Your results chapter is finished, your discussion is written, and the limitations and recommendations section is a blank heading with a submission date that will not move. It is the section students leave for last because it feels like confessing your study’s flaws right before an examiner reads them — but a limitations section that is scoped correctly and written with confidence does the opposite: it shows the examiner you understand exactly what your study can and cannot claim, which is a mark of rigour, not weakness.
South African education dissertations have their own recurring limitations — the provincial approval chain, the single-district sample, the self-report survey completed by teachers who know their principal might read it — and examiners recognise genuine, specific versions of these instantly, and generic ones just as instantly. What follows is the difference, five education-specific limitations worked through properly, and a three-tier recommendations structure that finishes the section tonight.
Limitations and delimitations are not the same thing
A limitation is a constraint you did not choose — something outside your control that shapes what your findings can claim: a lower-than-planned response rate, a self-report instrument’s inherent social-desirability bias, an unusually disrupted school term that affected data collection timing. A delimitation is a boundary you chose deliberately when you scoped your study — one province rather than the whole country, Grade 10 learners rather than the whole phase, quantitative survey data rather than a mixed-methods design. Confusing the two is one of the most common and most easily avoided errors in this section: apologising for a delimitation (“I only studied one district”) reads as a weakness you failed to justify, while stating it as a delimitation (“this study was delimited to one district, selected because…”) reads as a deliberate, defensible scoping decision. Sort your list into the two categories before you write a word of the section itself.
Five limitations examiners in South African education research actually recognise
Generic limitations (“more research is needed,” “the sample was small”) read as filler. Specific, field-grounded ones read as evidence you understand your own study.
1. The provincial approval chain constrained your sampling window
Education research in South African public schools runs through a fixed permission sequence — university ethics clearance, then provincial Department of Basic Education research approval, then the principal and school governing body, then parental consent and learner assent — and that sequence commonly takes longer than planned, compressing the time actually available for data collection into a narrower window than the research design ideally called for. If your data collection happened later in the school year than planned, or during a period with competing school priorities (examinations, sports events, staff shortages), state it as a limitation and note specifically what it may have affected — response rate, respondent fatigue, or the representativeness of the specific weeks sampled.
2. Self-report data from teachers or learners carries social-desirability bias
A teacher completing a questionnaire about their own classroom practice, knowing a colleague or the department distributed it, has a documented incentive to describe practice more favourably than it occurs; a learner reporting on a teacher’s conduct has the reverse incentive if anonymity feels uncertain. State this limitation specifically to your instrument and your population, not as a generic disclaimer about all self-report research, and note what steps you took to reduce it (anonymised distribution, an external rather than school-based collector) and what residual risk remains despite those steps.
3. A single-district or single-quintile sample limits generalisability
South African schools vary sharply by quintile, resourcing and language of instruction, and a study set in one district or one quintile band cannot claim its findings generalise across that range without evidence. State the specific quintile, district or language context your sample came from, and name explicitly which other contexts your findings should not be assumed to transfer to — a stronger and more examinable statement than a vague “results may not generalise.”
4. Load shedding or connectivity disruption affected data collection or response quality
Where power or connectivity interruptions affected an online survey’s completion rate, an interview’s scheduling, or a school’s own ability to participate during the study window, this is a genuine, specific South African limitation rather than an excuse — state which activity it affected and, where you can estimate it, how many potential respondents or sessions were lost to it.
5. Language of data collection differed from participants’ home language for some respondents
If your instrument or interviews were conducted in English or Afrikaans while some participants’ home language differed, note this as a limitation affecting response depth or interpretation accuracy, particularly for open-ended or qualitative data, and state what mitigation was used (a bilingual research assistant, translated materials, back-translation) and its residual limits.
Disclosing a limitation is not the same as apologising for one
The tone of this section matters as much as its content, and South African examiners notice the difference between the two immediately. An apologetic register hedges everything and undercuts findings that do not need undercutting: “unfortunately, due to time constraints, this study was only able to…” A disclosing register states the constraint plainly, its likely effect, and moves on: “data collection occurred within a six-week window; this may have limited the diversity of teaching contexts captured.” The second version gives the examiner exactly the information needed to judge your findings’ scope, without inviting doubt about findings the limitation does not actually touch. Read your draft limitations section aloud and strike every “unfortunately,” “sadly” or “the researcher regrets” — if the sentence still makes its point without the apology, the apology was never doing any work.
Writing recommendations in three tiers
South African education examiners expect recommendations sorted by audience, not one undifferentiated list. Structure yours in three tiers, each following directly from a specific finding rather than restating general good intentions.
For practice — what a teacher, school or subject department could act on directly, stated specifically enough to be implemented: not “teachers should receive more training” but “subject heads could schedule a structured peer-observation cycle targeting the specific practice gap this study identified in [named finding].”
For policy — what a district, provincial or national Department of Basic Education decision-maker could consider, scoped to what your evidence actually supports rather than a broader policy position your single study cannot carry alone: a specific curriculum-support resource gap your data identified, not a sweeping national curriculum reform claim.
For future research — the specific next study your findings point to, stated precisely: not “further research is needed” but “a longitudinal design following the same cohort across two academic years would establish whether the effect observed here persists,” naming the exact design gap your study leaves open.
Output: three short paragraphs, each recommendation traceable to a specific finding in your results chapter, none of them a generic statement that could be pasted into any education dissertation unchanged.
A worked example, limitation to recommendation
Limitation: “Data collection was delayed by the provincial approval process, compressing the fieldwork window to the final six weeks of the third term, a period that coincided with examination preparation in the sampled schools. This may have depressed the teacher questionnaire response rate and increased the proportion of rushed, incomplete responses relative to a less time-constrained collection period.” Recommendation for future research: “A study replicating this design with fieldwork scheduled in the first or second term, outside the examination-preparation period, would test whether the response patterns observed here reflect genuine practice or simply reflect the timing constraint under which this study operated.” The recommendation follows directly from the named mechanism in the limitation — not a generic call for more research, but the specific next study the limitation itself points to.
Finish the section tonight, not the week before submission
A limitations and recommendations section written at 11pm under deadline pressure tends to read exactly like that — generic, apologetic, disconnected from the specific study it belongs to — which is precisely the impression a rigorous piece of research does not deserve. Tesify drafts this section once you give it your actual constraints (the approval timeline you experienced, your instrument, your sample’s specific context) and your key findings, structuring limitations against delimitations correctly and building the three-tier recommendations directly from what your results chapter actually found. There is a free plan, and you can have a full draft of this section before the night is over.
Draft your limitations and recommendations section with Tesify tonight
If the discussion chapter above this section is also still unwritten, the four-move paragraph pattern that gets it done is covered in your results are done and the discussion chapter is blank, and the provincial approval chain referenced above is worked through in full in our guide to writing the methodology chapter of an education mini-dissertation.
Frequently asked questions
What does Tesify cost, and is there a free option?
Tesify offers a free plan to get started, with paid tiers for the full drafting and bibliography toolset a complete dissertation needs. Pricing displays in rand at signup, so you can see the free tier’s limits before deciding whether to upgrade.
Will using Tesify for this section count as plagiarism or academic misconduct?
Tesify drafts around the constraints and findings you supply, so the substance is your own study’s specifics, not generic or borrowed content; the final wording and judgment remain yours to review and adjust. Check your specific faculty’s AI-disclosure policy, since South African universities vary on what disclosure they require, and disclose accordingly.
Is my dissertation content kept private if I use Tesify?
Yes, your draft content and any data you input are treated as confidential to your account, not published or shared with third parties. Review the current privacy terms at signup for the specific commitments in force.
How many limitations should an education dissertation list?
Three to six specific, mechanism-stated limitations is typical for an honours or master’s education dissertation; a longer list of generic points is usually weaker than a short list of well-explained specific ones. Quality of explanation matters more than quantity.
Can I list a delimitation as a limitation if I am not sure which it is?
Ask whether the constraint was a choice you made when scoping the study (a delimitation) or something outside your control that affected execution (a limitation). If genuinely unsure, discuss the specific item with your supervisor rather than guessing, since misclassifying several items suggests to an examiner that the distinction was not understood.
Should recommendations for policy be ambitious or conservative?
Scoped tightly to what your specific evidence supports, not ambitious for its own sake. A single honours or master’s study rarely justifies a sweeping national policy recommendation, and an examiner will notice the mismatch between a modest study and an oversized policy claim.