Table of Contents  |  Search Technical Documentation  |  References

NAEP Analysis and Scaling → Estimation of NAEP Score Scales → NAEP Scales → Correlations Among NAEP Subject Area Subscales → Calculating Marginal Correlation Coefficients

NAEP Technical DocumentationCalculating Marginal Correlation Coefficients

Marginal correlation coefficients are estimates of the Pearson product-moment correlations between scores on the subscales within a content area. They are calculated using the computational aid of plausible values by computing

• separate Pearson product-moment correlation coefficients for each plausible value,
• the Fisher's z-transformation for each value,
• the average of the transformed values, and
• computing the inverse transformation of the average.

Both marginal and conditional correlations are estimated from scales created using both Item Response Theory (IRT) models for items and population models to account for subgroup membership.

Printer-friendly Version