at the center of traditional farming parenting strategies.
Results of a recent cross-cultural longitudinal study (Vierhaus et al., 2011) showed that Cameroonian Nso infants had an accelerated gross motor development at ages 3 and 6 months compared with German middle-class infants. However, these differences equaled out at 9 months of age. Besides mean-based, cross-cultural differences in motor development (variable-centered approach), considerable inter-individual differences of intra-individual change also exist within the specific eco-cultural contexts (person-centered approach). As these individual developmental trajectories are not well understood yet, the present study analyzed motor development in a sample of German middle-class and Cameroonian Nso infants longitudinally (at 3, 6, and 9 months) from a differential psychological perspective using hierarchical cluster analysis (HCA).
As longitudinal assessments revealed that growth in length and weight during the first year of life as well as parental education are related to the achievement of motor development milestones in diverse cultural samples (e.g., Koutra et al., 2011), we considered parental education and growth in weight as important covariates to explain fine and gross motor growth trajectories related to differential development within and across cultures. Taken together, the present study focuses on (a) extracting developmental groups from the overall, non-culture separated, sample; (b) mapping differential motor development cluster groups to cultural groups; and (c) testing differences in parental education and growth trajectories between developmental cluster groups.
Method
Sample, Testing Materials, and Procedure
A sample of 345 (n = 73 Cameroonian Nso, n = 272 German) 3-month-old (born at term, healthy) infants participated. 1 Infants were reassessed at 6 and 9 months of age. Complete longitudinal data were available for 69 Nso (35 female, 34 male) and 198 German middle-class (97 female, 101 male) infants. Drop-out analyses indicated no significant differences between participating and dropping-out infants. Most Nso infants were recruited through infant welfare events organized by the local Health Centers. In Germany, infants were recruited through newspaper ads and invitation letters sent to parents of newborn children. Parents received monetary compensation for study participation. Infants were tested within a range of 2 days before and 10 days after the respective testing ages (3, 6, 9 months).
Bayley Scales of Infant Development III (Bayley, 2006), administered in line with the standard procedure (observer rating immediately during testing), were used for assessing fine motor (i.e., prehension, perceptual-motor integration, motor planning, and motor speed); rtt(3M) = .89; rtt(6M) = .82; rtt(9M) = .84, and gross motor development (i.e., movements of the limbs and torso related to static positioning and dynamic movement; rtt(3M) = .89; rtt(6M) = .89; rtt(9M) = .94. In Germany, the assessments were based on German versions of the test items. In Cameroon, trained local multilingual (Lamnso, English, Pidgin English) assistants presented the items in the local language Lamnso. Socio-demographic characteristics were evaluated by parent self-report on formal school education and age. Weight trajectories (3, 6, 9 months) were evaluated by parent self-report (parents relied on regular pediatrician infant screenings while reporting weights) in Germany and a weighting procedure in Cameroon.
Results
Socio-Demographic Variables and Weight Trajectories
Table 1 depicts the socio-demographic and birth weight differences between the rural Nso and German middle-class families.
Socio-Demographic and Birth Weight Differences Between the Rural Nso and German Middle-Class Families.
Data show that reflecting the different cultural contexts the rural Nso infants had more siblings and younger mothers than the German middle-class infants and the German parents had a higher level of formal school education measured by years of schooling compared with the Nso parents.
Cluster Analytic Procedure
Cluster analyses (z-transformed data, squared Euclidian distance, Ward-algorithm) were run with the total data set (irrespective of cultural groups). Clusters were determined by using both Elbow criterion (i.e., an additional cluster does not increase explained variance) and inspection of cluster dendrograms. After cluster determination, cluster groups were cross-tabulated with cultural groups. To explain differential growth trajectories within and across cultures, the extracted cluster groups were compared with respect to socio-demographic status and weight trajectories differences.
Fine Motor Development
Cluster analyses of the fine motor scale extracted two clusters with differential growth (see Figure 1). Cluster 1, consisting of 56.3% (n = 112) German middle-class infants and 87.1% (n = 61) rural Nso infants, shows lower fine motor performance across all three measurement points, F(1, 268) = 237, p < .001 (all Bonferroni corrected post hoc t tests, p < .05), than Cluster 2, consisting of 43.7% (n = 87) German middle-class infants and 12.9% (n = 9) rural Nso infants, χ2 = 21.6, df = 1, p < .001.
Growth-curves, cross-tabulation Cluster × Culture and birth-weight trajectories for gross and fine motor development of German middle-class and Cameroonian Nso infants.
Gross Motor Development
Cluster analyses of the gross motor scale extracted four clusters with differential growth patterns, F(1, 261) = 151, p < .001.The first cluster consists predominantly of German middle-class infants (47.7% German vs. 4.3% Nso infants) and shows the lowest gross motor performance across all three measurement occasions. The second cluster also consists predominantly of German middle-class infants (40.5% German vs. 12.9% Nso infants) and shows second lowest performance at 3 and 6 months but increases relatively to the other groups at the age of 9 months, that is, performs almost as good as Cluster 3. The third cluster consists predominantly of rural Nso infants (10.3% German vs. 55.7% Nso infants) and has the highest performance across all three measurement occasions. The fourth cluster consists predominantly of rural Nso infants (1.5% German vs. 27.1% Nso infants) and shows second highest performance at 3 and 6 months but decreases to the level of the lowest cluster group (Cluster Group 1) at the age of 9 months. At 9 months of age, the second and third clusters converge to a “high gross motor group” (with 69% of Cameroonian Nso and 50% of German middle-class infants) and the first and fourth clusters converge to a “low gross motor group” (with 31% of Cameroonian Nso and 50% of German middle-class infants), χ2 = 127.2, df = 3, p < .001.
Differences in Socio-Demographic Status
Differences in socio-demographic status (age and education of mother/father) were tested with separate MANOVAs with cultural and cluster groups (fine and gross motor cluster groups) as between-subject factors and socio-demographic status as a dependent variable. Obtained main effects of cultural groups, FAgeMother = 40.8, p < .01; FAgeFather = 0.5, ns; FEducationMother = 523, p < .01; FEducationFather = 500, p < .01, reflect differences between the German middle-class and rural Cameroonian Nso sample. Main effects of cluster groups were found for education of mother (years of schooling) for fine motor development, F(1, 234) = 4.6, p < .05, as well as an interaction. Post hoc comparisons indicated that infants in the high-performing Cluster Group 2 had mothers with higher education (years of schooling) in the Cameroonian Nso sample (MC2 = 8.6, SDC2 = 1.9, nC2 = 9; MC1 = 6.9, SDC1 = 2.5, nC1 = 52; p < .5) but not in the German middle-class sample (MC2 = 12.4, SDC2 = 1.2, nC2 = 81; MC1 = 12.3, SDC1 = 1.2, nC1 = 98; ns). In addition, a significant interaction of education of mother for gross motor development, F(3, 234) = 2.9, p < .05, was found. Post hoc comparisons indicated that education of mother did not differ in the German middle-class sample. As sample sizes were too low in Cluster Group 1 (lowest performance) and Cluster Group 2 (increasing performance) in the Nso sample, we only compared Cluster Group 3 (highest performance) and Cluster Group 4 (decreasing performance). This analysis indicated that infants in Cluster Group 3 (highest performance) had mothers with marginally significant, p < .10, lower education (M = 7.44, SD = 1.5, n = 36) than infants in Cluster Group 4 (M = 8.33, SD = 1.8, n = 15).
Differences in Weight Trajectories Between Motor Development Cluster Groups
Differences in weight trajectories (3, 6, and 9 months) between gross and fine motor cluster groups were analyzed with repeated-measures ANOVAs. For fine motor development, the linear trend was significant, F(1, 159) = 1,089, p < .001 but neither main effect, F(1, 159) = 0.12, ns nor interaction effect, F(1, 159) = 0.9, ns. For gross motor development, the linear trend, F(1, 175) = 1,027, p < .001, was significant but not the main effect, F(1, 175) = 0.48, ns. The interaction effect, F(1, 175) = 7.71, p < .05, however, was significant. As sample sizes for both German middle-class and Cameroonian Nso infants differ tremendously in the four gross motor cluster groups, another ANOVA was re-run with reduced samples (Clusters 1 and 2 only contain German middle-class infants and Clusters 3 and 4 only contain Cameroonian Nso infants). Paralleling the ANOVA results of the total sample linear trend, F(1, 157) = 950, p < .001 and interaction, F(1, 157) = 8.30, p < .01, were significant but not the main effect, F(1, 157) = 0.79, ns. The German middle-class Clusters 1 and 2 increase their weight rather parallel and linearly. In contrast, weight trajectories of Cameroonian Nso clusters show a different trend. Whereas Cluster Group 3 (highest gross motor performance) increases almost linearly, the Cluster Group 4 (decreasing gross motor performance) slows relatively down in gaining weight.
Discussion
Results of the present longitudinal study testing fine and gross motor development (3, 6, and 9 months old infants) from two eco-cultural contexts (German middle-class and Cameroonian Nso infants) demonstrate that two clearly distinct cluster groups were extracted within the fine motor development domain. These groups, however, were not clearly separated by eco-cultural context. The high-performing cluster predominantly consisted of German middle-class infants, whereas the low-performing cluster consisted of both German middle-class and Cameroonian Nso infants (with a higher percentage of Cameroonian Nso infants). In the Cameroonian sample, these two clusters differed according to educational level of mothers with infants performing high having mothers with higher education. In the German sample, however, cluster groups did not differ. We reason speculatively that with higher parental education the focus within traditional farming communities shifts toward a socialization strategy of psychological autonomy focusing on exploration and mastering of objects, thereby fostering fine motor development. This is in contrast with fostering gross motor development through socializing action autonomy through gross motor development (e.g., walking). As we did not measure object availability or parenting behaviors (e.g., encouragement of object play) in the different eco-cultural contexts, further empirical research is needed to study this relationship more extensively.
For gross motor development, four cluster groups were extracted (two predominantly German cluster groups and two predominantly consisting of rural Nso infants). At 3 and 6 months, the predominantly German groups showed lower gross motor performance than the predominantly Nso groups. At 9 months, however, one of the German groups relatively increased performance to the highest performing group, whereas one Nso group relatively decreased performance to the level of the lowest performing group. The decreasing Nso cluster had mothers with higher educational level than the highest performance cluster. In addition, for the Nso clusters weight trajectories parallel the gross motor developmental trajectories. To speculate, higher educated Nso mothers, tending to work extramural, from gross motor Cluster 4 might have weaned their infants earlier leading to a slowed down infant growth trajectory affecting gross motor development. Unfortunately, we do not have any information on weaning or quantity and quality of supplementary food of Nso mothers.
Admittedly, applying a test developed and normed in Western contexts (Bayley-III) renders its application in a traditional rural subsistence-based community a test-theoretical challenge. Although familiarity with test materials or settings differs between cultures and valid norms for Cameroonian Nso infants do not exist, test items used mainly provoked certain movements or body positions (in the gross motor scale). The fine motor scale used several objects to observe grasping and hand coordination. But even though German middle-class infants might have more toy experience, the Nso infants were very interested in the red blocks and other toys and touched them without reservation. Nevertheless, future research should try to develop culture-sensitive test instruments to disentangle real inter-individual differences from mere testing effects.
Taken together, the present study analyzed relations between eco-cultural context, motor development, weight growth, and maternal education. Looking at inter-individual differences in intra-individual developmental trajectories facilitated some understanding of the disappearing differences at the cultural group level at the age of 9 months that had been reported in preceding studies. Parental strategies as a mediating factor between eco-cultural contexts and inter-individual developmental trajectories should be included in future research. Generally, the study shows that the consideration of intra-cultural variation is important to interpret cross-cultural comparisons and that developmental trajectories are related to certain covariates in culture-specific ways.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by a grant from the German Research Foundation (Deutsche Forschungsgemeinschaft) to Heidi Keller, Monika Knopf, Arnold Lohaus, and Gudrun Schwarzer (KE 263/53–1, KN 275/6–1, LO 337/20–1, and SCHW 665/9–1).