You already know that examiners care about interpretation, not just numbers. This guide is the next step — exactly how to write that interpretation, sentence by sentence, for every result type across every engineering branch. Not theory. Not what good writing looks like. What to actually type.
Fig. 1 — How to write results and discussion 2026: the presentation rules, the 4-move discussion framework, branch-specific examples, and the exact sentences that separate an average chapter from a distinction-level one
Writing the results and discussion chapter is not about showing how much data you collected. It is about proving you understand what your data means. Results present your findings factually — measured values, graphs, and tables — without interpretation. Discussion explains the engineering mechanism behind each finding, compares it to a published standard or benchmark, and states honestly where the conclusion holds and where it stops. Every sentence in discussion should answer one question: so what does this tell an engineer?
- What This Chapter Is Actually For — And Why Most Students Write It Wrong
- The Boundary Between Results and Discussion — Where Each Sentence Goes
- How to Present Data That Examiners Can Actually Evaluate
- The 4-Move Discussion Framework — Applied to Every Branch
- Real Before-and-After Examples — The Exact Rewrite
- The "So What" Test — One Question That Fixes Every Weak Paragraph
- The Six Mistakes That Cost the Most Marks
- Frequently Asked Questions
There is a companion guide to this one — How External Examiners Evaluate Project Results and Conclusions — which explains the examiner's mindset in depth: why they treat results as behavioural evidence, how they evaluate risk in conclusions, and why technically correct projects still receive average grades. If you have not read it, the short version is this: examiners are not checking your numbers. They are checking whether you understand what your numbers mean.
This guide picks up where that one ends. Once you understand what examiners look for, the next problem is practical — how do you actually write a chapter that delivers it? What sentence structure works? What do you write after presenting a graph? How do you compare to a standard without it sounding like you are just listing values? How do you acknowledge a limitation without it sounding like an apology?
Those are the questions this guide answers, with real examples from real engineering branches, written at the level of specificity that actually helps someone sitting down to write this chapter today.
Section 01What This Chapter Is Actually For — And Why Most Students Write It Wrong
The results and discussion chapter has one job: to show that your data answers your research question, and that you — the engineer who collected it — understand what the answer means in practice.
Most students write it as a data reporting exercise. They present every graph, describe every trend, and move on. The chapter ends up being a catalogue of what happened, with no explanation of why it happened and no connection to what it means for the engineering system being studied. An examiner reading that chapter learns nothing about the student. They see output, not understanding.
The mistake is not carelessness — it is a misunderstanding of what the chapter is supposed to demonstrate. You are not submitting evidence that you ran the experiments. You are submitting evidence that you understood what the experiments showed. Those are different things, and the writing reflects which one you did.
Section 02The Boundary Between Results and Discussion — Where Each Sentence Goes
The most consistent structural problem in this chapter is mixing results and discussion. Students interpret inside the results section, or re-describe data inside the discussion section. Both weaken both sections. The boundary is simple and absolute: results report facts, discussion explains them.
| Sentence Type | Belongs in Results | Belongs in Discussion | Example — Wrong / Right |
|---|---|---|---|
| Measured value with condition | Yes — state the number, unit, and test condition | No — do not repeat numbers already in the table | Wrong in discussion: "As shown in Table 3, the tensile strength was 3.8 MPa." | Right in discussion: "The 31% increase over control reflects enhanced crack-bridging at 0.5% fibre content." |
| Graph description | Brief — one sentence directing reader to the key trend | No — never re-describe what is already visible in the figure | Wrong in discussion: "Figure 4 shows compressive strength increasing with curing age." | Right in discussion: "The accelerated gain in Mix B reflects the pozzolanic reaction of fly ash, which proceeds more fully by 28 days than Portland cement hydration." |
| Comparison to standard or study | No — this is interpretation, not reporting | Yes — your value, the benchmark, the gap, and what the gap means | Right in discussion only: "BOD removal of 87% exceeds the CPCB Class B standard of 80%, confirming the system meets statutory requirements for river discharge without additional polishing treatment." |
| Engineering mechanism explanation | No — belongs in discussion | Yes — explain the physical, chemical, or computational principle behind the pattern | Right in discussion only: "The baffle plate forces flue gas through a longer contact path, increasing the number of transfer units (NTU) and driving the observed 13-point efficiency improvement." |
| Unexpected result | Yes — report it. Do not suppress it. | Yes — specific possible explanation; honest uncertainty if cause is unclear | Results: "Sample 4 showed 40% lower strength than the batch mean." | Discussion: "The anomaly is attributed to a visible surface void identified post-testing, consistent with localised aggregate segregation during casting." |
If your institution allows it, write a combined Results and Discussion chapter — present one result, discuss it immediately, then move to the next. This is more readable than separating all results from all discussion, and it prevents the most common problem: a discussion section that re-describes findings the examiner read fifteen pages ago.
Section 03How to Present Data That Examiners Can Actually Evaluate
Data presentation is not formatting — it is communication. A graph without labelled axes, a table without units, or a figure you never reference in text are not minor oversights. They tell the examiner you have not thought about how your evidence is read. These are the rules that matter, and why they matter.
Tables — Caption Above, Reference Always
Caption goes above the table, not below. Every column header must include units in brackets — "Compressive Strength (MPa)", not just "Strength". Before every table in your text, write one sentence that directs the reader to what matters: "Table 3 presents the effect of w/c ratio on 28-day compressive strength across three mix proportions." After the table, your discussion begins. Never present a table and then say nothing about it in the text — an unreferenced table is invisible to an examiner scanning your chapter.
Graphs — Axes, Legend, Error Bars
Both axes labelled with units — no exceptions. If you have more than one data series, a legend is not optional. If you repeated measurements, error bars are not optional — they tell the reader how much to trust the trend. A graph without labelled axes cannot be read, verified, or cited. It will be flagged. A graph with error bars that are small and consistent tells the examiner your methodology was controlled. A graph with error bars that are large and erratic tells the examiner to ask you about it in viva — make sure you have an answer.
Statistical Reporting — The One Sentence Most Students Skip
If you ran any test more than once, report mean ± standard deviation. If you used a statistical comparison test — ANOVA, t-test, chi-square — report the test statistic and p-value, and then write one sentence explaining what it means: "The difference between Mix A and Mix B was statistically significant (F = 8.34, p = 0.02), confirming that the modified water-cement ratio produced a reproducible and non-random improvement in compressive strength." That sentence is worth more than a full paragraph of trend descriptions.
Unlabelled graph axes. Figures presented in the chapter but never referenced in text — "as shown in Figure 5" with no explanation of what Figure 5 shows. Results presented in a different order than your methodology described them. Raw Excel tables pasted without formatting. Fix all four before you submit. They signal that the chapter was assembled rather than written.
Section 04The 4-Move Discussion Framework — Applied to Every Branch
Every strong discussion paragraph — every single one, regardless of branch, topic, or result type — makes four moves. In order. You can vary the sentence structure, but the four moves must all be present. If any one is missing, the paragraph is incomplete from an examiner's perspective.
Move 1 — Anchor the Finding
One sentence. The specific value, the condition it was measured under, and the comparison that makes it meaningful. Not "strength improved" — but "the modified mix achieved 42.3 MPa at 28 days, a 22% improvement over the control mix (34.7 MPa), exceeding the M35 design target of 35 MPa specified in IS 456:2000." The anchor sentence does three things: states the fact, quantifies the change, and immediately contextualises it against something the reader can evaluate.
Move 2 — Explain the Mechanism
Why did this happen? Name the engineering or scientific principle that caused the result. This is the move most students skip — and it is the one examiners are most specifically looking for. "The strength improvement is attributed to the secondary pozzolanic reaction of the 15% fly ash replacement, which generates additional calcium silicate hydrate at later ages — evidenced by the steeper strength gain curve between 7 and 28 days relative to the control mix." You cannot write this sentence by reading your data table. You have to understand your project. That is the point.
Move 3 — Compare to Published Evidence
Is your result consistent with what others have found, or does it differ — and if it differs, why? "This finding aligns with Naik et al. (2018), who reported 18–25% strength gains for similar fly ash proportions in M30 concrete. The higher value in the present study may reflect the lower w/c ratio of 0.42 compared to 0.47 in the referenced work, which reduces total porosity and increases effective binder density." If your result is different from published work, that is a finding worth explaining. If it confirms published work, that is validation worth stating.
Move 4 — State Implication and Limit
What does this mean for engineering practice, and where does that meaning stop? "The result supports the use of 15% fly ash replacement in M35-grade structural concrete for moderate exposure applications as defined by IS 456:2000 Table 3. Results are limited to specimens cured at 27°C ± 2°C — higher field curing temperatures would alter hydration kinetics and should be validated separately before specifying this mix for outdoor construction in high-temperature climates." A limitation stated this specifically does not weaken your project. It demonstrates that you understand your own evidence.
| Branch | Move 1: Anchor | Move 2: Mechanism | Move 3: Compare | Move 4: Implication + Limit |
|---|---|---|---|---|
| Civil / Structural | Deflection reduced 23% with CFRP wrap at same load | CFRP increases effective moment of inertia, raising flexural rigidity without added dead load | Consistent with ACI 440.2R-17 predicted range for CFRP-wrapped RC beams | Supports CFRP as retrofitting option for under-strength beams. Static loading only — fatigue and seismic behaviour untested |
| Mechanical / Thermal | Thermal efficiency improved from 31% to 44% after baffle addition | Longer flue gas path increases NTU; LMTD increases due to improved temperature difference profile | 44% within the 40–47% range reported by Kumar et al. (2021) for equivalent baffle configurations | Applicable to industrial waste heat recovery. Efficiency improvement measured at rated flow — performance at 3× rated flow not tested |
| Electrical / Power | Power factor improved from 0.72 to 0.91 with capacitor bank | Capacitor bank supplies reactive power locally, reducing lagging current draw from the supply and cutting I²R losses in the distribution cable | 0.91 exceeds IE Rules 2023 minimum of 0.85 — statutory requirement met | Viable for the tested industrial load profile. Assumes stable load — variable or non-linear loads require dynamic reactive compensation |
| Environmental | BOD removal = 87% at 6-hour hydraulic retention time | Extended aeration maintained DO above 2 mg/L throughout, sustaining aerobic microbial activity at the tested organic loading rate | Exceeds CPCB Class B discharge standard of 80% — suitable for river discharge | Validated at tested influent BOD of 250 mg/L. High-strength industrial effluent above 500 mg/L may require pre-treatment |
| CS / Machine Learning | Crop disease classifier: 91.3% test accuracy, F1 = 0.91 (class-weighted) | MobileNetV2 transfer learning retains low-level texture features from ImageNet that are relevant to leaf disease morphology — fine-tuning preserves these while adapting to crop-specific patterns | Lab accuracy comparable to Mohanty et al. (2016). Field accuracy drops to 74% — consistent with lighting and background variation reported in field-deployment literature | Viable for controlled lab screening. Field deployment requires training augmentation with field images under Indian lighting conditions |
Section 05Real Before-and-After Examples — The Exact Rewrite
The difference between a pass-grade and a distinction-grade discussion paragraph is almost never the data — it is the writing. These examples use identical data. The strong version does not add new information. It adds interpretation of the information that was already there.
Civil Engineering — Concrete Strength
"The compressive strength of the modified mix was 42.3 MPa at 28 days. This is higher than the control mix which showed 34.7 MPa. The results are shown in Table 4. The modified mix performed well in all tests and met the required grade."
"The modified mix achieved 42.3 MPa at 28 days — a 22% improvement over the control (34.7 MPa), exceeding the IS 456:2000 M35 target of 35 MPa and satisfying requirements for moderate exposure conditions. The strength gain reflects the secondary pozzolanic reaction of the 15% fly ash replacement, which generates additional C-S-H gel at later ages. This is consistent with the steeper 7-to-28-day gain curve in Mix B relative to the control, where hydration was largely complete by day 14. Results are valid for the tested curing regime of 27°C and 95% RH — elevated field temperatures would accelerate early strength gain but may reduce long-term performance, which should be verified for outdoor applications."
Electrical Engineering — Power Factor Correction
"The power factor before correction was 0.72. After adding the capacitor bank, the power factor improved to 0.91. This shows that the capacitor bank was effective. The system now meets the required standard."
"The capacitor bank improved power factor from 0.72 to 0.91, exceeding the IE Rules 2023 minimum of 0.85 for industrial loads and reducing reactive current draw by approximately 37%. At the tested load of 45 kVA, this corresponds to a reduction in apparent power demand of 6.3 kVAR — directly reducing I²R losses in the 25-metre distribution cable between the panel and the load. The improvement was measured at steady-state loading; under fluctuating or non-linear loads such as variable-speed drives, dynamic reactive compensation would be required to maintain this correction across all operating conditions."
Section 06The "So What" Test — One Question That Fixes Every Weak Paragraph
After writing any discussion paragraph, read it back and ask: so what? What should a practising engineer do differently because of this finding? If your paragraph cannot answer that question, it is not finished.
This test exposes the most common type of weak discussion — the paragraph that explains a result but stops short of connecting it to any engineering consequence. "The modified beam showed lower deflection" is a result. "The lower deflection indicates the CFRP wrap is sufficient to upgrade the beam from serviceability class S2 to S1 under IS 456 Table 16, removing the need for re-plastering of the ceiling below" is an engineering finding. The data is identical. The difference is whether you followed the result to its practical conclusion.
The "so what" test also catches overclaiming. If your paragraph answers "so what" with something the data does not actually support — "this design can be used for all industrial applications" when you only tested one load case at one temperature — you have crossed into the risk territory that the companion guide describes. The answer to "so what" must be traceable to your actual results, under your actual test conditions.
Before moving to the next result, confirm your discussion paragraph has: an anchor sentence with a specific value and comparison benchmark; a mechanism sentence naming the engineering principle that explains the result; a comparison to a published study or standard with explicit reference; an implication statement answering "so what for an engineer"; and a limitation that defines the specific condition under which the conclusion holds. Five elements. Every paragraph. Non-negotiable.
Section 07The Six Mistakes That Cost the Most Marks
| Mistake | Why It Actually Costs Marks | The Specific Fix |
|---|---|---|
| Describing graphs instead of interpreting them | The examiner can see the graph. "The graph shows an upward trend" tells them nothing they did not already know. It proves you noticed the trend but not that you understand what caused it. | Write the engineering mechanism that produced every visible trend. One sentence per trend. If you cannot name the mechanism, that is a gap in your project understanding — not a writing problem. |
| Results with no comparison baseline | A number without a reference point cannot be evaluated as good, acceptable, or poor. The examiner cannot grade a result that exists in isolation. | Every key result needs a comparison — a published standard (IS, ASTM, CPCB, IEEE), a previous study, or your own control group. State the comparison explicitly and interpret the gap. |
| Vague limitations | "A larger sample size would improve results" is true of every engineering project ever conducted. It demonstrates awareness of the concept of limitations, not understanding of this specific one. | Name the specific condition under which your conclusion breaks down and the mechanism by which it does. The more specific, the stronger the mark — specificity is the signal of genuine understanding. |
| Suppressing anomalous results | Experienced examiners notice when data disappears between the methodology description and the results. It raises authenticity concerns that affect the entire project evaluation. | Report every result, including anomalies. Address each one in discussion with a specific possible cause. Unexplained anomalies are often the most interesting engineering findings — do not waste them. |
| Discussion that re-describes results | If your discussion section just restates what was in the results section in slightly different language, it adds zero value. It proves data was collected but not understood. | A simple rule: never write a sentence in discussion that could have been written in results. If the sentence does not contain a mechanism, a comparison, an implication, or a limitation, it does not belong in discussion. |
| No connection to research objectives | The examiner cannot determine whether your project answered what it set out to answer. The most fundamental evaluative question remains open at the end of the chapter. | At the start of each major result block, state which research objective it addresses. At the end of your discussion, confirm explicitly whether each objective was met — and if not, explain why not. |
The before-and-after examples, the 4-move framework, and the "so what" test in this guide are built from analysis of examiner feedback across engineering project evaluations in India, the UK, Singapore, and Australia. The most consistent finding across all of them: students who describe data receive average marks, students who interpret it receive distinction. The framework makes that shift repeatable rather than accidental.
Section 08Frequently Asked Questions
Follow your institution format. If none is specified, one combined chapter is better — you interpret each result immediately after presenting it, which reads more naturally and is harder to fault than a separate discussion that re-describes data from memory.
Results state what happened — measured values and graphs, presented as facts. Discussion explains why it happened — the mechanism, the comparison to a standard or study, and what it means for engineering practice. Results are the evidence. Discussion is the reasoning that makes that evidence matter.
For undergraduate projects, 1,500 to 3,000 words excluding figures and tables is typical. The right length is whatever it takes to fully apply the 4-move framework to each of your key findings — padding with extra data descriptions adds words but removes value.
Describing what a graph shows instead of explaining what engineering behaviour caused that pattern. The examiner can see the graph. What they cannot see is whether you understand what produced it — and that is what the discussion section exists to demonstrate.
A specific limitation always makes a project stronger — it shows the examiner you understand the boundaries of your own evidence. "Results are valid for specimens cured at 27°C — higher field temperatures would alter hydration kinetics" is a sign of engineering maturity. "A larger sample size would improve results" impresses no one.
- How External Examiners Evaluate Project Results and Conclusions — Why Interpretation and Judgement Decide Final Grades 2026
- How to Write the Methodology Chapter for Engineering Projects — Complete Guide 2026
- How to Write a Literature Review for Engineering Projects — All Branches, UG to PhD 2026
- How to Write an Abstract for Engineering Projects — All Branches, UG to PhD 2026
- Why Civil Engineering Project Results Fail in Viva — Even When the Numbers Are Correct 2026
- How Examiners Evaluate Civil Engineering Projects — Hidden Criteria Students Never See
- The Complete Guide to Engineering Project Viva 2026
- 50 Most Common Engineering Project Viva Questions and How to Answer Them
- Feasibility and Measurement Framework for Engineering Projects
- How to Write an Engineering Project Report That Impresses Examiners — All Branches 2026