Session 4 — Michigan Test for Teacher Certification
Presenters: Bridgett McDowell (OEE), Nancy Hahn and Heather Klesch (Evaluation Systems – Pearson / Evaluation Systems), Jackie Thompson (Pearson Professional Assessments) · Day: April 23, 2026
The longest session of the meeting — scheduled 1:00–4:00pm with a break. No deck survives in the corpus; this page is reconstructed from the minutes alone.
What OEE brought
Two problems under one contract.
First, the new Pearson contract requires OEE to “maximize the number of meaningfully different test forms for each field,” and OEE needed the committee to tell it whether its working definition of meaningfully different was defensible.
Second, whether to build constructed-response tests for the optional science and social studies endorsements.
Full threads: MTTC Meaningfully Different Forms and MTTC Constructed-Response Endorsements.
The definition problem
OEE’s working definition: a single high ceiling on overall item overlap between test forms, with a slightly lower ceiling on the scorable set, applied uniformly across every field. Rationale offered was standard practice in educator licensure testing, program integrity, and legal defensibility. Goal: a minimum of three forms by contract end.
The committee did not accept it, saying the ceiling was far too high to constitute meaningful difference — and then putting it in terms of the contract’s own horizon: a handful of forms produced over several years, each sharing most of its items with the others, cannot reasonably be called meaningfully different.
The specific ceilings are held in the restricted layer; see restricted/mttc-form-security-detail.md.
The constraint that makes it hard
Low-incidence fields do not generate examinees fast enough. Statistical equating needs a minimum of 100 examinees per form. Re-using a single form raises both security and fairness risk. Only a small number of forms can be live in any one testing window, and a candidate may repeat after a fixed interval — those two figures together are the most sensitive detail in the corpus and are held in the restricted layer.
OEE’s response was to sort the certification tests into four groups by incidence. The two higher-volume groups build a new form once a field reaches the 100-examinee threshold, differing in equating method. The two lower-volume groups do not equate at all and are standard-setting driven, with Pearson conducting the standard settings.
Per-group form counts are in the restricted layer. One detail matters publicly: asked why the permitted form range for the highest-volume group was what it was, the answer was that this was what the old contract said — not that it followed from Michigan’s volumes.
The validity question underneath
Early in the session the committee asked what inferences OEE can defend for each of these tests. The answer given was procedural: a passing candidate can be certificated, which allows a district to interview them.
The committee’s response is the most consequential thing said in the session — Define the inference the MTTC actually defends. OEE should be able to answer substantively, not procedurally: what can you infer when a candidate does not reach the cut score, what does performance mean, and what do MTTC scores represent?
What the committee advised on forms
- Define ‘meaningfully different’ per incidence group — one uniform overlap rule across four very different groups is the wrong instrument
- Start with a standard-setting form and grow the pool — begin with a standard-setting form and grow the pool as non-scorable items prove out
- Stick with 100 examinees per form, not 200 — 100 per form, not 200; waiting buys precision at the cost of forms
- Escalate a field to the next group after two years below 100 — if 100 cannot be reached in two years, move the field to the next group’s process
- Reduce item overlap between forms — target roughly half the working ceiling
- Prefer fewer forms with less overlap — fewer forms with less overlap beats many forms with lots of overlap
- Use automated item generation to enlarge the pool — automated item generation to build the pool
- Define incidence categories for Michigan — define incidence categories for Michigan rather than inheriting them
The two-year escalation rule carries its own reasoning: educator preparation programs will not change enough in under two years to move the numbers, but after two years the absence of change is itself information.
What the committee advised on constructed response
The committee advised against building them: Do not build constructed-response endorsement tests. “This creates a lot of work, will be hard to do, and will be expensive; suggest that this isn’t necessary.”
It offered four alternatives instead — Consider transcript review to verify coursework, Use higher conjunctive cut scores on the existing test, Build one open-ended prompt from a randomly selected standard, and Give districts an interview guide instead of a test.
And it returned to purpose: Define the purpose of the endorsement assessment. What does having the endorsement on your certificate signify? OEE’s own answer — that you know the content in the standards — is a claim a cheaper instrument could support.
A complication noted in the room: higher education institutions may design their educator preparation programmes as they wish and are not aligned across programmes, which is precisely why transcript review is imperfect and why the endorsement test exists at all.
Open questions
How long does it take to accumulate enough examinees to equate? · Can equating data be too old?