Humphreys's attenuated test of Spearman's hypothesis

Humphreys's test of Spearman's hypothesis (viz., that the size of the standardized black-white differences on various psychometric tests is positively related to the tests' loadings on g, the general intelligence factor) is methodologically weak. It is based on comparison of a fairly representative sample of the black population of U.S. school children with a highly selected sample of the white school population, representing the lowest 15 to 20% of whites in socio-economic status (SES). A fair test of the hypothesis requires that the black and white samples should not be selected on any g-correlated variable, such as SES. Selection attenuates and distorts the relationship between tests' g loadings and the magnitude of the standardized mean black-white differences on the tests. Other unorthodox conditions in Humphreys's study, such as performing factor analysis on the test-score means of various arbitrary demographic groups instead of on individual test scores inflates tests' g loadings and biases the test of Spearman's hypothesis by restriction of reliable variance in g loadings. Humphreys's study cannot be considered a proper replication of Jensen's examination of Spearman's hypothesis in 11 different studies comprising 74 different tests, which consistently bears out Spearman's hypothesis.     

Predicting young women's role preference for parenting and work

Young women's motivations for two sex-role options were predicted with scores describing perceived rewards/costs of parenting. Using a sequence of multiple regression analyses it was found that the reward-cost perceptions of parenting and work were predictors of the two motivational options. Individual rewards/costs of parenting and work items were better predictors than summed items. It was found that a preference for the career choice was produced by not viewing parenting as rewarding. The preference for parenting choice was influenced most by perceiving parenting as rewarding. The perceived cost and rewards of work were not predictors of a career choice. The items found to be the best predictors for either the parenting or career choice all dealt with perceptions of parenting. It was concluded that these parenting perceptions are the key elements for young women's role motivations.     

The development of compensatory stepping skills in children

The development of the ability to use the step for balance recovery was studied among twenty-five 9- to 19-month-old children. The children were grouped according to walking experience (4 levels) and exposed to backward support surface translations, 8 cm in amplitude, under 3 velocity conditions: 15, 20, and 25 cm/s. New walkers (up to 2 weeks' walking experience) used the step infrequently and ineffectively in response to threats to balance. Intermediate walkers (1-3 months' walking experience) showed an increasing use of the step and significant improvement in step execution compared with new walkers. Advanced walkers (>3; months' walking experience) experienced no falls throughout the protocol, capturing balance with feet-in-place or step responses under all perturbation conditions. A significant developmental transition in the emergence of the compensatory step occurred between the new walker and the intermediate walker experience levels, that is, within the first 3 months of walking experience. Three to 6 months' experience was required for the development of an effective stepping response. A concomitant change in mediolateral stability paralleled the emergence of compensatory stepping.     

Learning and transfer from a simple dynamic system

The amount of learning gained from being assisted in completing the task of bringing a predator-prey system into equilibrium by controlling the predator population was investigated. Learning was explored both by post-task questioning and by testing for transfer to another predator-prey task. Participants were 28 undergraduate psychology students, all female. They were randomly and evenly split into an experiment group that was subjected to a learning session with the first task before being tested in the second task, and a control group that only performed the second task. What was most needed in the first task was help in sticking to analytically derived conclusions by resisting "common sense" responses. There was a significant transfer effect on performance to the second task, stemming from learning shown by half the participants in the experiment group. The other half showed hardly any learning. Learning in about half the subjects has proven a stable finding.