Acronyms and terms of art appearing in the Michigan TAC record. Resolve terms here rather than inferring them — several are Michigan-specific and at least one (AGP) is routinely mistyped in source documents.
Assessment programmes and instruments
Term
Meaning
M-STEP
Michigan Student Test of Educational Progress. The state summative assessment. Assesses the same content as Smarter Balanced but on a converted scale — hence M-STEP / Smarter Balanced Concordance.
MI-Access
Michigan’s alternate assessment for students with significant cognitive disabilities. The only assessment using the logistic regression growth target method.
Smarter Balanced instructional resource tied to the interim assessments.
Growth measurement
Term
Meaning
SGP
Student Growth Percentile. The percentage of students with similar score histories having lower growth than this student. Describes growth achieved, normatively.
AGP
Adequate Growth Percentile. The SGP a student must consistently meet or exceed to reach proficiency within a set timeframe. A criterion target on a normative scale. Rendered APG in places in the April minutes.
Growth target timeframe
The number of years in which a student is expected to reach proficiency. Michigan: 1–3 years for content areas, 1–5 for English language proficiency.
NGA
Nearest Grade Average. Proposed growth target method borrowing from the nearest grade where a full model is estimable. See Nearest Grade Average Method.
R SGP package
The R implementation used for Michigan’s quantile regression growth targets.
NTILE(50)
Division into fifty ordered groups. Used in NGA to define norm-group buckets — chosen so each averages 1,000+ R-package AGPs.
Growth-to-standard
Growth measured against a proficiency criterion, as opposed to purely normative growth.
Subscore reporting and psychometrics
Term
Meaning
Claim
ELA and mathematics subscore unit. Michigan reports three of four mathematics claims, combining Claims 2 and 4.
Domain
Science subscore unit — Earth Science (ES), Life Science (LS), Physical Science (PS).
Subscore performance indicator
The reported band: Above Standard, At/Near Standard, Below Standard, relative to the overall content cut.
The Smarter Balanced subscore indicator method, currently in use in Michigan.
K-modes
Clustering method used as a reference point in the deviance comparison.
GMM
Gaussian Mixture Model. Used to generate true theta values in the simulation.
QWK
Quadratic Weighted Kappa. The agreement statistic used to report classification accuracy and consistency.
Theta (θ)
The latent ability estimate.
R2SS
Raw to Scale Score conversion.
Unidimensionality
The property of a test measuring a single latent construct. Checked by the Wainer et al. (2001) procedure — Cronbach’s α per subtest, then eigendecomposition of the error-corrected covariance matrix, looking for one dominant eigenvalue.
Classification accuracy
Agreement between assigned category and true category.
Classification consistency
Agreement between categories assigned on two independent administrations.
Contractual requirement that MTTC forms differ substantively. OEE’s working definition set a single high overlap ceiling across all fields and was rejected by the committee. Specific ceilings are in the restricted layer. See MTTC Meaningfully Different Forms.
Equating
Statistical adjustment placing scores from different forms on a common scale. Requires a minimum of 100 examinees per form.
Linear / z-score equating
The two equating approaches specified for MTTC Groups 1 and 2.
Standard setting
Expert judgement process establishing cut scores. Used for Groups 3 and 4, where equating is infeasible.
Scorable / non-scorable items
Items that count toward a score, versus embedded field-test items that do not.
Automated item generation
Producing families of “cousin” items from a common item model.
Conjunctive cut score
A requirement to reach a minimum on each sub-area, not only on the total.
Incidence
Candidate volume in a certification field. Determines which of four groups a test falls into.
Every Student Succeeds Act. Federal law requiring growth reporting. Michigan’s approved ESSA plan constrains changes to the AGP — see ESSA Plan Constraints.
School Index
Michigan’s high-stakes school accountability system.
Points Earned / Points Possible. How social studies subscores are reported.
Break in trend
An announced discontinuity in a reported series following a methodology change, after which longitudinal comparison is invalid.
Cited literature
Wainer et al. (2001) — the unidimensionality check procedure used in the subscore simulation.
Sinharay (2014); Gorney and Sinharay (2025) — cited on the limits of the existing subscore reporting literature, particularly its multivariate normality assumption.