Interpretation Problems of Age and Grade Equivalents

none - publisher technical guidance

Users of Pearson-published norm-referenced tests (the PPVT-III worked example is given)

Publisher guidance, not research, and undated beyond a 1996-2026 copyright range - grade D and the year is an approximation of retrieval, not of authorship. The independence flag is what makes it worth recording: this is the TEST PUBLISHER advising against a score type its own instruments report. A developer-led recommendation normally gets discounted for the direction it pushes; here it pushes against the developer's interest, which is the same logic that makes a developer-led null stronger than a developer-led positive.

Four stated defects of age and grade equivalents, and one flat recommendation. (1) Reliability is limited by the raw-score relationship underneath the equivalent. (2) The scale is radically non-linear: on the PPVT-III a 5-raw-point change equals about 4 MONTHS of development at age 4, and nearly 2 YEARS at age 16+ - the same score gap means wildly different things at different points. (3) "Age and grade equivalents are not a ratio or interval scale of measurement. They cannot be added, subtracted, or averaged" - which invalidates the most common uses, including differencing two subtests' GEs or averaging them. (4) "The reliability for age-equivalent scores is much poorer for advanced test-takers", i.e. exactly where a high-scoring child's GE gets quoted. The recommendation is unambiguous: "These scores should not be used for making diagnostic or placement decisions."

Not applicable - measurement guidance with no estimates.

Not an empirical finding. Consistent with the long-standing measurement-community position and with the publishers' equivalent guidance elsewhere.