Learning Readiness Report™ · Case A · 16 August 2026

Case A, mandatory compliance course

The full report behind the published case study, in the format a client receives. 73 content pages · 15 assessment items · publisher blinded.

This is a static analysis run, not an instrumented playback

Every finding comes from inspecting the package. The six learner profiles were applied to that content analytically. Each profile is a defined set of failure conditions, and we checked the course against them.

We did not drive this course in a browser and we are not reporting timings. There are no session durations, no click paths and no dwell times in this report, because we did not measure any. A client engagement adds an instrumented pass. This published run does not have one.

Verdict

Not Ready for Deployment

Triggered by the hard gate, which is absolute. A learner can complete this course and be recorded as compliant without answering a single question correctly. Under our rubric that fails the course regardless of every other number, and no remediation of the other dimensions changes it.

Three critical issues were found. The dimension scores below are published for remediation planning, not to soften the verdict.

Scorecard

Objective alignmentNot assessable

The course states no learning objectives. Not weak ones: none. Across all 73 pages there is no instance of "objective", "you will be able to", "upon completion" or any equivalent. The denominator for this dimension does not exist, so we reverse-engineered six objectives from the content and scored the rest of the report against those. They are our inference, not the publisher's statement.

Capability coverage, at the required level38

The content covers the subject matter thoroughly and accurately at the level of explain. The audience's actual task is decide. These are the people who approve or deny personnel actions. Five of the six derived objectives require applying a rule to a fact pattern; the instruction presents the rule and stops. One item in the entire course asks the learner to apply anything to a scenario.

Practice adequacy22

15 knowledge checks exist, which is not nothing. But 32 of 46 feedback strings are verdicts rather than explanations. Under our rubric practice without explanatory feedback does not count as practice. Worse, all feedback is delivered through a popup window, which modern browsers block by default, so in current conditions a wrong answer produces a red highlight and silence.

Assessment alignment17

Every item tests recall or recognition against capabilities that require application. Two items resolve to "All of the above" where all options are plainly in scope. The cumulative review item is an earlier question with one word changed. And the correct answer to all 15 is declared in the page source, so the items cannot discriminate between a learner who knows the material and one who read the HTML.

Synthetic learner performance31

Median across the three instructional profiles. Cold Start, The Practitioner and Minimum Effort. Reported with the spread: 18 to 44. The minimum-effort learner is excluded from the median and reported separately below, because including a learner who passed without engaging would flatter the number.

Access and readability26

Reported separately from the instructional score and excluded from the guarantee. Operability: no viewport meta tag on any of 73 pages, a fixed 796-pixel container and zero media queries in either stylesheet. Readability: statutory register throughout, and a glossary that defines terms from a different course entirely.

Critical issues3

A count, not a percentage. Three of the three qualifying conditions are met.

Step one, the objectives, reverse-engineered

Because the course states none, we derived them from what the content teaches and what the stated audience does. Everything downstream is scored against these six. If the publisher disagrees with the derivation, that disagreement is itself the most useful output of this report.

Derived objectiveLevel requiredInstructionPracticeAssessed at level
1. Determine whether a given employee and period of service are coveredApplyPresentRecall onlyNo
2. Carry out the agency's obligations when an employee gives noticeApplyPresentRecall onlyNo
3. Decide whether a returning employee has restoration rightsApplyPresentOne scenarioPartly
4. Determine the correct reemployment position and deadlineApplyPresentRecall onlyNo
5. Route a complaint correctly between the two available channelsApplyPresentRecall onlyNo
6. Apply pay, leave and benefit-continuation rules during serviceApplyPartialNoneNo

The pattern this table shows

Instruction is present for all six. Assessment at the required level is present for none. This is the most common shape we find, and it is why completion rates on courses like this one look healthy: the content is genuinely there, and the test never asks for the thing the content was for.

The panel

01Cold StartNo prior knowledge of the subjectBlocked
  • Twelve specialist terms are used without definition anywhere in the course.
  • The glossary belongs to a different course. It contains 14 entries: abuse of authority, age discrimination, color, disability, gender, national origin, whistleblower and similar, drawn from an equal-opportunity module. Zero of this course's own terms appear in it, and the course's own subject is not mentioned once. A learner who hits an unfamiliar term, clicks Glossary as instructed, and lands on definitions for an unrelated topic has been actively misdirected.
  • The second instructional page defines a core concept in a single 74-word sentence lifted from statute, listing seventeen categories without an example.

Traced to: overview, m1/p2, m1/p3, m1/p5, m1/q7, m3, glossary

02The PractitionerFifteen years in the roleCompletes, distrusts
  • A statutory change is described as "a new law effective March 15, 2009." The package was last updated in 2016 and is still in circulation. A practitioner reads "new law" against a seventeen-year-old date and begins discounting everything else, correctly, because they have no way to tell which other provisions have moved.
  • The five-year service limit is taught, then immediately undercut with "however, there are numerous exceptions", and the exceptions are not enumerated. Determining whether an exception applies is the practitioner's actual job.
  • Nothing distinguishes what changed since the learner last took it. For mandatory annual training this is the difference between a refresher and a re-read.

Traced to: m1/p7, m1/p10

03The Mobile LearnerTouch only, small screenBlocked
  • No viewport meta tag on any of the 73 pages. A mobile browser applies its default virtual viewport and scales the page down.
  • A fixed 796-pixel content container and zero media queries in either stylesheet. There is no responsive path at all.
  • The consequence is pinch-zoom and horizontal panning through 73 pages of statutory prose. Not impossible; reliably abandoned.

Traced to: css/main.css, all 73 pages

04The Second-Language ReaderFluent, but not nativeCompletes, comprehension unverified
  • The register is statutory throughout: long subordinate clauses, nested conditions, and definitions that resolve into other undefined terms. One prerequisite is defined in terms that are circular for any reader who does not already know the underlying regulations.
  • The wrong glossary compounds this. The support mechanism a second-language reader relies on most is the one that has been pointed at another course.
  • No plain-language summary, no worked example, and no equivalent path through the material.

Traced to: m1/p2, m1/p7, glossary

05The Keyboard-Only UserNo mouse, no touchCompletes

The clearest positive result in this run, and not the result we expected.

  • Answer options are real anchor elements, so every assessment item is reachable and operable by keyboard. The pattern is dated. A button element would be correct, but the outcome is fine.
  • Every page carries a skip link and an access key to a keyboard-shortcuts list. Someone thought about this.
  • Of 627 image tags, exactly one lacks an alt attribute; 280 carry alt="", which is correct for decorative images.

Traced to: m1/q1 markup, all page templates

06Minimum EffortWants to pass and leavePassed, hard gate triggered

This learner succeeded, which fails the course outright. Three independent routes, each sufficient alone:

  • Read the key. Every correct answer is declared in the page source. Fifteen questions, fifteen visible answers.
  • Exhaust the options. The check is evaluated client-side with unlimited attempts, and the code that would re-disable the Next button after a wrong answer is commented out.
  • Skip the questions. No course page reports a score or status. The only file containing SCORM calls is the generic API wrapper. Completion is issued for reaching the final page.

Traced to: all 15 question pages, functions.js, APIWrapper.js

Critical issues, three of three conditions met

01 · A learner can get credit without demonstrating the capability

Answer keys are in the page source, retries are unlimited, and nothing is scored or reported. This is the hard gate and it is why the verdict is what it is.

02 · A prerequisite is required but never taught

Twelve specialist terms carry the reasoning in this course and none is defined in it. The glossary that exists is for a different course. A learner without prior experience in this domain cannot resolve them from the material provided.

03 · An assessment item tests something the course never covered

One item requires the learner to identify entitlements that go beyond the statute's requirements, including two named benefit programs that appear nowhere in the instruction, only in the question and its options. The item is answerable by elimination, which is how it survived.

Remediation, ranked by what unblocks the most

Ordered so that the first item alone changes the verdict. Effort estimates assume access to the source project, which we did not have.

  1. Move scoring server-side, or out of the package entirelyNothing else on this list matters while the answer key ships to the browser. The realistic options are an LMS-hosted assessment, a rebuilt SCO that reports a score against a declared mastery score, or an external proctored check. Any of the three closes critical issue 01 and lifts the hard gate.Changes the verdict on its own
  2. Replace the glossary with this course's own termsTwelve terms, already identified. This is the cheapest item on the list and it closes critical issue 02, which currently blocks the single largest learner segment.Under a day
  3. Rewrite the assessment to the required cognitive levelFive of six derived objectives require applying a rule to a fact pattern and are assessed by recall. The course already contains one item that does this correctly. It is the pattern to copy. Remove the duplicate and the two 'All of the above' items while you are there.The main body of work
  4. Replace popup feedback with in-page feedback, and write itTwo problems in one fix: the delivery mechanism is blocked by modern browsers, and 32 of 46 strings say only that the learner was wrong. Explanatory feedback is what makes the 15 existing items count as practice.Moderate
  5. Add a viewport meta tag and a responsive stylesheetThe container is fixed at 796 pixels with no media queries. Contained, mechanical work, and it is what stands between the course and a substantial share of its audience.Moderate
  6. Date the content and enumerate the exceptionsRemove 'new law' from a 2009 provision, add a revision date, and list the exceptions to the five-year rule rather than referring to them.Low
  7. HousekeepingFifteen question script files are captured error pages from an unrelated system, and editor backup files are present in four directories. Neither affects a learner. Both indicate nothing inspected the artifact before it shipped.Low

Retest is included

A Preflight includes one retest of the revised course with a before-and-after delta on every dimension. On this course the delta that matters is binary: either the hard gate clears or it does not.

What this report does not establish

It is a structured risk screen, not a measurement of learning effectiveness. It identifies specific, traceable, fixable instructional defects. It does not predict how real learners will perform, and no score here should be read as evidence that the training does or does not work. We have not run a validation study comparing our findings to human reviewers or to learner outcomes, and until we have we will not make an accuracy claim about the score.

We tested the package, not the deployment. An LMS can impose mastery rules on a SCO that reports nothing. If the publisher's platform does that, the completion record may mean more than the package alone supports. Nothing inside this package would supply such a rule with data.

The access and readability dimension is a screen, not a conformance audit. We do not test with assistive technology, we do not issue a VPAT or ACR, and we do not evaluate every success criterion. A score of 26 means our profiles found serious obstruction. It is not a conformance result in either direction.

The derived objectives are our inference. The course states none. Every downstream score depends on that derivation being reasonable, and the publisher is better placed than we are to correct it.

If you think this might be your course

Email hello@actyra.com with the subject “Is this our course?” and we will tell you, privately, within two business days. If it is yours you get the identity, the package hash and the complete working notes at no charge, and no follow-up unless you ask for one.

That offer stands for any public-sector publisher whether or not one of these runs is theirs. If you want your own course checked against the same three critical-issue conditions, ask and we will do one free, for a course of up to 3 hours.


The published case study for this run · All published runs · How Preflight works