Choosing Your 7CO04 Research Question

No single decision in 7CO04 matters more than the research question. It determines what you read, what design makes sense, whose time and data you need, what analysis is possible and what your conclusions can legitimately claim. A strong question makes every later stage easier; a weak one guarantees months of struggle that no amount of effort at the writing stage can fully repair. Learners who end up referred on this unit have very often lost the project in week one, at the question, without noticing.

The good news is that question design is a learnable skill with a clear method, not a flash of inspiration. This guide covers what makes a question researchable, how to narrow from a broad topic to a precise question, how to test the question against the access you really have, what different question types commit you to, and the common faults to check for before you commit. Your centre may need to approve your question or proposal, so treat everything here as method and check the formal requirements against your own current brief and centre guidance.

In this guide you'll learn

• The four properties that make a question researchable
• The narrowing ladder that takes you from topic to question
• The access test: whose data, whose time, whose permission
• What descriptive, comparative, explanatory and evaluative questions commit you to
• The five common question faults, with rewrites that fix them

What makes a question researchable

A researchable question has four properties. It is specific: it names a defined phenomenon in a defined context, rather than gesturing at a topic area. It is answerable with data you can realistically obtain: the evidence that would answer it exists, and you can reach it within your timescale and permissions. It is bounded: the scope is small enough to investigate properly within the word count and months available, which for most learners means one organisation or a defined part of one. And it is genuinely open: you do not already know the answer, and the question is not phrased so that only one answer is possible.

Openness is the property learners miss most often. A practitioner usually arrives at 7CO04 with convictions, and it is tempting to design a question that will confirm them. But research that cannot surprise you is not research, and markers notice when a question has been built to reach a destination. If every plausible answer to your question would be worth reporting, including the ones you would find uncomfortable, you have an open question.

The narrowing ladder: from topic to question

Most learners start with a topic, and a topic is not a question. The reliable way down is a ladder with four rungs. First the topic: the broad area of interest, such as hybrid working, wellbeing or line manager capability. Second the angle: the specific aspect of the topic you care about, such as its effect on development, retention or trust. Third the population and context: who, where and when, such as early-career staff in one organisation, or line managers in one division. Fourth the question itself: a single sentence that combines angle, population and context into something a study can answer.

Each rung should feel like a loss of grandeur and a gain of precision. The impact of hybrid working becomes hybrid working and professional development, becomes early-career employees at one firm, becomes a question about how those employees experience the effect of hybrid working on their development. The final question looks modest next to the original topic, and that is the point: a modest question answered rigorously is a masters-level project, while a grand topic surveyed loosely is not.

The access test: whose data, whose time, whose permission

Before you commit to a question, put it through the access test, which has three parts. Whose data would answer this question, and does that data exist in a form you can use. Whose time does the study need, how much of it, and why would those people give it to you. And whose permission is required, both formally from the organisation and through your centre's ethics process, and is there any realistic doubt about getting it.

Run the test pessimistically. Surveys get lower response rates than you hope, interviewees cancel, and gatekeepers who are enthusiastic in conversation can slow down when a formal request arrives. A question that only works if everything goes smoothly is a question that does not work. The strongest projects are usually built on access the learner already has, which is why researching your own organisation is so common on this unit, and even then the permission must be real, documented and obtained before any data is collected.

Question types and what they commit you to

The grammar of your question commits you to a kind of study, so it pays to know what you are signing up for. A descriptive question asks what is happening: what are the experiences, patterns or practices in a context. It commits you to systematic measurement or careful qualitative description, and it suits surveys, interviews and secondary data. A comparative question asks how two or more groups, sites or periods differ. It commits you to collecting equivalent data on each side of the comparison, which quietly doubles the access you need. An explanatory question asks why something is happening or what relationship exists between things. It is the most ambitious type, because claims about causes and relationships need designs and analysis strong enough to carry them, and masters projects often need to soften pure explanation into exploration. An evaluative question asks how well something is working against some criterion. It commits you to defining that criterion defensibly before you judge anything against it, which is harder than it sounds.

There is no best type. The mistake is mismatch: asking an explanatory question and then running a study that can only describe, or asking an evaluative question without ever saying what good would look like. Choose the type whose commitments you can actually meet.

Common question faults, with rewrites

Five faults account for most weak 7CO04 questions, and all five can be fixed by rewriting rather than starting again. Too broad: the question spans whole fields, such as what is the impact of technology on HR, and the fix is the narrowing ladder, pinning an angle, a population and a context. Double-barrelled: the question is really two questions wearing one question mark, such as what do employees think of the appraisal system and how should it be redesigned, and the fix is to keep the researchable half and move the other into your aims or recommendations. Unanswerable: the question needs data you cannot get, such as comparing confidential practices across competitor firms, and the fix is to rebuild it around the access you have. Foregone conclusion: the question presumes its own answer, such as why does poor management cause turnover, and the fix is to reopen it, asking what factors are associated with turnover and letting management take its chances with the evidence. Consultancy brief: the question asks what the organisation should do, which is a deliverable rather than a research question, and the fix is to ask what is actually happening first and let recommendations follow from findings.

The last fault deserves emphasis because it looks so professional. Practitioners are trained to move straight to solutions, and a question like how can we improve onboarding feels natural. But research earns its recommendations. Ask what new starters actually experience and where the process diverges from its design; the improvement question then answers itself from evidence rather than instinct.

A method for developing and stress-testing your question

1. Write down your topic and, in one sentence, why it matters to an organisation you can access; if you cannot name the organisational stake, the topic is a reading interest, not a project.
2. Walk the narrowing ladder: fix an angle, then a population and context, then draft the question as a single sentence.
3. Classify the draft as descriptive, comparative, explanatory or evaluative, and check you can meet what that type commits you to.
4. Run the access test pessimistically: whose data, whose time, whose permission, and what happens to the project if a third of it falls through.
5. Check the four properties: specific, answerable, bounded, genuinely open, and rewrite until all four hold.
6. Test it against the five faults: too broad, double-barrelled, unanswerable, foregone conclusion, consultancy brief.
7. Draft two or three aims underneath the question and confirm they cover it without smuggling in a second project.
8. Say it aloud to a colleague who knows the organisation and a friend who does not; if either cannot repeat back what you are investigating, simplify.
9. Sit with it for a week before committing, because the flaws in a question tend to surface once the excitement of having one fades.
10. Take it through whatever approval your centre requires before building anything else on it.

Worked examples: weak questions rewritten

These examples are invented and deliberately simplified, but the pattern of each rewrite is real. A learner interested in wellbeing drafts how does wellbeing affect performance. It is too broad and quietly explanatory, demanding causal evidence no masters project can gather. Rewritten down the ladder it becomes: how do employees in one contact centre experience the relationship between the organisation's wellbeing programme and their day-to-day performance. Descriptive, bounded, answerable through interviews she can arrange.

A learner in resourcing drafts why is our recruitment process failing to attract diverse candidates. It is a foregone conclusion twice over, assuming both failure and its location. Rewritten open, it becomes: what factors influence candidates' decisions to apply, or not to apply, for roles at the organisation, with attraction of underrepresented groups as a stated aim. Now the data can speak, and if the process is indeed the problem, the findings will show it with authority instead of assumption.

A learner in learning and development drafts how should the organisation redesign its induction programme. That is a consultancy brief. Rewritten as research it becomes: how do recent joiners experience the current induction programme, and where do their experiences diverge from its intended design. The redesign question has not been abandoned; it has been moved to the recommendations chapter, where it can stand on findings.

Where to go next

Once your question is settled, the hub article, 7CO04 Assignment: The Complete Guide, walks through the rest of the project, from literature and design through ethics, analysis and recommendations, and 7CO04 FAQs: Your Questions Answered deals quickly with the questions learners ask most. For building the reading base your question deserves, Finding Credible Sources for CIPD Assignments with Google Scholar covers the practical craft of literature searching.

If you would like tailored, ethical help at this stage, our 7CO04 support includes question scoping and feasibility coaching, methods guidance and careful review of your proposal drafts. The question you choose, and the research you build on it, are always your own.