Richard N. Walker

Prediction of adult Sheldon somatotypes I and II from ratings and measurements at childhood ages

The photographs of 82 boys from the Harpenden Growth Study were assigned somatotype ratings at ages 5, 8, 11, 14, and 18 years, using both Sheldons earlier, anthroposcopic method and his revised, objective method (somatotype II). Inter-judge correlations for the anthroposcopic ratings of the 18-year-olds ranged from 0.79 to 0.93 for the three components; correlations for the somatotype II ratings ranged from 0.94 to 0.99. The three components of the somatotype II ratings showed greater independence of one another than did those of the anthroposcopic method, which tended to collapse towards two dimensions. Correlations for corresponding components between the anthroposcopic and somatotype II ratings at the same age were mostly in the low 0.80s. Mean somatotype ratings changed little with age in either method, but the somatotype II ratings were consistently higher in endomorphy and mesomorphy and lower in ectomorphy than the anthroposcopic ratings. Patterns of inter-age correlations were similar within methods: endomorphy showed lower age-to-age correlations than did mesomorphy and ectomorphy. Correlations of anthroposcopic component ratings with ratings at age 18 increased distinctly from age 5 to age 8, less sharply thereafter. Between ages 8 and 18, within observer, they were 0.72, 0.83, and 0.82, for endomorphy, mesomorphy, and ectomorphy. These 8-to-18 correlations for mesomorphy and ectomorphy are similar in magnitude to those for height.

Standards for somatotyping children: I. The prediction of young adult height from children's growth data

SummaryLongitudinal series of measurements on 143 boys and 80 girls were used to develop equations for predicting height from three combinations of variables: (1) from height at age alone, (2) from height at age plus growth rate over the preceding year, (3) from the preceding plus determination of whether or not the subject had passed his age of peak velocity of puberal height growth. Effectiveness of these equations was tested in a cross-validating sample of Berkeley Guidance Study subjects. From early adolescence onward, predictions based on (2) were somewhat more accurate and those based on (3) appreciably more accurate than those based on single observations. In the original sample, after age 6 height could be predicted with an average error under an inch, this error decreasing rapidly with onset of puberty. Use of a predicted height as one of the three parameters for objectively determining somatotype appears possible by mid-childhood. An appendix gives the prediction formulae.

Standards for somatotyping children: II. The prediction of somatotyping ponderal index from children's growth data

SummaryThe somatotyping ponderal index (SPI) — height divided by cube root of weight, measured at the time of the subjects greatest massiveness — is one of three parameters used in determining somatotype. Longitudinal series of measurements on 143 boys and 80 girls were used to develop equations for predicting SPI. The effectiveness of these equations was tested in a cross-validating sample of Berkeley Guidance Study subjects. After age 5 in boys, age 10 in girls, predictions based on repeated observations were more accurate than those based on single observations, especially so after early adolescence. After age 6, in the original sample, SPI could be predicted from longitudinal data with an average error under 0·20, this error decreasing steadily as age increased. Growth trends in ponderal index are pressented. An appendix gives the prediction formulae.

The Gesell Development Assessment: Psychometric Properties.

Abstract Tests comprising about 80% of the Gesell Developmental Assessment (GDA) were given to 400 4- to 6-year-olds—40 girls and 40 boys at five age levels, stratified by parent occupation. Four experienced judges gave global developmental ratings and a grade placement recommendation for each child. For three main judges, mean developmental age scores fell below chronological age by from 2 to 7 months; distributions suggested a different model than the bell-shaped one. For Composite scores, mean interjudge correlations by age groups ranged from .74 to .85 for single-judge ratings, from .85 to .89 for two-judge averages. Mean sex differences were small but significant, all favoring girls. The test (including no verbal items) showed no differences by socioeconomic level. At age 8 1 3 , 182 of the children were rested, on multitrait, multimethod measures. Highest correlations of GDA ratings were with later Adaptive ability; intermediate ones appeared with Number, Fine Motor, Language, and Reading abilities and general school Competence. Lowest, but still significant, were correlations with Gross Motor and Personal-Social development. Children aged 4 1 2 to 6 whose ratings fell in the lowest 25% on Recommended Grade particularly showed differences in distribution on outcome scores from children with ratings in each other quartile.