Generating Diagnostic Assessments for Spanish With AI
No published, scored, CINS-style diagnostic instrument exists for K-12 Spanish gender agreement or conjugation errors, but that isn't because nobody has studied these mistakes. Second-language-acquisition research has catalogued both in real depth for decades. The gap sits somewhere no earlier subject in this series has landed: rich, quantified misconception research that was simply never built into a scored instrument a teacher could hand out.
Quick Answer: Spanish is a sixth shape in this diagnostic-assessment series. Unlike biology's clean yes (Anderson, Fisher, & Norman, 2002) or computer science's thin, emerging research base, decades of Spanish-specific second-language-acquisition research has already documented exactly which gender-agreement and conjugation errors English-speaking learners make, and why (Franceschina, 2005; White, Valenzuela, Kozlowska-Macgregor, & Leung, 2004; Andersen, 1991). That research was never assembled into a published, scored diagnostic tool. The two real, reachable K-12 Spanish instruments — ACTFL's AAPPL and Avant Assessment's STAMP4S — pass every access and grade-band bar but measure overall proficiency level, the identical construct mismatch this series already found in ESL's WIDA Screener.
Say you're two days from opening a unit on the Spanish preterite with a class that's already sat through a year of present-tense verbs. You don't actually know, walking in, whether this group's likeliest mistake is applying English's one-size-fits-all past tense, defaulting every irregular stem to a regular pattern, or something else entirely. A diagnostic exists to answer that question before the unit's first explanation gets built around a guess.
This guide covers:
- Why a pre-unit Spanish diagnostic asks a different question than the hub's scored quiz item or its formative-check sibling
- The honest answer to whether a CINS/FCI-equivalent instrument exists for Spanish's two core mechanics, checked against AAPPL and STAMP4S directly
- Why Spanish is a genuinely new, sixth shape in this series — deep, decades-old research that never became an instrument, sitting behind a construct-mismatched placement test
- Worked pre-unit probes built from the hub's own conjugation and gender-agreement examples, a step-by-step AI workflow, a tools comparison, and the mistakes that undercut a Spanish diagnostic
This entry joins five earlier ones in the series: biology, where the Conceptual Inventory of Natural Selection transfers straight into a classroom; music, where no comparable research exists at all; coding, where a real instrument exists but fails K-9 teachers on access; ESL, where a reachable instrument measures the wrong construct; and computer science, where the underlying research itself is thin and new. None of those five shapes fits Spanish cleanly.
A Diagnostic Asks a Different Question Than the Hub's Quiz or Its Formative Sibling
A pre-unit Spanish diagnostic doesn't sit earlier on the same timeline as the hub's scored quiz item or its formative-check sibling — it answers a question neither of those formats is built to ask. Both assume a lesson already happened this year. A diagnostic assumes nothing did.
What a Scored Item and a Formative Check Get to Assume
The hub's conjugation table builds hicimos against three deliberately wrong forms, each traceable to a taught concept the item is checking (Norris & Ortega, 2000). A formative check assumes even less lead time, but still assumes today's specific lesson just happened.
What a Diagnostic Has Nothing to Check Against
A pre-unit probe has no lesson behind it at all. It measures what a student walked in with this year: a prior grade's coverage, a language app, a bilingual household, or English's own grammar leaking through. That's a fundamentally different question.
| Design Question | Scored Quiz Item | Formative Check | Pre-Unit Diagnostic |
|---|---|---|---|
| Timing | After a concept is taught | During or right after a lesson | Before the unit's first lesson this year |
| What it measures | Can the student now apply it correctly? | Is today's teaching landing right now? | What does the student already believe or default to? |
| Best prompt shape | Four-option, traceable distractors | Fast production or judgment | Open self-report or an ungraded single judgment |
| A wrong response means | A specific, gradeable error | A gap in this lesson, right now | A pre-existing pattern the unit needs to plan around |
A diagnostic that copies the hub's tight, single-right-answer shape throws away the one thing a pre-unit check exists to gather: an honest read on what a class defaults to before anyone has taught against it this year.
Does Spanish Have a CINS/FCI-Equivalent Diagnostic? The Honest Answer
No validated, scored, misconception-derived diagnostic instrument exists for Spanish gender agreement or conjugation errors specifically — but two real, K-12-appropriate, validated Spanish assessments do exist, and checking them directly against that question is the only honest way to answer it. Both pass bars earlier entries in this series found genuinely hard to clear.
AAPPL and STAMP4S Pass the Bars Coding's Instrument Failed
Coding's FCS1/SCS1 are real, validated instruments a K-9 teacher can't reach — restricted, university-normed, and shaped as one fixed battery. Spanish's two reachable instruments clear all three of those hurdles cleanly.
- ACTFL's AAPPL (Assessment of Performance toward Proficiency in Languages) is a web-based K-12 assessment built for grades 3-12, scoring Interpersonal Listening/Speaking, Presentational Writing, and Interpretive Reading and Listening against ACTFL's own Performance Descriptors (ACTFL, 2024).
- Avant Assessment's STAMP4S and its younger-grade counterpart, STAMPe, are computer-adaptive Spanish proficiency tests measuring vocabulary, grammar, and verb use across reading, writing, listening, and speaking, reported on the same Novice Low through Advanced High scale (Avant Assessment, 2025).
- Both are normed for the K-12 grade band, administered routinely by schools and districts every year, and available on demand — none of the access or format problems that ruled out FCS1/SCS1 for a K-9 classroom apply here.
The Same Construct Mismatch ESL Already Found
Neither AAPPL nor STAMP4S was built to name a specific grammatical misconception, and that's a design choice, not an oversight. The ESL diagnostic guide found the identical shape in WIDA Screener: a real, reachable, K-12-validated placement instrument that measures overall proficiency level, not which specific belief a student holds.
A STAMP4S "Intermediate Low" result and an AAPPL "Intermediate 3" rating both tell a teacher roughly where a student sits across four skills. Neither tells a teacher whether that student defaults to the present tense under time pressure, over-regularizes an irregular preterite stem, or drops gender agreement on adjectives specifically rather than articles. That's a different, more granular question than either instrument was built to answer.
| CINS (Biology) | AAPPL / STAMP4S | A Real Spanish Misconception Diagnostic | |
|---|---|---|---|
| What a result names | A specific, documented misconception about a mechanism | A proficiency level, Novice Low through Advanced High, across four skills | A specific, documented error pattern (gender-assignment default, tense default, stem over-regularization) |
| Built and validated for | Secondary biology students | K-12 language learners, for placement and Seals of Biliteracy | Not standardized; teacher-built per unit |
| Passes K-12 grade-band/access/format bars | Yes | Yes | N/A — no single instrument to test |
| Diagnoses a targetable classroom gap | Yes, directly | No — measures overall level, not a specific error | Yes, when scoped to one documented pattern |
The Sixth Shape: Decades of Research That Never Became an Instrument
Spanish's answer doesn't match any of the five shapes this series has already found, because the two halves of the question point in different directions at once. The placement side matches ESL's construct mismatch exactly. The research side looks nothing like ESL's — or CS's, biology's, music's, or coding's.
Gender Agreement Has a Deep, Quantified Research Base
The hub already cites Franceschina's (2005) finding that grammatical gender is one of the most persistent L2 Spanish error types for English speakers. That single citation understates how much has actually been measured.
- White, Valenzuela, Kozlowska-Macgregor, and Leung (2004) tested English-speaking L2 Spanish learners and found a specific, repeatable asymmetry: accuracy on determiner-noun agreement consistently outpaces accuracy on adjective-noun agreement.
- Montrul, Foote, and Perpiñán (2008) compared adult L2 learners against Spanish heritage speakers and found errors skew in one predictable direction: masculine gender gets overgeneralized to feminine nouns far more often than the reverse.
- Bruhn de Garavito and White (2002) used grammaticality-judgment tasks, the same methodology the true/false sibling already grounds in Bialystok (1979) and Ellis (2005), to argue persistent gender errors trace to lexical gender assignment — not knowing a noun's gender — more than to a broken agreement rule.
Conjugation Has Its Own Named, Competing Theories
Conjugation errors have an equally mature research thread, built around a genuine scholarly disagreement rather than one settled finding. Roger Andersen's Aspect Hypothesis proposed that L2 Spanish learners initially attach the preterite to punctual, completed actions and the imperfect to ongoing states, based on longitudinal data from English-speaking learners (Andersen, 1991).
M. Rafael Salaberry's Default Past Tense Hypothesis offers a competing account: beginning learners default to the preterite broadly, treating it as equivalent to English's simple past, regardless of a verb's inherent aspect (Salaberry, 1999; Salaberry, 2008).
Two named researchers, using real classroom and longitudinal data, disagree about which specific default an English-speaking Spanish learner reaches for first. That's not thin evidence — it's the kind of active, cited scholarly debate a genuinely mature research area produces, closer to what preceded CINS in biology than to CS's handful of 2024 interview studies.
Why This Research Never Crystallized Into a CINS-Equivalent
Every study above measures error patterns; none of them built a scored, multi-item, publicly available test a teacher can administer and grade the way CINS's 20 items or MG-CSCI's 24 items work. The research stayed inside academic grammaticality-judgment batteries, elicited-production tasks, and small-sample longitudinal studies, never packaged into a norm-referenced instrument.
That gap is worth naming precisely against the five prior shapes in this series.
| Subject | Underlying Research | Instrument Status |
|---|---|---|
| Biology | Deep, mature | Built into CINS, a scored, validated 20-item instrument (Anderson, Fisher, & Norman, 2002) |
| Music (notation/theory) | Thin to none | No instrument; Gordon's batteries measure a different construct entirely |
| Coding | Deep, mature | Built into FCS1/SCS1, but restricted and university-normed |
| ESL | Deep, mature (interlanguage/error analysis) | No single instrument; WIDA Screener exists but is construct-mismatched |
| Computer science (non-code) | Thin, emerging | No instrument; research itself is still being built |
| Spanish (gender/conjugation) | Deep, mature, decades-old | No instrument at all — and the reachable placement tests are construct-mismatched too |
Turning Documented Error Patterns Into a Pre-Unit Probe
Since no packaged instrument exists, the realistic classroom tool borrows the same two moves earlier entries in this series landed on: an open, ungraded Background Knowledge Probe (Angelo & Cross, 1993) for a broad first read, and a single, untimed grammaticality-judgment item for a sharper one. Both reuse sentences the hub already built rather than inventing new ones.
A Gender-Agreement Probe Built From the Hub's Own Sentence
The hub's correct sentence is La chica alta lee el libro nuevo. A diagnostic version doesn't score it — it asks a student to judge one unscored variant before any gender-agreement lesson this year, exactly the kind of untimed judgment Bruhn de Garavito and White (2002) used to separate assignment errors from agreement errors.
"Before we study this, read this sentence and tell me: does it sound right to you, and if not, which word feels off? 'La chica alto lee el libro nuevo.'"
A student who flags alto has already noticed an adjective-agreement break unprompted. A student who says nothing sounds wrong reveals exactly the pattern White et al. (2004) documented: weaker sensitivity to adjective agreement than to determiner agreement.
A Conjugation Probe Built From the Hub's Own Verb
The hub's preterite item uses hacer, with hicimos as the target nosotros form. A pre-unit version turns the same verb into an open production prompt, with no options to select from and nothing at stake.
"Before we study the preterite, finish this sentence however feels natural to you: 'Ayer, mis amigos y yo ___ la tarea juntos.'"
- An answer like hacemos signals a present-tense default, worth knowing before a preterite unit opens.
- An answer like hacimos signals stem over-regularization, the exact pattern the hub's third distractor surface already names.
- A correct hicimos tells a teacher this student needs less scaffolding on this verb than the rest of the class.
A Grade-Calibrated Illustration
Say you teach a grade 7 Spanish II class about to open a unit on irregular preterite verbs after a year of steady present-tense work. Before a single lesson, you could hand out five open completion prompts built around five different irregular verbs, unscored, framed honestly as "write whatever feels right."
Reading twenty-eight short responses that evening shows whether the class defaults to present tense, over-regularizes stems, or splits roughly evenly, a genuinely different first-lesson plan for each pattern. A grade 4 class just starting gendered articles can run a lighter version: one picture, two labels, "which one looks right to you, el or la gato?"
A Step-by-Step AI Workflow for a Spanish Diagnostic
Building a defensible pre-unit Spanish diagnostic starts with naming the exact documented error pattern, not the grammar topic, since a vague prompt returns a vague, ungradeable probe.
- Name the specific pattern, not the topic. "Gender agreement" isn't specific enough; "over-generalizes masculine gender to feminine nouns" is (Montrul, Foote, & Perpiñán, 2008).
- Choose gender or conjugation, and one documented sub-pattern within it, rather than a broad, unscoped probe covering both mechanics at once.
- Write the prompt as an open production or an untimed single judgment, never a four-option quiz item, since a diagnostic has no answer key to defend.
- Reuse a sentence or verb your class has already seen, the way the examples above reuse the hub's hicimos and La chica alta sentence, so the probe isn't the first time students meet the vocabulary.
- Set your class profile's grade level and course sequence, so verb and sentence complexity match what the class has actually covered.
- Ask AI to draft three or four prompts at once, spanning both mechanics, so the diagnostic samples broadly instead of narrowing to one pattern.
- Sort responses into three buckets before planning instruction: already correct, shows the documented pattern, or unsure — and let that sort, not a score, shape the unit's first days.
A Sample Prompt
Grade 7 Spanish II, opening a unit on irregular preterite verbs after
a year of present-tense work. Generate four short, open, ungraded
pre-unit prompts: two sentence-completion items using different
irregular verbs (one testing stem over-regularization, one testing a
present-tense default), and two single-sentence gender-agreement
judgments (one testing article agreement, one testing adjective
agreement), reusing vocabulary this class has already covered.
Explicitly frame every prompt as ungraded.
You could use EduGenius to hold a class profile's grade level and course sequence while drafting a batch of open, pattern-targeted pre-unit prompts anchored to a named error type, then export the finished probe as a one-page handout to hand out before the unit opens.
Tools for Building Spanish Diagnostics
A tool for this job needs open-ended prompt generation and the ability to name a documented error pattern explicitly, since a diagnostic's value comes from the pattern it surfaces, not a graded score.
| Tool | Open-Ended Prompt Support | Anchors to a Named Error Pattern | Class-Context Memory | Starting Price |
|---|---|---|---|---|
| EduGenius | Yes, via open-response format selection | Yes, if the pattern is named in the prompt | Yes, via saved class profiles | 25 welcome credits free; Starter $7.99/month (500 credits); Professional $15.99/month (1,000 credits) |
| General-purpose AI chat tools | Yes, if explicitly requested | Only if the source is supplied in the prompt | No, unless re-typed each time | Varies by provider |
| AAPPL / STAMP4S | No — fixed, standardized proficiency instrument | Not applicable; not misconception-focused | Built as a placement instrument, not a diagnostic | District/school licensing |
| Manual, teacher-written probe | Yes, inherently | Only as well as the teacher researches it | No | No cost |
For the timing distinctions this guide narrows, see Generating Formative Assessments for Spanish With AI. For the true/false format's own grammaticality-judgment grounding, see Generating True/False Quizzes for Spanish With AI. The cross-subject speed problem sits in How to Write Better Diagnostic Assessments Faster With AI, and the wider assessment landscape is in AI-Powered Assessment and Quiz Generation.
Pro Tips for Spanish Diagnostics
- Name the documented pattern, not just the grammar point. "Tests whether a student over-regularizes an irregular stem" beats "a hard question about hacer."
- Reuse the exception-noun list from the matching worksheets guide. Words like mapa and problema are exactly where an assignment-error probe earns its keep.
- Keep every probe genuinely ungraded, and say so out loud. A student who suspects a diagnostic is secretly scored defaults to a safe guess, not an honest one.
- Sample both mechanics in the same sitting. You don't yet know whether this class's widest gap is gender assignment, agreement itself, or a conjugation default.
- Reuse a verified probe every year and compare results. A pattern that recurs across cohorts earns a dedicated first lesson, not just a mention.
What to Avoid
Mistaking AAPPL or STAMP4S for a Misconception Diagnostic
A proficiency-level result is real and useful for placement, but it can't say whether a student's specific gap is gender assignment, agreement, or a tense default. Confirming a level isn't the same work as surfacing a documented pattern.
Scoring a Diagnostic Like a Quiz
Grading an ungraded probe defeats its purpose. A student who suspects "what feels right to you" is secretly scored starts guessing safely instead of answering honestly.
Treating an Untested Claim as a Documented Pattern
Ask AI which study a claimed error pattern traces to, and verify it against real sources like White et al. (2004) or Andersen (1991), the same discipline the true/false sibling already applies to a claimed grammar rule.
Assuming Deep Research Means an Instrument Already Exists
A rich citation trail on gender-assignment errors and preterite defaults doesn't mean a scored test exists to hand a class. That research has to be turned into a teacher-built probe by hand, the same gap the ESL guide found on the placement side of its own subject.
Key Takeaways
- A pre-unit Spanish diagnostic asks a different question than a scored quiz item or a formative check — what a student already defaults to, before any lesson this year gives them something to apply.
- No CINS/FCI-equivalent instrument exists for Spanish gender agreement or conjugation, even though real, K-12-reachable Spanish assessments do: ACTFL's AAPPL and Avant Assessment's STAMP4S (ACTFL, 2024; Avant Assessment, 2025).
- AAPPL and STAMP4S repeat ESL's WIDA Screener finding exactly: they pass every K-12 access and grade-band bar, but measure overall proficiency level, not a specific documented error.
- Spanish is a sixth shape in this series: decades-deep, quantified research on gender-assignment and conjugation-default errors (Franceschina, 2005; White et al., 2004; Andersen, 1991; Salaberry, 1999) that was simply never assembled into a scored instrument.
- The realistic classroom tool combines two borrowed methods: an open Background Knowledge Probe (Angelo & Cross, 1993) and an untimed grammaticality-judgment item, the same method the true/false guide already grounds in real SLA research.
- AI speeds up drafting pattern-anchored prompts; verifying every claimed research pattern and reading every response stays a human task.
Frequently Asked Questions
Is there a validated diagnostic test for Spanish, like biology's CINS or physics's FCI?
No. Despite decades of documented research on gender-assignment and conjugation-default errors among English-speaking learners (Franceschina, 2005; White, Valenzuela, Kozlowska-Macgregor, & Leung, 2004; Andersen, 1991), that research was never built into a scored, published instrument the way CINS was for biology.
What's the difference between AAPPL, STAMP4S, and a Spanish diagnostic assessment?
AAPPL and STAMP4S are real, validated K-12 placement instruments that report a proficiency level, Novice Low through Advanced High, across reading, writing, listening, and speaking (ACTFL, 2024; Avant Assessment, 2025). A diagnostic is a teacher-built, ungraded probe targeting one specific, documented error pattern before a unit starts — a narrower, different question.
Does English-speaking learners' Spanish gender-agreement errors follow a predictable pattern?
Yes. Research has found masculine gender gets overgeneralized to feminine nouns more often than the reverse, and accuracy on adjective agreement consistently lags behind accuracy on determiner agreement (Montrul, Foote, & Perpiñán, 2008; White, Valenzuela, Kozlowska-Macgregor, & Leung, 2004).
Can AI build a real Spanish diagnostic assessment?
Yes, for open, pattern-targeted probes grounded in documented research, once a prompt names the specific error pattern and reuses vocabulary the class has covered. It works best paired with a human check that any claimed research pattern is real, and a clear, spoken commitment that the probe is genuinely ungraded.
Sources
- ACTFL (2015). World-Readiness Standards for Learning Languages (4th ed.). American Council on the Teaching of Foreign Languages.
- ACTFL (2024). AAPPL: Assessment of Performance toward Proficiency in Languages. American Council on the Teaching of Foreign Languages.
- Anderson, D. L., Fisher, K. M., & Norman, G. J. (2002). Development and evaluation of the conceptual inventory of natural selection. Journal of Research in Science Teaching, 39(10), 952-978.
- Andersen, R. W. (1991). Developmental sequences: The emergence of aspect marking in second language acquisition. In T. Huebner & C. A. Ferguson (Eds.), Crosscurrents in Second Language Acquisition and Linguistic Theory (pp. 305-324). John Benjamins.
- Angelo, T. A., & Cross, K. P. (1993). Classroom Assessment Techniques: A Handbook for College Teachers (2nd ed.). Jossey-Bass.
- Avant Assessment (2025). STAMP4S: Standards-based Measurement of Proficiency. Avant Assessment.
- Bialystok, E. (1979). Explicit and implicit judgments of L2 grammaticality. Language Learning, 29(1), 81-103.
- Bruhn de Garavito, J., & White, L. (2002). The second language acquisition of Spanish DPs: The status of grammatical features. In A. T. Pérez-Leroux & J. M. Liceras (Eds.), The Acquisition of Spanish Morphosyntax (pp. 153-178). Kluwer Academic Publishers.
- Ellis, R. (2005). Measuring implicit and explicit knowledge of a second language: A psychometric study. Studies in Second Language Acquisition, 27(2), 141-172.
- Franceschina, F. (2005). Fossilized Second Language Grammars: The Acquisition of Grammatical Gender. John Benjamins Publishing.
- Montrul, S., Foote, R., & Perpiñán, S. (2008). Gender agreement in adult second language learners and Spanish heritage speakers: The effects of age and context of acquisition. Language Learning, 58(3), 503-553.
- Norris, J. M., & Ortega, L. (2000). Effectiveness of L2 instruction: A research synthesis and quantitative meta-analysis. Language Learning, 50(3), 417-528.
- Salaberry, M. R. (1999). The development of past tense verbal morphology in classroom L2 Spanish. Applied Linguistics, 20(2), 151-178.
- Salaberry, M. R. (2008). Marking Past Tense in Second Language Acquisition: A Theoretical Model. Continuum.
- White, L., Valenzuela, E., Kozlowska-Macgregor, M., & Leung, Y.-K. I. (2004). Gender and number agreement in nonnative Spanish. Applied Psycholinguistics, 25(1), 105-133.