WF-14 — Employer and Supervisor Feedback

Both of these KPIs are currently unreported at MBRU institutionally, and both are recoverable by IoL without a single curriculum change. Neither is a performance problem. KPI 2.3 fails because no named workplace supervisor exists to survey. KPI 2.4 fails because Institute - Employers.xlsx has almost certainly been populated, if at all, as a partnership list rather than as a list of the employers of last year's graduates. Two registers close both gaps.

1. Purpose

To ensure that the people who supervise IoL learners in the workplace, and the people who employ IoL graduates, are identified by name and contact at the moment the relationship exists, are surveyed on the Ministry's instruments within the Ministry's timetable, and return enough responses at a high enough quality that MBRU can report a defensible employer feedback score for both programmes, with the raw per-respondent data retained and producible on demand.

2. Scope

Scope statement. This process manages employer and supervisor feedback for IoL programmes from the identification of a named respondent, through survey issue, reminder and collection, to the return of results to MoHESR and the retention of the raw evidence.

Applies to. Workplace supervisors of PGDipHPE and MScHPE learners on approved placements, whether the site is a Dubai Health entity or an external partner organisation; and the employers of PGDipHPE and MScHPE graduates of the previous academic year, wherever those employers are located. Both the Ministry's Employer Work-placement Survey (EWS) and the Ministry's Employer Satisfaction Survey (ESS).

Does not apply to. Student satisfaction with a placement, which is a different construct captured in Internship_satisfaction and belongs to WF-06. It also does not apply to partnership relationship management, alumni relations correspondence, or employer engagement for recruitment purposes; those are WF-24 and WF-25. Feedback from an organisation that has neither supervised a learner nor employed a graduate is outside scope entirely, however valuable the relationship.

Applicable requirements. OBEF guide v11.5 KPI 2.3 and KPI 2.4; CHEDS/HEDB Master API templates Institute - Employers.xlsx, Students - Internship.xlsx and Students - Graduates.xlsx; Appendix B evidence provisions, which name employer work-placement survey raw data explicitly; Appendix C small-sample and grouping rules; MBRU data protection policy and the UAE personal data protection regime, because this workflow transfers named third-party contact details to a ministry.

The division of labour that governs everything below. These two KPIs are Track B, meaning the Ministry computes them, but the operational burden is not symmetrical and confusing the two is the commonest way to lose them.

Formulae, quoted. KPI 2.3 is the sum of employer evaluation scores for all students during the work placement, divided by the number of students evaluated by employers. KPI 2.4 is the sum of employer evaluation scores, divided by the number of completed Employer Satisfaction Surveys. Note that the 2.3 denominator counts students, while the 2.4 denominator counts completed surveys. They are not the same kind of unit and cannot be reconciled to each other.

The programme-level propagation rule for 2.4, stated plainly. Graduates whose employers completed the ESS are linked back to the programmes they graduated from, and those programmes receive the same score the institution received from that employer. A single employer response therefore propagates to every programme that employer's hires came from. For IoL this is consequential in both directions: one large Dubai Health employer that hires from both PGDipHPE and MScHPE sets the score for both programmes at once, and a single poor response is not diluted by the size of the cohort.

The work-placement definition, which constrains the 2.3 population. Accepted placement forms are clinical training, practical training, apprenticeship, co-op programmes and diplomatic postings. Each must be formally approved by the institution and a formal part of the programme structure, curriculum or study plan. A supervisor of an activity that fails either test is not a KPI 2.3 respondent, and surveying them anyway pollutes the numerator with responses that cannot be evidenced.

3. Trigger, boundary and endpoint

4. SIPOC

5. Accountability

Process owner. IoL Director of Programmes, or the role able to require respondent capture inside WF-13 and WF-10 and to authorise the transfer of third-party contact data to MoHESR.

Escalation.

6. Process steps

1. Confirm the KPI 2.3 respondent population. Take from WF-13 every placement episode completed in the reporting period and its named workplace supervisor. Test each against the placement definition: was the placement institutionally approved, and is it in the approved study plan? Episodes failing either test are excluded from the survey population and the exclusion is recorded with its reason. [CAPTURE] respondent name, position, organisation, email, the placement episode it relates to, and the learner and Program Code it attributes to. Serves 2.3. Decision point. If an episode has no named supervisor, it cannot generate a 2.3 response. Route the defect back to WF-13 rather than substituting a departmental contact. A generic address returns a response that cannot be attributed to a supervisor and cannot be evidenced under Appendix B.

2. Confirm the KPI 2.4 respondent population. Take from WF-10 the graduate cohort of the previous academic year, and from WF-25 the employer of record for each graduate at 12 months. One row per employer, with the graduates and Program Codes that employer accounts for carried alongside. [CAPTURE] Name of the employer, Position, Sector, Industry, Company Name, Location_Country, Location_Emirate, Email, Phone (Optional), Academic Year, plus the graduate-to-employer linkage that is not a CHEDS column and must live in an IoL register. Serves 2.4.

3. Apply the population test that decides whether KPI 2.4 is reportable at all. Institute - Employers.xlsx must be populated as employers of graduates, not as a generic partnership or engagement list. These are different populations, they overlap only partially, and only the second serves this KPI. Before submission, assert that every row traces to at least one named graduate of the previous academic year. Rows that do not are removed to the partnership register, which is WF-24's asset, not this one's.

4. Validate and normalise. Every respondent row needs a syntactically valid, current email and a named individual with a position. Organisation names are normalised against the same controlled list of Dubai Health entities and aliases that WF-13 uses for placement sites and WF-25 uses for employers of record. [CAPTURE] normalised organisation identifier. Serves 2.4 and 3.1 matching. Exception route. A bounced or unreachable contact is corrected once and, if still unreachable, is recorded as unreachable with a date. It stays in the population count. Silently dropping unreachable respondents inflates the apparent response rate and is a misstatement.

5. Agree and document the Appendix C grouping, before any survey runs. IoL's supervisor and employer populations for a single programme will be small. Appendix C permits merging programmes that share degree level and graduation year, similar content area, or similar employment fields. PGDipHPE and MScHPE satisfy the content-area and employment-field principles on their face. [CAPTURE] the grouping decision, its rationale, the second signature from Quality and IQA, and the date it was made. Serves 2.3, 2.4 and the Mechanism of collecting the data field. This step has a date test, not just a content test. A grouping decided after results are visible is not a grouping, it is a selection. The date is the evidence that it was not.

6. Freeze the population. Record the population as at a stated date, with counts by programme and by organisation. Nothing enters or leaves after the freeze except a documented correction. [CAPTURE] frozen population with its date and counts.

7. Issue the Employer Work-placement Survey (2.3). MBRU distributes this one. Issue within 10 working days of placement completion, or in a single scheduled wave where the placement calendar is synchronised, whichever the programme operates. Two reminders at 7-day intervals, then a personal follow-up by the placement coordinator for non-responders. [CAPTURE] issue date, reminder dates, collection date, respondent identity, and the full per-question raw response. Serves 2.3. Retain the raw per-question data. Appendix B permits the Ministry to demand exactly this, by name, for this KPI: participant contact details, per-participant per-question raw data, and survey issue and collection dates.

8. Submit the employer list and chase responses (2.4). MoHESR distributes this one, so IoL's levers are the quality of the list and the chase. Notify each employer contact in advance that a ministry survey is coming and why, which is both good practice and the single largest determinant of response rate for an externally distributed instrument. Follow up after the Ministry's issue date. [CAPTURE] date the list was supplied, date the Ministry issued, the advance-notice and follow-up record. Serves 2.4. Note the dependency. IoL cannot see the Ministry's response file directly. Response chasing is therefore blind unless MoHESR reports back. Ask, in writing, for a response manifest by employer, and record the answer.

9. Measure the response rate against the chosen Appendix C route and report it before the fieldwork window closes, not after. A projected shortfall is actionable; a measured shortfall is a post mortem. [CAPTURE] responses received, population, response rate, the route applied.

10. Compute and return the KPI 2.3 result to MoHESR. Sum the employer evaluation scores across all evaluated students and divide by the number of students evaluated. Where a student was evaluated by more than one supervisor, decide once how the student's score is formed, record the rule, and hold it constant. [CAPTURE] numerator, denominator, result, and the method note for the Mechanism of collecting the data field. Serves 2.3. Never submit Internship_satisfaction as the KPI 2.3 value. That field is the student's satisfaction with the placement. It is a different respondent, a different construct and a different KPI, and the mistake is easy to make because the two live in the same template.

11. Consider substitute data for KPI 2.4. The guide explicitly invites institutions to submit their own data directly to MoHESR for a defined set of KPIs including 2.4, where they consider it more accurate than the standard survey. Decide deliberately, with a second signature, and record the methodology. Serves 2.4. See section 9 for the test that applies.

12. Close the loop and feed back. Return themes, not scores, to WF-04 annual programme monitoring, and acknowledge every respondent. An employer who receives nothing after responding will not respond next year, and this KPI is a three-year rolling average.

13. Lodge the evidence. Every raw response file, with issue and collection dates and respondent identity, goes to the Evidence Register at the point of collection, not at submission time. [CAPTURE] evidence artefact reference against the KPI and academic year. Serves Appendix B and cross-KPI assertion C14.

7. OBEF data generated

Capture rule. Four things cannot be reconstructed after the fact. The named supervisor with a working email, captured while the placement is live and the relationship is current. The employer of record at 12 months with a named contact, captured at the point of contact and not inferred later from a job title on a professional network. The raw per-question responses with their issue and collection dates, which cease to exist the moment a survey platform is closed or a licence lapses. The date of the grouping decision, which is the only thing distinguishing a legitimate Appendix C grouping from a post-hoc selection. Everything else in the table above can be rebuilt from records; these four cannot.

Reproducibility test. Today, a second analyst could reproduce neither KPI at IoL. For 2.3 there is no supervisor register, so the respondent population cannot be constructed at all, which is the capture defect this workflow exists to close. For 2.4 the graduate-to-employer linkage does not exist as a held record, so even if the Ministry returned a set of employer scores, IoL could not reproduce which programme each score propagates to. Both defects are register defects, and neither requires anybody's permission to fix.

8. Max-score design

Read the anchors carefully. The gap between the start of High and the top is half a point on a five-point scale. A supervisor population that rates IoL learners at 4.5 scores 75, and at 4.75 scores about 87. The difference between scoring nothing and scoring 75 is a register; the difference between 75 and 100 is half a point of rating. Effort should follow that ratio, which means every hour spent on respondent capture and response rate is worth more than an hour spent trying to move the rating.

What this workflow must do to reach the target.

1. Obtain a named supervisor with a working email for every completed placement, sourced from WF-13 at approval, not chased at survey time. Contact details captured after the placement ends are materially less accurate, and in a health system where staff rotate they decay within months.
2. Build Institute - Employers.xlsx as an employers-of-graduates list and defend it as such. This is the whole of KPI 2.4's precondition, and it is a five-minute test applied to every row: name the graduate.
3. Hold the graduate-to-employer-to-programme linkage in an IoL register, because CHEDS has no column for it and without it the programme-level propagation rule cannot be applied or reproduced.
4. Decide and document the Appendix C grouping before fieldwork, with a Quality and IQA signature and a date.
5. Set the response expectation at induction for supervisors and at graduation for employers. Announcing a survey at the moment of issue is the difference between a return rate that clears Appendix C and one that does not.
6. Retain raw per-respondent data with dates from the first cycle onward, because both KPIs are three-year rolling averages and the evidence obligation is retrospective.

[REDESIGN] actions.

Sequencing note. R3 gates R1 and R2 in practice, because both produce third-party personal data that must reach a ministry. It is the least visible action on this list and the one most likely to be discovered late, at the point of transfer, when the survey window has already opened. Start it first even though R1 is the item with the point value attached.

9. Indirect strategy where data cannot be collected

Category 2 and Category 4 both apply, and the two KPIs sit in different places.

Category 4, activity exists but is invisible, applies to KPI 2.3. IoL learners are supervised in the workplace by identifiable senior professionals every day. The supervision is real, the supervisors are known to the programme, and the evaluations would be returned if they were requested. What is missing is a register of who the supervisor was, in a form that lets a survey be sent and a response attributed to a learner and a Program Code. The response to Category 4 is capture, not tactics, and here the capture is action R1: one mandatory field at the point of placement approval.

Category 2, the population is too small for the instrument, applies to both. IoL's supervisor population for a single programme, and its employer population for a single programme's graduating cohort, will both fall below the Appendix C thresholds that govern the survey KPIs. Two sanctioned routes exist and IoL should use both deliberately.

• Grouping. Appendix C permits merging programmes that share degree level and graduation year, similar content area, or similar employment fields, in order to reach a usable population. PGDipHPE and MScHPE share content area and employment field plainly and obviously: they are both health professions education qualifications whose graduates work as educators inside health systems. The grouping must be decided, documented, signed and dated before the survey runs. That sequencing is the entire difference between a legitimate small-sample remedy and a selection of the favourable answer, and it costs nothing to observe.

• Substitute data, for KPI 2.4 only. The guide explicitly invites institutions to submit their own data directly to MoHESR where they consider it more accurate than the standard survey, and 2.4 is named in that provision. IoL is a small, high-contact postgraduate department whose graduates and their employers are personally known to the programme team. A documented near-census of employers, with retained per-respondent data, is plausibly more accurate than an externally distributed sample that returns a handful of responses.

  The test the guide sets is accuracy, not favourability. The question to answer, in writing, before submitting is: would this method give a better estimate of employer satisfaction even if the result were worse? If the honest answer is no, do not submit it. And be prepared to produce the raw per-respondent data with dates, because Appendix B permits the Ministry to demand exactly that for a substitute submission. A substitute submission that cannot survive an evidence request is worse than no submission.

The substitute-data route does not exist for KPI 2.3, which is not in the guide's named list, and which in any case MBRU already distributes itself. For 2.3 there is nothing to substitute for: the failure is entirely internal, and so is the fix.

Ambiguities in the guide, stated rather than resolved.

1. Whether continuing employment counts as "hired" for 2.4. The ESS goes to employers who hired graduates of the previous academic year. Most IoL learners are already employed, frequently by the same organisation, throughout their studies. Whether an employer who retained rather than hired a graduate belongs on the list is not addressed. This is the same definitional question WF-13 raises for KPI 3.1 and it should be put to MoHESR in the same letter. Do not assume the favourable reading. Record the assumption applied, in the Mechanism of collecting the data field, until an answer arrives.
2. The unit of the 2.4 denominator under the propagation rule. The denominator is the number of completed surveys. If one employer completes one survey and that employer hired graduates from both PGDipHPE and MScHPE, both programmes receive that employer's score, but whether the institutional denominator counts one completed survey or two is not stated. It matters only when programme results are rolled up and checked against the institutional result under cross-KPI assertion C10. Flag it to WF-26 rather than resolving it locally.
3. Whether one respondent may serve both KPIs. In the IoL case a learner's workplace supervisor and a graduate's employer contact will frequently be the same person at the same organisation. Nothing in the guide prohibits this, and the two KPIs measure different things at different times, so on the plain reading it is acceptable. It is nonetheless worth confirming, because the respondent burden is doubled and the perception of a single voice setting two KPIs is worth managing.
4. When the Ministry shares the EWS instrument. MBRU is obliged to distribute an instrument the Ministry designs, but no timetable for the Ministry to release it is stated. IoL cannot plan a fieldwork window around an undated dependency. Request the instrument and its issue calendar in writing, early, and record the request as the evidence that the delay was not IoL's.

Boundary check. This workflow must never:

• survey a supervisor of an activity that is not an institutionally approved placement in the study plan, or count such a response in the KPI 2.3 numerator;
• submit Institute - Employers.xlsx as a partnership, engagement or prospective-employer list, or leave a row in it that cannot be traced to a named graduate of the previous academic year;
• submit Internship_satisfaction, which is the student's view of the placement, as the KPI 2.3 result;
• revise the Appendix C grouping, or the composition of the survey population, after any result is visible;
• drop unreachable or non-responding respondents from the population in order to raise the reported response rate;
• select which supervisors or employers are surveyed on any basis other than the frozen population rule;
• submit own data for KPI 2.4 because it is more favourable rather than more accurate, or submit it without the raw per-respondent data retained and producible;
• transfer supervisor or employer contact details to any party, including MoHESR, without the lawful basis established under action R3;
• assume the favourable reading of the "hired" question at 2.4 in the absence of a Ministry answer.

10. Service standards

11. Records and evidence

Appendix B readiness. Today, IoL could produce almost none of this within 15 working days, for the simple reason that most of it is not being created. There is no supervisor respondent register, no employers-of-graduates register, no raw response data because no survey has been run, and no dated grouping decision. That is not a retention failure, it is a capture failure, and it is why both KPIs currently report null. From the first cycle in which R1 to R3 are operating, the position reverses completely: every artefact in the table above is generated as a by-product of running the survey, and readiness becomes a matter of storing it rather than finding it.

12. Risks and controls

13. Performance measures

The first four measures are the ones that decide whether these KPIs exist at all. A supervisor population can be entirely satisfied with IoL learners and the KPI will still report null, and these four measures are the difference.

14. Change control