How to Write Recommendations for Future Research in 2026: A Complete Guide
Knowing how to write recommendations for future research is the difference between a dissertation that feels finished and one that feels genuinely complete. This section — tucked at the end of your discussion or conclusion — is where you stop being a reporter of results and become a contributor to your field: you tell the academic community where to look next, and why. Examiners read it carefully. A vague list of stock phrases will land badly; four or five precise, evidence-rooted proposals will leave a strong final impression.
The good news is that this is one of the most formulaic sections in any thesis. Once you understand the three core types of recommendation, the five-step derivation process, and a handful of proven sentence starters, you can write a polished future-research section in a single focused session. This guide walks you through every stage, with worked examples and templates you can adapt straight away.
What Are Recommendations for Future Research?
Video: Dissertation Conclusion Chapter: 6 Simple Steps + Examples — Grad Coach
Recommendations for future research are short, evidence-based proposals that point other scholars toward the gaps your study could not close. They are distinct from the rest of your conclusion in a crucial way: your conclusion summarises what you found; your recommendations explain what someone else should do next, and why your work makes that next step worthwhile.
Think of them as passing the baton. You have run your leg of the race — you have defined a problem, collected data, and drawn conclusions. The recommendations section formally hands the baton to the next researcher, complete with a note about which direction to run in.
Recommendations also serve a second function that students often overlook: they signal intellectual maturity. Acknowledging that your study has boundaries and that the field still has open questions is not an admission of weakness — it is evidence that you understand your place in a larger conversation. Oxford and Cambridge examiners consistently cite well-constructed recommendations as a marker of a strong final chapter.
Where Do Recommendations Go in a Thesis?
Placement depends on the length and nature of your project. The three most common conventions are:
| Project type | Typical placement | Approximate length |
|---|---|---|
| Undergraduate dissertation (8,000–12,000 words) | Final paragraphs of the conclusion | 150–300 words |
| Master’s dissertation (15,000–25,000 words) | End of discussion chapter or start of conclusion | 300–600 words |
| PhD thesis (60,000–100,000 words) | Dedicated sub-section or standalone chapter | 600–1,200 words |
If your dissertation separates the discussion from the conclusion into two distinct chapters — which is the dominant convention at UK research universities — the recommendations section sits within or immediately after the discussion. If your department combines them into a single “Discussion and Conclusion” chapter, recommendations come near the end of that combined chapter, just before your closing paragraph. For guidance on that structural distinction, see our article on how discussion and conclusion chapters differ.
Before drafting recommendations, it helps to revisit what you found — see the guide on how to write a research limitations section for how every limitation maps naturally to a future-research proposal.
The Three Core Types of Recommendation
Most future-research recommendations fall into one of three broad categories. A strong recommendations section typically includes proposals from at least two categories, and often all three.
1. Theoretical Recommendations
These propose that future researchers develop, test, or refine a theoretical framework. They are most common in social sciences, humanities, and business research. A theoretical recommendation typically follows from a finding that the existing theory only partially explained your results — perhaps the model you applied was developed in a different cultural or temporal context.
2. Methodological Recommendations
These suggest that future studies use a different or improved research design to address a question that your methodology could not fully answer. Common prompts include a small or non-representative sample, reliance on self-report measures, a cross-sectional design where longitudinal data would be more informative, or a single-site study that may not generalise.
3. Practical and Policy Recommendations
These direct practitioners, organisations, or policymakers to act on your findings. They are most prominent in applied disciplines such as nursing, education, public health, social work, and business management. A practical recommendation is grounded in a specific finding and names a concrete action and a concrete actor.
Six Specific Future-Research Categories
Beyond the three broad types, the academic literature identifies six more granular categories that help you pinpoint what kind of future study would be most valuable. Drawing on the taxonomy developed by Laerd Dissertation, these are:
- Addressing a research limitation — directly proposing a study that fixes a design flaw you acknowledge in your limitations section (e.g. a larger sample, a different data-collection instrument).
- Building on a particular finding — exploring an unexpected result that opened a new question you did not set out to answer.
- Testing a framework for the first time — recommending that a conceptual model you constructed or synthesised be empirically validated in a real-world setting.
- Testing a framework in a new context — proposing replication of your study in a different geography, culture, sector, or population group.
- Expanding a framework — suggesting that future researchers add variables, mediators, or moderators that your study excluded for scope reasons.
- Re-evaluating a framework — calling for reassessment of a theory in light of a significant external change (new legislation, a technological shift, a social event) that emerged during or after your study.
Using this taxonomy as a checklist when you review your own findings and limitations ensures that you are not simply listing generic suggestions but offering targeted proposals that advance the field.
The Five-Step Derivation Process
The most reliable way to generate strong recommendations is to derive them systematically from your findings and limitations, rather than inventing them at the end. This five-step process works for any discipline.
Step 1: List Your Limitations
Open a fresh document and write down every limitation you acknowledged in your thesis — sample size, sampling strategy, geographical scope, timeframe, measurement instruments, access constraints, and theoretical boundaries. Each limitation is a seed for a recommendation.
Step 2: Identify Your Unanswered Sub-Questions
Review your research questions and look for aspects you could not fully address. Perhaps you found a relationship between two variables but lacked the data to determine causality. Perhaps an unexpected pattern emerged in a sub-group you had not planned to analyse. Write these down alongside the limitations.
Step 3: Map Gaps to Recommendation Types
For each limitation or unanswered question, decide which recommendation type it generates. A sample limitation → methodological recommendation. An unexpected finding → “building on a finding” recommendation. An untested theoretical model → theoretical recommendation.
Step 4: Draft Each Recommendation in Two Sentences
Each recommendation needs (a) a rationale rooted in your findings, and (b) a clear, actionable proposal. Two sentences is usually enough. Resist the urge to write a paragraph — examiners prefer precision over expansion here.
Step 5: Order by Priority
Place your most significant recommendation first. The most significant is usually the one that most directly advances the central argument of your thesis, or the one that addresses your most consequential limitation. Ordering by importance signals academic judgement and makes the section easier to read.
This process connects naturally with the results you present in your data chapters. Before drafting recommendations, it helps to revisit what you found — and if you need a refresher on structuring those findings clearly, the guide on how to write your thesis results chapter covers organising quantitative and qualitative data in detail.
Phrasing Templates and Sentence Starters
One of the most practical things you can do before writing this section is to arm yourself with a set of reliable sentence starters. Below are templates for each recommendation type, drawn from analysis of successfully defended dissertations at UK and US research universities.
Opening Frames (to establish the basis for the recommendation)
- “Based on the findings of this study, it is recommended that…”
- “The results of this research suggest that future studies should…”
- “Given the limitations identified in Section [X], subsequent research would benefit from…”
- “This study’s finding that [result] raises the question of whether…, warranting further investigation into…”
- “The present study was constrained by [limitation]; future research should therefore…”
Theoretical Recommendation Templates
- “Future researchers should examine whether [theory/framework] extends to [new context/population], given that the current study tested it exclusively within [original context].”
- “Subsequent work might explore the theoretical relationship between [variable A] and [variable B], which this study could not fully establish due to [reason].”
- “It would be valuable to test [named theoretical model] longitudinally to determine whether the observed effects persist over time.”
Methodological Recommendation Templates
- “Future studies should employ a larger, more representative sample to confirm whether the patterns observed here generalise to [broader population].”
- “A longitudinal design would enable researchers to establish temporal precedence between [variable A] and [variable B], which the cross-sectional nature of this study precluded.”
- “Incorporating objective measures alongside the self-report instruments used in this study would improve construct validity and address the social-desirability bias identified as a limitation.”
- “A mixed-methods approach — combining the survey data analysed here with semi-structured interviews — could illuminate the mechanisms underlying [finding].”
Practical/Policy Recommendation Templates
- “[Organisation/profession] should consider [specific action] in light of this study’s evidence that [finding].”
- “Policymakers in [sector] are advised to [action], particularly given this study’s finding that [result] was associated with [outcome].”
- “Practitioners working with [population] may wish to incorporate [intervention/approach], as the data indicate that [finding].”
Worked Examples by Discipline
The following examples show how a study limitation maps to a specific recommendation across four disciplines. Each example follows the two-sentence formula: rationale first, then proposal.
Psychology (Undergraduate dissertation)
Limitation: “Participants were recruited via social media, resulting in a self-selected sample skewed toward younger, digitally active adults.”
Recommendation: “Future research should recruit participants through community organisations and GP surgeries to achieve a more age-diverse sample. This would allow researchers to determine whether the link between social media use and anxiety symptoms reported here holds across older adult populations.”
Business and Management (Master’s dissertation)
Limitation: “The study focused on a single mid-sized UK retail firm, limiting the generalisability of findings to other sectors or national contexts.”
Recommendation: “Subsequent studies should replicate this analysis across multiple industries and at least two national contexts — for example, the UK and Germany — to test whether organisational agility correlates with employee retention in the same way across different institutional environments.”
Public Health (Master’s dissertation)
Finding: “Peer-support group attendance was the strongest independent predictor of continued medication adherence at 12-month follow-up.”
Recommendation: “NHS trusts should pilot the integration of structured peer-support sessions into standard medication review appointments for patients with long-term conditions. A randomised controlled trial with 18-month follow-up would be needed to establish whether this association reflects a causal pathway.”
Environmental Science (PhD thesis)
Limitation: “Soil carbon measurements were confined to a single growing season, preventing assessment of inter-annual variability.”
Recommendation: “Future field campaigns should capture at least three consecutive growing seasons across climatically contrasting years to distinguish genuine treatment effects from inter-annual climate-driven fluctuations in soil carbon stocks. Multi-year datasets would substantially increase confidence in the flux estimates reported here.”
Common Mistakes and How to Avoid Them
Even strong researchers stumble on this section. Here are the seven most common errors, and the fix for each.
1. Generic, unjustified suggestions
Mistake: “Future research should examine this topic further using a larger sample.”
Fix: Always anchor the recommendation to a specific finding or limitation: why does the sample need to be larger, and what would a larger sample reveal?
2. Introducing new concepts
Recommendations must be grounded in your study. Do not introduce theories, variables, or concepts that you never discussed in your literature review or analysis — examiners will flag this as a structural error.
3. Only one type of recommendation
Writing only methodological recommendations (e.g. “use a bigger sample next time”) suggests limited analytical depth. A balanced section includes at least two types.
4. Proposing unfeasible research
Recommending a 20-year longitudinal cohort study for a topic that could be addressed with a well-designed three-year trial undermines your credibility. Apply the SMART framework (see below) to every proposal.
5. Repeating the limitations section verbatim
Limitations describe what your study did not do; recommendations describe what a future study should do. The framing shifts from retrospective to prospective. Do not simply restate the limitation without converting it into a forward-facing proposal.
6. Over-long paragraphs
Each recommendation should be no more than two to four sentences. Long paragraphs dilute impact. If you find yourself writing a full paragraph for one recommendation, split it into two distinct proposals.
7. Vague actor attribution
For practical and policy recommendations, name the stakeholder. “It should be recommended that steps are taken” is meaningless. “NHS clinical leads” or “secondary school curriculum designers” or “future experimental researchers at institutions with access to longitudinal cohort data” are meaningful.
The SMART Check for Every Recommendation
Before finalising your recommendations section, run each proposal through the SMART framework, which is widely endorsed by academic supervisors at institutions including UCL, the University of Edinburgh, and Harvard’s Graduate School of Arts and Sciences as a quality-check for actionable academic writing.
| Criterion | Question to ask yourself | Fix if failing |
|---|---|---|
| Specific | Is the proposed study precisely defined in terms of population, method, or variable? | Replace “a larger study” with “a stratified sample of at least 500 participants across three regions” |
| Measurable | Could a future researcher operationalise this recommendation clearly? | Add outcome criteria or measurable targets where possible |
| Achievable | Is this realistically feasible given current methods and resources? | Scale back proposals that require resources unavailable to most research groups |
| Relevant | Does this recommendation connect directly to your findings or limitations? | Delete any recommendation that cannot be traced back to a specific finding or limitation |
| Time-bound | Can a plausible timeframe be implied (e.g. longitudinal vs. cross-sectional, short-term pilot vs. multi-year trial)? | Specify a design type that implies duration even if you cannot name an exact number of years |
A Full Recommendations Section Template
Below is a ready-to-adapt template that shows how four recommendations fit together as a cohesive section. Replace the bracketed placeholders with the specifics of your own study.
Recommendations for Future Research
Several directions for future inquiry emerge from the findings and limitations of this study. First, and most immediately, [Methodological recommendation: propose a study that addresses your primary limitation — e.g. a longitudinal design, a larger or more diverse sample, or an alternative measurement instrument]. This would enable researchers to [specific outcome this makes possible].
Second, the unexpected finding that [unanticipated result] warrants further investigation. Future research should examine [what the next study should do to explore this finding], preferably drawing on [type of data or method that would be most informative].
Third, this study tested [theoretical framework] exclusively within [your context]. Researchers working in [alternative context] should examine whether [the same relationships / the model’s predictions] hold under different [cultural / sectoral / regulatory] conditions, thereby contributing to the generalisability of [theory].
Finally, [Practical/policy recommendation: name a specific actor and action grounded in your most significant finding]. A [pilot study / randomised trial / systematic evaluation] would provide the rigorous evidence base needed before widespread adoption.
Notice how each paragraph is distinct — it is grounded in a different aspect of the study, it names a specific type of future research, and it explains why that research would be valuable. That variety is what separates a strong recommendations section from a list of generic afterthoughts.
Putting It All Together: A Checklist
Before you submit, run through this checklist for your recommendations section:
- Every recommendation traces back to a specific finding or acknowledged limitation — not invented independently.
- At least two of the three core types (theoretical, methodological, practical/policy) are represented.
- Each recommendation is two to four sentences maximum.
- Practical/policy recommendations name a specific stakeholder group.
- No new concepts are introduced that did not appear earlier in the thesis.
- You have proposed four to six recommendations in total (three is thin; seven or more is unwieldy).
- The section reads in prospective, forward-looking language (“should”, “would”, “could”) rather than retrospective language.
- Every recommendation passes the SMART check.
This is the final analytical move in your dissertation. By the time an examiner reaches this section, they want to feel that your research has genuinely advanced the conversation — and a confident, precise recommendations section does exactly that. For a full picture of how this section fits into your closing chapters, see the structural differences between your discussion and conclusion chapters — understanding that boundary is the clearest map to placing your recommendations correctly.
Frequently Asked Questions
How many recommendations for future research should I include?
Four to six recommendations is the widely accepted range for a Master’s-level dissertation. Undergraduate dissertations typically include three to four, while PhD theses may include five to eight across a dedicated section. Fewer than three signals limited reflective depth; more than seven usually indicates insufficient prioritisation. Quality and specificity matter far more than quantity — four well-reasoned, evidence-rooted recommendations will always outperform eight generic ones.
What is the difference between limitations and recommendations for future research?
Limitations describe what your current study did not or could not do — they look backwards at your methodology. Recommendations for future research look forwards: they translate each limitation (and each unexpected finding) into a specific proposal for what a subsequent researcher should do. Every limitation should ideally generate at least one recommendation, but recommendations are not limited to limitations — they also emerge from novel or unanticipated findings.
Should recommendations be in the discussion or conclusion chapter?
For most UK and US universities, recommendations appear at the end of the discussion chapter or at the start of the conclusion chapter. In shorter undergraduate dissertations where discussion and conclusion are combined, they appear near the end of that combined chapter. For longer Master’s and PhD projects at institutions such as UCL, Edinburgh, or Harvard, a standalone “Recommendations and Future Research” sub-section within the conclusion is common. Always check your department’s guidelines first — some disciplines (such as clinical medicine) have specific conventions.
Can I recommend future research even if my study had strong findings?
Absolutely — in fact, strong findings often generate the most interesting recommendations. Robust results in one context immediately invite replication in other contexts. Significant correlations invite causal designs. Newly confirmed theories invite application to different populations. The recommendations section is not a remediation exercise for a weak study; it is a forward-looking intellectual contribution regardless of your results. Even a near-perfect study operates within a bounded scope, and articulating what lies beyond that scope demonstrates exactly the kind of scholarly awareness examiners reward.
How do I write a recommendation for future research when I only have one limitation?
If you genuinely have only one significant limitation, supplement your methodological recommendation by reviewing your findings for unexpected sub-patterns (generate a “building on a finding” recommendation), checking whether your theoretical framework was applied in a limited context (generate a contextual replication recommendation), and considering whether your findings have practical implications for a specific stakeholder group (generate a practical recommendation). You will almost always find at least four substantive directions even when your core limitation is singular.
Is it acceptable to recommend that my own approach be replicated by future researchers?
Yes — recommending contextual replication is a legitimate and frequently valued category of future-research proposal. If your study tested a model in the UK, recommending replication in Australia or Canada is a valid intellectual contribution because it points toward establishing the generalisability of your findings. The key requirement is that you articulate why replication in that specific context would be informative — not just that “more research is needed.”
Ready to write your dissertation’s final chapters?
Tesify is an AI writing assistant built for university students. It helps you structure your discussion and conclusion chapters — including the recommendations section — with guidance tailored to your discipline, university, and citation style. Thousands of students at UK and US universities have used it to finish their thesis with confidence.
Write your thesis with AI
Structure, draft, cite, and format your thesis faster with Tesify’s AI writing tools, automatic bibliography, and plagiarism checker. Free to start, no credit card required.
Leave a Reply