Item Response Theory (IRT) Scale Scores. The ECLS-K direct cognitive assessment employed a two-stage design. As such, within any given domain, children received a routing set of items (stage 1) and then, based on their performance on the routing items, proceeded to a second set of items of a certain difficulty level (stage 2). Because not all children received all items, the assessment scores in the ECLS-K study were modeled using Item Response Theory (IRT). Based on childrenís performance on the items they received, an ability estimate (theta) was derived for each domain. The IRT scale scores represent estimates of the number of items children would have answered correctly if they had received all of the scored questions in a given content domain. They are useful in identifying cross-sectional differences among subgroups in overall achievement levels and provide a summary measure of achievement useful for correlational analysis with status variables. The IRT scale scores are also used as longitudinal measures of overall growth. Gain scores may be calculated by subtracting childrenís scale scores at two points in time.
Standardized Scores (T–scores). These scores are IRT based and derived from the childís ability estimate theta. They provide norm-referenced measurements of achievement; that is, estimates of achievement level relative to the population as a whole. A higher mean Tscore for a particular subgroup indicates that the groupís performance was high in comparison to that of other groups. A change in mean T-scores over time reflects a change in the groupís status with respect to that of other groups.
Proficiency Probability Scores. These scores are IRT based and derived from the childís ability estimate theta. They provide information on performance on clusters of items of similar difficulty along the overall scale. The scores measure the probability of correct responses in each cluster and can take on any value between 0 and 1. Because each proficiency probability score targets a particular set of skills, they can be used for studying the details of achievement. They are useful as longitudinal measures of change because they show not only the extent of gains, but also where on the achievement (or development) scale the gains are taking place.
Race/ethnicity. Office of Management and Budget guidelines for collecting information on race and ethnicity were followed. A respondent could select one or more of five dichotomous race categories when reporting their own race or that of their child. Each respondent was also asked to identify whether he or she (as well as the study child if the respondent was a parent) was Hispanic. The study data files include several variables indicating race and ethnicity. There are six dichotomous race variables indicating whether a respondent or study child was of a certain race (White, Black, Asian, Native Hawaiian or Other Pacific Islander, American Indian or Alaska Native, and more than one race) as well as one dichotomous ethnicity variable indicating whether a respondent or study child was Hispanic. These variables were used to create one race/ethnicity composite variable with mutually exclusive categories: White, not Hispanic; Black, not Hispanic; Hispanic of any race; Asian, not Hispanic; Native Hawaiian or Other Pacific Islander, not Hispanic; American Indian or Alaska Native, not Hispanic; and Two or more races, not Hispanic.
Socioeconomic Status (SES). Each ECLS-K data file includes a measure of SES reflecting the SES of a childís household at the time of data collection. The components used to create the SES variable are father/male guardianís education, mother/female guardianís education, father/male guardianís occupational prestige, mother/female guardianís occupational prestige, and household income. In households with two mothers or two fathers, education and occupational prestige for both mothers/fathers were used. Each parentís occupation was scored using the average of the 1989 General Social Survey (GSS) prestige scores for the 1980 census occupational category codes that correspond to the ECLS-K occupation code.