A workplace report can contain accurate figures and still support an unreliable recommendation. The difficulty is often the step between describing what happened and explaining what should be done next.
For CIPD 5CO02 Evidence-based practice, connect the people-practice question, the evidence, its limitations and the proposed decision. Calculate carefully, distinguish a pattern from an explanation and show why the available evidence supports this action more strongly than the alternatives. A neat chart is useful only when its meaning survives inspection.
This guide uses the public CIPD qualification specification checked on 24 October 2026. Its organisation, records, people and calculations are fictional teaching material. They are not a completed assignment, a current assessment brief or evidence about an actual employer.
Find your next learning step
Connect a people-practice question with clear measures, cautious interpretation and a proportionate recommendation in CIPD 5CO02 learning.
ROUTE 1 · CHAPTERS 1–3
Define the decision
Locate the unit, frame the question and define the measures.
ROUTE 2 · CHAPTERS 4–6
Inspect the evidence
Check arithmetic, explanations and qualitative scope.
ROUTE 3 · CHAPTERS 7–9
Choose a response
Work through the case, alternatives and people affected.
ROUTE 4 · CHAPTERS 10–12
Review the reasoning
Connect outcomes, communication and changed conditions.
ROUTE 5 · CHAPTERS 13–15
Check independent understanding
Try a new case, read the FAQs and return to primary guidance.
Full chapter index · Independent learning check · Education parent guide
Full chapter index
Define the decision · Chapters 1–3
Inspect the evidence · Chapters 4–6
Choose a response · Chapters 7–9
Review the reasoning · Chapters 10–12
Check independent understanding · Chapters 13–15
CHAPTER 1 OF 15 · Define the decision
1. Locate the unit before choosing a method
The current Associate Diploma in People Management specification identifies 5CO02 as Evidence-based practice. Its scope includes interpreting quantitative and qualitative evidence and presenting findings that inform people-practice decisions.
Use the specification alongside your study centre's current instructions. A public unit description does not establish the exact question, dataset or submission requirements of an individual assessment.
The learning task is broader than choosing a familiar business model and applying its labels. You need to know what decision is being considered, what information would help and what the evidence can actually establish.
CIPD's evidence-based practice factsheet provides a wider professional framework. This article concentrates on the learner's reasoning from records to a defensible recommendation.
The existing Organisational Development guide retains the wider discussion of organisational change. A 5CO02 learning exercise should not become an attempt to redesign an entire organisation when the evidence addresses a narrower issue.
Begin by stating the decision in one sentence. If the sentence names only a topic, such as retention or training, it probably needs further work.
CHAPTER 2 OF 15 · Define the decision
2. Replace a broad concern with an answerable question
“Staff turnover is a problem” does not yet identify what needs investigating. Which staff, over what period, compared with what expectation, and with what consequences?
In an original fictional case, Rowan Services is concerned about employees leaving its customer-support team. A manager wants to purchase a training package. The learning question is whether the available evidence justifies that response.
Several different problems could sit behind the same headline. New starters might leave before becoming confident. Experienced employees might leave because working patterns no longer fit their circumstances. A temporary contract might end as planned.
These possibilities require different information. A total count of departures cannot distinguish them by itself.
A useful question might ask whether early departures are concentrated among a particular entry cohort, and what the available records suggest about their experience. That wording creates a route to investigation without deciding the explanation in advance.
Keep the question proportionate to the evidence you can legitimately use. A learner should not invent confidential records, imply access to real employees or expand a small supplied dataset into a claim about an entire sector.
An answerable question improves both analysis and writing. It gives each figure a job and makes irrelevant detail easier to remove.
CHAPTER 3 OF 15 · Define the decision
3. Check what a number counts before calculating
A departure count needs a definition. Does it include voluntary resignations, completed temporary contracts, internal transfers or all recorded exits? Different definitions answer different questions.
The observation period also matters. A count for one month should not be compared directly with a count for a year and described as a change in the same measure.
In a fictional training exercise, Department A records six departures and Department B records four. It is tempting to say that A has the greater retention problem. But the teams may be very different sizes.
If A has an average headcount of sixty and B has twenty, the illustrative departure-to-average-headcount ratios are ten per cent and twenty per cent respectively. These calculations describe the supplied example; they are not a benchmark or a universal official turnover formula.
Write down the numerator, denominator, population and period. Follow the definition required by the actual task when one is provided.
This is a useful point for a teacher to pause. Ask the learner what a result means in a sentence without using the word percentage. If they cannot explain what the numerator and denominator represent, the calculation may have become detached from the question.
Check the consequences of a changed denominator
Return to the fictional department comparison. Department A had six departures and an average headcount of sixty; Department B had four departures and an average headcount of twenty.
Now imagine that the task's records actually define Department B's twenty as headcount at one date rather than an average for the period. The earlier calculation used a value with the wrong meaning.
The arithmetic can be flawless while the measure is wrong. This is why a report needs more than a correct formula.
A learner should go back to the supplied definition and establish which denominator is available. If the needed value cannot be established, identify the limitation rather than silently relabel a convenient number.
Suppose an exercise supplies opening and closing headcounts but also notes a substantial expansion halfway through the year. A simple average may still be the required calculation, but its limitations deserve interpretation.
The right response depends on the task. Do not replace an instructed method with a preferred one without explaining why. Do not describe an instructed estimate as if it perfectly captures every change during the period.
A clear report can distinguish the calculation required for the exercise from the additional information that a real organisational decision might need.
For teaching, ask the learner to create two plain-language descriptions of the same number: one accurate and one that quietly overstates its scope. Comparing them helps expose how definitions can change without any visible arithmetic error.
CHAPTER 4 OF 15 · Inspect the evidence
4. Make arithmetic transparent enough to check
Keep an inspectable path from the supplied values to the result. A reader should be able to identify the source numbers, the operation and the unit.
For Rowan Services, suppose a fictional exercise provides forty employees at the start of a period, fifty at the end and nine departures during it. Using the simple average of opening and closing headcount gives forty-five. Nine divided by forty-five gives twenty per cent.
That result depends on the stated method. It does not establish why employees left, whether departures were avoidable or whether training would improve retention.
An average based only on the opening and closing figures may also conceal substantial variation within the period. If monthly records are available, the task may support a more informative denominator.
| Record or calculation | What it can contribute | What it cannot establish alone |
|---|---|---|
| Count of departures | The number of recorded events | Whether team sizes make comparisons fair |
| A clearly defined rate | A size-adjusted description for the period | The cause of the pattern |
| Exit comments | Accounts of some employees' experiences | A complete account of everyone who left |
| Training attendance | Participation in an activity | Independent performance after it |
The table is an original reasoning aid. Use the actual definitions and instructions in your assessment, rather than copying its examples into a submission.
Inspect an aggregate that hides two different populations
Consider a second original training case involving support sessions for fictional new starters. The records group employees into novice and experienced entrants. These labels describe the invented dataset, not a judgement about actual people.
Among novice entrants who attended support sessions, twelve of twenty completed a defined task independently during the observation period. Among novice entrants who did not attend, two of five did so.
The illustrative proportions are sixty per cent and forty per cent. Within this category, the observed task outcome is more common among attendees.
Among experienced entrants, nine of ten attendees completed the task independently. Eighteen of twenty-five non-attendees did so. The corresponding proportions are ninety per cent and seventy-two per cent.
Now combine the categories. Attendees account for twenty-one completions among thirty people, or seventy per cent. Non-attendees account for twenty completions among thirty people, or approximately sixty-seven per cent.
Even where the overall direction is similar, the combined result conceals important differences in who attended. A simple pooled percentage does not explain why participation differed or whether the groups faced comparable tasks.
The learner should first check the definitions. Does independent completion mean the same thing in each category? Were people observed for the same period? Did attendance occur before the outcome was recorded?
Then inspect the composition. The attendee group contains more novice entrants than the non-attendee group. Comparing the pooled results alone mixes the support question with differences in the populations.
None of these figures proves that the sessions caused improvement. The example is observational and fictional. It teaches the learner to inspect the comparison before giving the total an explanation it cannot support.
A useful recommendation might therefore seek a clearer account of eligibility, task difficulty and baseline capability before expanding the programme. The numerical work changes the next question rather than automatically settling the decision.
This is the practical value of a calculation that does not produce a dramatic conclusion. It can show which comparison is missing.
CHAPTER 5 OF 15 · Inspect the evidence
5. Keep a pattern separate from a causal explanation
Suppose the fictional records show that several departures occurred among employees who had attended an induction course. That does not establish that induction caused the departures.
Perhaps nearly all new employees attended the course. Perhaps people in the most demanding roles received additional training. Perhaps the observation periods differ.
Ask what comparison would be meaningful. Who was eligible, who attended, when did the activity happen and what other conditions changed?
This does not require a learner to conduct a formal experiment for every workplace question. It requires the strength of the conclusion to match the evidence.
You can state that a pattern warrants closer examination without claiming that its cause has been proven. You can recommend a limited diagnostic step when the data do not yet justify a large intervention.
The Sengkang guide to observed, inferred and unknown makes this boundary practical. A recorded departure is an observation. A proposed reason may be an inference. Some details remain unknown.
Using those categories does not weaken a report. It makes the reader's decision more dependable by preventing an interpretation from quietly becoming a fact.
CHAPTER 6 OF 15 · Inspect the evidence
6. Ask whose experience the qualitative evidence represents
Comments can reveal something a numerical summary misses. A person may describe unclear expectations, inconsistent support or a scheduling difficulty. Such accounts can suggest questions worth investigating.
However, a small collection of comments has a scope. Who was invited to respond, who chose to respond and whose experience is absent?
In Rowan Services, suppose three departing employees provided exit comments. Two mentioned difficulty getting timely help. The comments support a concern about those reported experiences. They do not establish that two-thirds of all employees lack support.
The same phrase can also mean different things. “Not enough training” might refer to technical knowledge, access to a supervisor, unclear procedures or confidence when facing an unfamiliar task.
Before coding comments into themes, inspect their context. Avoid converting every negative remark into the theme that best supports a preferred recommendation.
A teacher can give learners two short fictional comments that use the same broad label for different practical problems. Ask what additional question would help distinguish them.
The purpose is to learn how interpretation changes a decision. It is not to turn a handful of invented comments into research findings or to claim that any employer has a particular problem.
Turn a comment into an investigable question
Suppose one fictional exit comment says, “There was nobody to help when a difficult case arrived.” This is an important account, but its meaning still needs to be understood.
The learner could interpret it as insufficient training, unavailable supervision, unclear escalation arrangements or a mismatch between assigned tasks and readiness. Each interpretation points toward different evidence.
A useful first question is what kind of help the person needed. Did they need knowledge of a procedure, permission to make a decision, reassurance about an unusual situation or access to a colleague with relevant expertise?
The comment should not be dismissed because it is qualitative. It should also not be treated as a complete causal explanation.
Now add another invented comment: “The training was clear, but I did not know who could authorise an exception.” This narrows one part of the problem. Repeating the training may leave the authorisation question unresolved.
A third comment might describe an entirely different difficulty, such as a schedule that no longer fits personal circumstances. Combining every departure under one theme would erase those distinctions.
The learner can group related comments while preserving differences that matter for the decision. A theme should help understanding rather than flatten the evidence.
In a report, explain how the theme was derived and what it includes. If the available comments are few or self-selected, state that scope.
The next action may be a focused review of the relevant process, undertaken through legitimate organisational channels. An educational exercise does not itself authorise interviews with real employees or access to private records.
CHAPTER 7 OF 15 · Choose a response
7. Work through a recommendation that runs ahead of its evidence
The fictional manager at Rowan Services proposes a full-team training programme because early departures have increased. The learner's first report agrees and supplies a chart of total departures.
The chart may be correct, but the reasoning has skipped several links. It has not shown which employees left, what difficulty the training addresses or why this intervention is preferable to another response.
The learner revisits the records. In this invented exercise, the departures are concentrated among recent starters assigned to a late shift. The available comments mention difficulty obtaining help when unusual cases arise.
This does not prove that shift support is the cause. It does make a general classroom training package a less immediate conclusion.
The learner now proposes checking how support operates on that shift and whether new starters can obtain help with unfamiliar cases. The recommendation is smaller, but more closely connected to the evidence.
A later training response may still be appropriate. The important change is the order: investigate the suspected learning or support problem before selecting the package.
This resembles the distinction in BTT's guide to foundation and method errors. Different causes can produce a similar visible difficulty. A useful repair depends on finding the relevant link.
CHAPTER 8 OF 15 · Choose a response
8. Compare alternatives using criteria that matter
A recommendation becomes easier to judge when the alternatives are visible. Doing nothing, gathering further information, changing access to support and delivering training may have different benefits and costs.
Choose criteria connected with the actual issue. For the fictional late-shift problem, useful considerations might include availability of timely help, feasibility within existing working arrangements, burden on colleagues and the ability to review what changes.
Do not invent a precise financial return when the necessary information is unavailable. A clearly identified missing cost is more honest than a plausible-looking figure presented as measured fact.
Likewise, avoid scoring every option on an arbitrary numerical scale and treating the total as an objective answer. If you use a decision aid, explain the criteria, assumptions and meaning of the ratings.
Two people may reasonably weigh priorities differently. The report should make that difference discussable rather than hide it behind a total.
For a learner, this is a valuable transition from naming advantages and disadvantages to evaluating them. The relevant question is which consequence matters for this decision, given the evidence and constraints.
An option with several small benefits is not automatically preferable to one that addresses the central problem more directly.
Check whether the recommendation has a feasible owner
A recommendation can be well reasoned and still leave implementation vague. In the fictional support case, “improve access to help” does not identify who can organise the change or what resources it requires.
The learner can distinguish the person making a recommendation from the person authorised to decide or implement it. An educational report should not imply that its author can alter real working arrangements.
Describe the proposed next action at an appropriate level. A review of the support process may require a manager's agreement, relevant staff input and suitable access to records. These are considerations to establish, not permissions supplied by the assignment.
Think about what happens if the action is not feasible in its original form. Could the question be investigated through a smaller legitimate step? Would a different option address the same issue with fewer demands?
Feasibility does not mean selecting the easiest action regardless of the problem. It means recognising whether the proposal can actually be carried out in a way that preserves its purpose.
For learning, compare a recommendation with an activity list. The recommendation explains why a response fits the evidence. The activity list identifies work that might follow. Neither replaces the other.
A useful final inspection asks whether someone unfamiliar with the case could identify the issue, the reason for acting, the proposed decision and the uncertainty that remains. If those are clear, the report offers a basis for discussion rather than a demand for agreement.
Keep language proportionate. “The available evidence supports investigating this route” may be more accurate than “this will solve retention.” The stronger statement needs stronger evidence.
A learner who can make this distinction has developed something useful beyond an assessment. They can contribute a clear, revisable proposal to a decision while remaining honest about the limits of their information.
Before finalising, check that the recommendation would still make sense if the organisation name were removed. Its justification should depend on the particular evidence and constraints, rather than a generic story about what all employers ought to do.
CHAPTER 9 OF 15 · Choose a response
9. Include the people affected by the decision
A people-practice recommendation affects people whose experience may not be fully visible in the supplied records. Consider whose views need to be understood and how the proposed action could create different effects.
For Rowan Services, adding an experienced colleague to late-shift support may help new starters but place additional demands on that colleague. A proposed review should consider that workload rather than assume that support is costless.
An apparently convenient arrangement may also work differently for staff with different responsibilities or access needs. These considerations should be examined through the legitimate information and processes available.
This guide does not provide employment-law advice or establish the legal requirements of a real organisation. The educational point is that an ethical lens changes what counts as a sufficient analysis.
Do not collect personal information merely because it might be interesting. Use the information required by the task and the setting's actual permissions and safeguards.
A learner can show good judgement by identifying a question that needs an appropriate organisational process, rather than pretending that an assignment exercise authorises them to investigate real colleagues.
CHAPTER 10 OF 15 · Review the reasoning
10. Distinguish an activity measure from an outcome
If Rowan Services introduces a support arrangement, counting the number of support sessions tells you that something happened. It does not establish that the original problem improved.
An outcome measure should connect with the problem the action was intended to address. In a fictional learning exercise, the review might examine whether new starters can obtain timely help and handle appropriate cases more independently.
Even then, interpretation requires care. Improved performance might coincide with a quieter period, changes in case difficulty or other support. A before-and-after comparison does not automatically isolate the intervention's effect.
Choose a review period and evidence that fit the proposed action. Avoid a promise that a short pilot will establish every long-term outcome.
The existing DELTA Module Three course-priorities guide explores another education setting where a plan needs to connect needs, learning and assessment. Its qualification requirements are different, but the distinction between delivering activity and building capability remains useful.
For 5CO02, keep the conclusion tied to the specific people-practice issue and the evidence your task actually supports.
Make the review capable of changing the plan
A proposal to “monitor the results” leaves several decisions unresolved. What result matters, when could it reasonably appear and what would cause the action to be revised?
For the fictional support-access intervention, an early review might examine whether new starters know where to seek help and whether appropriate help is available when needed. A later review could examine a more independent capability connected with the original problem.
These are different questions. An improvement in awareness does not automatically establish improved performance, and improved performance does not automatically establish that departures will fall.
Consider a possible review outcome: staff report knowing the support route, but unusual cases still wait for an authorised decision. The intervention may have clarified access without addressing the bottleneck.
A useful response would inspect the unresolved decision stage rather than declare the entire programme successful or unsuccessful.
Another possible outcome is that help becomes faster but the supporting colleague's workload becomes unmanageable. The review should notice that consequence as well. Improving one part of the system by overwhelming another may not be a dependable solution.
Before implementation, explain what evidence would support continuing, adapting or stopping the proposal. Those are learning questions for the fictional case, not promises about an actual employer.
CHAPTER 11 OF 15 · Review the reasoning
11. Present a finding before decorating it
A useful finding states what the evidence shows, why it matters and what limits its interpretation. Formatting then helps the intended reader see that reasoning.
A chart should have a clear population, period, unit and label. A table should make the comparison easier to inspect. Neither should require the reader to guess what the values represent.
For the fictional support case, a brief statement might identify the concentration of departures among recent late-shift starters, acknowledge the small number of comments and explain why support access deserves further examination.
That is more useful than a large chart followed by “therefore, training is needed.”
Keep a clear distinction between the supplied record and your interpretation. A reader should be able to trace each important claim back to the relevant evidence.
The qualification-as-a-signal essay helps explain the educational purpose. The qualification should represent judgement that can be inspected, not only a polished document.
A learner's own argument matters. Do not use this fictional report as a ready-made answer, replace its organisation name and present its analysis as your own experience.
Give a stakeholder a conclusion they can inspect
Imagine explaining the recommendation to someone who has not seen your working. Begin with the issue, the most relevant finding and its limits. Then state the proposed next action.
The explanation should survive a reasonable challenge. If asked why training is not yet the first recommendation, the learner should be able to point to the particular uncertainty that remains unresolved.
If asked what could change the recommendation, they should name evidence rather than merely repeat their preferred action.
This is a useful rehearsal for writing. Speaking the logic aloud often reveals a missing connection that polished prose can conceal.
A teacher can ask three neutral questions: what does this figure describe, what conclusion does it support, and why does that conclusion favour this next step? These questions inspect reasoning without supplying an answer to the actual assessment.
The learner should then return to their own task and make the argument in their own words. The aim is not to memorise Rowan Services. It is to build a dependable way of thinking that remains useful when the organisation, evidence and decision change.
CHAPTER 12 OF 15 · Review the reasoning
12. Change one condition and inspect the recommendation
A strong learning check changes the case. Suppose the same total departure figure now comes mainly from temporary contracts ending as planned. The earlier recommendation about late-shift support may no longer address the central issue.
Alternatively, suppose new starters report good access to help but struggle with a particular procedure. A targeted learning intervention could become more relevant than the support-access investigation proposed earlier.
The purpose is not to make the learner guess a hidden correct answer. It is to test whether their reasoning depends on the facts they claim to be using.
Ask which evidence changed, which interpretation it affects and which part of the recommendation should now be reconsidered. Keep unchanged parts of the case visible as well.
The Punggol guide to using case details in an explanation develops the same learning habit in another subject: a detail becomes meaningful when it changes the explanation.
Transfer is present when the learner can make that adjustment independently, rather than reuse the previous recommendation with slightly different wording.
For a changed professional context, the England early-years observation guide follows the move from an observed action to a teaching response. Its practitioner responsibilities are different; the comparison helps you notice how the purpose and limits of evidence change with the decision.
CHAPTER 13 OF 15 · Check independent understanding
13. Try an independent evidence-to-decision check
Take a fresh fictional people-practice problem with a small numerical record and several short comments. State the decision before choosing a method.
Define the measures, perform any necessary calculation and explain its meaning in a sentence. Then identify which comments add useful context and whose experience remains absent.
Write one finding that is directly supported, one interpretation that needs qualification and one question that remains unanswered.
Compare at least two plausible next actions. Explain the criterion that makes one more appropriate at this stage, and state what evidence could change your recommendation.
Finally, choose a review question connected with the intended outcome. Do not substitute attendance or completion for capability without explaining why that measure is relevant.
This is an original learning exercise, not an official assessment checklist. Follow your actual brief and produce your own analysis.
For teaching, the most useful feedback identifies the earliest unsupported link. A correct calculation cannot repair an undefined question, and a clear recommendation cannot repair evidence that does not address it.
CHAPTER 14 OF 15 · Check independent understanding
14. Questions learners often ask
Is more data always better?
No. Relevant, interpretable information is more useful than a large collection whose definitions and scope are unclear. Identify what the decision needs.
Does a percentage establish the cause of a problem?
No. It describes a relationship under a stated definition. Causal interpretation requires further reasoning and suitable evidence.
Can qualitative comments be used alongside numerical records?
Yes, when their source and scope are clear. Comments can add context without representing every employee's experience.
Must I recommend a large intervention?
No such requirement follows from this guide. A proportionate next investigation can be more defensible when the available evidence is limited.
Can I copy the fictional calculations into my assignment?
Use them to understand the method. Your actual task may supply different records and definitions. Submit your own work under the applicable instructions.
Is this an official current assessment brief?
No. It is an educational explanation linked to the public unit specification. Obtain the correct assessment instructions from your study centre.
Does passing the unit prove competence in every workplace decision?
No. A qualification represents particular assessed learning. Applying that learning in a new setting still requires attention to context and evidence.
What should I check first when my report feels unclear?
Check whether the question, measure, finding and recommendation address the same people-practice issue. A break between them often explains the confusion.
CHAPTER 15 OF 15 · Check independent understanding
15. Sources and the next learning decision
Primary sources checked on 24 October 2026: CIPD's Associate Diploma in People Management specification, version 2.6 and evidence-based practice factsheet. The Rowan Services case, calculations, table and practice check are original fictional material.
For the broader organisational setting, read the Organisational Development guide. It retains the wider subject role while this article develops a qualification-linked learning task.
Return to the Education parent when judging whether a course is building transferable capability. A useful next step is to select one recommendation in your own practice work and identify the exact evidence that makes it reasonable.