Concluding & Evaluating — Concluding and evaluating - IB Physics

SECTION 1

Historical Context & Motivation

Science has always depended on more than just collecting data — the real power lies in what you do with that data. Throughout history, the greatest breakthroughs came not from running experiments alone, but from carefully concluding and evaluating — interpreting results, identifying weaknesses, and refining methods. The scientific method itself evolved over centuries as thinkers realized that honest self-assessment of experimental work was just as important as the experiment itself.

1620 • Francis Bacon's Novum Organum

Bacon formalised the idea that scientific investigation must include systematic evaluation of evidence and identification of sources of error, laying the groundwork for modern experimental methodology.

1800s • Rise of Error Analysis

Mathematicians such as Gauss developed rigorous methods for quantifying uncertainty. Scientists began distinguishing between systematic and random errors, making evaluation of data quality a standard practice.

1935 • Karl Popper & Falsifiability

Popper argued that a scientific conclusion is only meaningful if it can, in principle, be proven wrong. This shifted focus towards evaluating whether conclusions are genuinely supported by evidence.

1960s–Present • Peer Review & Reproducibility

Modern science relies on independent evaluation of experimental claims through peer review. The reproducibility crisis of recent decades has reinforced why rigorous concluding and evaluating skills are essential.

In your IB Physics course, concluding and evaluating is the final stage of the internal assessment cycle. After you design, collect, and process data, you must ask: Does the data actually support my hypothesis? What went wrong, and how could I do better? Mastering this skill turns you from someone who merely follows instructions into someone who thinks like a real physicist.

SECTION 2

Core Principles of Concluding & Evaluating

Concluding and evaluating in IB Physics involves two interconnected skills. First, you draw a conclusion that directly addresses your research question, supported by the processed data. Second, you evaluate your procedure and results by identifying weaknesses, sources of error, and realistic improvements. These are not separate afterthoughts — they form the intellectual core of any investigation.

Stating a Valid Conclusion

A conclusion must directly answer your research question and be justified by your processed data (graphs, calculations). It should reference the relationship found — for example, whether the data shows a linear, proportional, or inverse relationship.

Comparing with Theory

You should compare your experimental results with accepted scientific values or theoretical predictions. Calculate the percentage error to quantify how close your result is to the literature value.

Identifying Sources of Error

Distinguish between random errors (scatter in data) and systematic errors (consistent offset from the true value). Explain specifically how each source affected your results — avoid vague statements like 'human error.'

Evaluating the Procedure

Assess whether the method was appropriate. Consider the range and quantity of data collected, the control of variables, and whether the equipment was precise enough for the measurements needed.

Suggesting Improvements

Propose realistic, specific modifications that address the weaknesses you identified. Each improvement should be linked to a specific source of error and explain how it would reduce uncertainty.

✦

KEY TAKEAWAY

Think of concluding and evaluating like reviewing a recipe you just tried. The conclusion is your verdict — 'The cake rose properly, so the baking powder ratio works.' The evaluation is your honest critique — 'The oven ran hot, making the edges crispy, so next time I should calibrate the temperature first.' In physics, your data is the cake, and you need both the verdict and the critique to improve.

SECTION 3

Visual Explanation — The Concluding & Evaluating Process

This flowchart shows the five sequential steps of concluding and evaluating. The first two steps (highlighted in violet and cyan) fall under concluding, while the final three steps (pink, amber, emerald) fall under evaluating. Each step builds on the previous one — you cannot meaningfully suggest improvements without first identifying errors.

Notice in the diagram above how concluding and evaluating are separated into two zones. The concluding zone requires you to interpret your processed data and connect it to accepted physics. The evaluating zone demands critical thinking — you must be honest about what went wrong and thoughtful about how to fix it. In the IB assessment, both zones carry marks, and many students lose points by rushing through the evaluation or offering only generic comments.

SECTION 4

Mathematical Framework — Percentage Error & Uncertainty

While concluding and evaluating is largely a qualitative skill, there is an important quantitative tool that anchors your conclusions: percentage error. This calculation tells you how far your experimental result deviated from the accepted or theoretical value. A small percentage error suggests your procedure was sound; a large one signals that significant errors affected your results.

PERCENTAGE ERROR

Percentage Error = |Experimental Value − Accepted Value| ÷ Accepted Value × 100%

The vertical bars indicate absolute value — the result is always positive. Experimental Value is what you measured; Accepted Value is the literature or theoretical value.

You should also consider the percentage uncertainty of your measurements. If your percentage error is within the percentage uncertainty, then the discrepancy is likely due to random error and your result is consistent with the accepted value. If the percentage error is significantly larger than your uncertainty, a systematic error is probably present.

PERCENTAGE UNCERTAINTY

Percentage Uncertainty = (Absolute Uncertainty ÷ Measured Value) × 100%

Absolute Uncertainty is typically ± half the smallest division of your instrument, or the range of repeated readings divided by two.

💡

Decision Rule

If % error ≤ % uncertainty, your result agrees with the accepted value within experimental limits — random error likely accounts for the difference. If % error > % uncertainty, a systematic error is probably present and you should identify its source.

SECTION 5

Detailed Breakdown — Random vs. Systematic Errors

A strong evaluation section requires you to distinguish clearly between the two main categories of experimental error. Understanding this distinction is essential because the type of error determines the type of improvement you should suggest.

The target analogy illustrates the difference between the two error types. Random errors produce scattered results around the true value (left target), while systematic errors shift all results consistently away from the true value (right target). The improvement strategy differs for each type.

Feature	Random Error	Systematic Error
Effect on data	Causes scatter above and below the true value	Shifts all readings in one direction (too high or too low)
Visible on graph	Data points scattered around best-fit line	Best-fit line has wrong gradient or y-intercept
Reduced by	Repeating measurements and calculating the mean	Recalibrating instruments, improving technique
Example	Slight variations in timing a pendulum with a stopwatch	A ruler with a worn-down zero mark, causing all length measurements to be 2 mm too short

Key differences between random and systematic errors

SECTION 6

Worked Example — Evaluating a Pendulum Experiment

Suppose you investigated the relationship between the length of a simple pendulum and its period. You measured g (acceleration due to gravity) from your data and obtained g = 10.2 m s⁻². The accepted value is 9.81 m s⁻². Your percentage uncertainty from error propagation is ± 3.0%. Let us work through the full concluding and evaluating process.

Concluding and Evaluating a Pendulum Experiment

1
Step 1 — State the Conclusion

From the graph of T² against L, the relationship between the period squared and the length of the pendulum is linear and proportional, passing through the origin. This is consistent with the theoretical relationship T² = (4π²/g) × L. The gradient of the best-fit line was used to calculate the experimental value of g.

Conclusion: T² is directly proportional to L, confirming the theoretical prediction.
2
Step 2 — Calculate Percentage Error

Percentage Error = |10.2 − 9.81| ÷ 9.81 × 100% = 0.39 ÷ 9.81 × 100%

Percentage Error ≈ 4.0%
3
Step 3 — Compare % Error with % Uncertainty

The percentage error (4.0%) is greater than the percentage uncertainty (3.0%). This means the discrepancy cannot be fully explained by random error alone. A systematic error is likely present in the procedure.

4.0% > 3.0% → systematic error suspected
4
Step 4 — Identify Specific Sources of Error

The experimental value of g was higher than expected. Since g = 4π²/(gradient), a higher g means the gradient was too small — that is, T² values were too low relative to L. This could be caused by measuring the pendulum length from the wrong point (e.g., from the top of the bob instead of the centre of mass), making effective lengths systematically too long. Another source is reaction time when starting and stopping the stopwatch, introducing random error into period measurements.

Systematic: incorrect length measurement point. Random: reaction time with stopwatch.
5
Step 5 — Suggest Realistic Improvements

To address the systematic error, measure the pendulum length from the pivot to the centre of mass of the bob, and use a set square to ensure the measurement is vertical. To reduce the random timing error, use a light gate and data logger to measure the period electronically, or time 20 complete oscillations instead of 10 to reduce the fractional uncertainty in the period measurement.

Improvement 1: Measure to centre of bob mass. Improvement 2: Use light gate for timing.

SECTION 7

Strong vs. Weak Evaluation Statements

One of the most common pitfalls in IB Physics is writing vague evaluation statements that don't demonstrate genuine understanding. The difference between a weak and a strong evaluation is specificity: you need to name the exact error, explain its direction of effect on your results, and propose a concrete fix. The table below shows real examples of this contrast.

Category	Weak Statement ✗	Strong Statement ✓
Source of error	"Human error affected results."	"Reaction time (≈ 0.2 s) when starting the stopwatch caused random uncertainty in period measurements, contributing ± 1% to each timing."
Direction of effect	"The results were inaccurate."	"The measured value of g was 4% too high, indicating that the gradient of the T²–L graph was systematically too low."
Improvement	"Use better equipment."	"Replace the manual stopwatch with a light gate connected to a data logger to eliminate reaction time and reduce timing uncertainty to ± 0.001 s."
Data range	"More data should be collected."	"Extend the range of pendulum lengths from 0.20–0.80 m to 0.10–1.20 m, and add two more data points to better constrain the gradient."

Comparing weak and strong evaluation statements for IB Physics

✦

KEY TAKEAWAY

Think of 'human error' like saying 'something went wrong with my car.' A mechanic would never accept that — they need to know if it's the brakes, the engine, or the tyres. Likewise, an IB examiner needs you to pinpoint the specific cause, explain its effect, and describe the exact fix. Specificity is what separates a passing evaluation from an excellent one.

SECTION 8

Connection to the IB Internal Assessment & Beyond

The concluding and evaluating skills you develop now directly feed into the IB Physics Internal Assessment (IA), which is worth 20% of your final grade. The IA rubric explicitly assesses your ability to state a conclusion with justification, evaluate your procedure, and suggest realistic improvements. Beyond the IB, these same skills are used by professional scientists when writing the discussion section of research papers.

Aspect	IB IA (High School)	University Research Paper
Conclusion	State the relationship found; compare with theoretical expectation	Discuss findings in context of existing literature and competing models
Error analysis	Percentage error and percentage uncertainty comparison	Statistical tests (χ², t-tests), confidence intervals, error budgets
Evaluation	Identify 2–3 specific weaknesses and their impact	Comprehensive methodology critique, control experiments, blind protocols
Improvements	Realistic, linked to identified weaknesses	Future work section; often informs follow-up experiments and grant proposals

How concluding and evaluating scales from IB to professional research

As you advance in science, the tools become more sophisticated — you'll encounter statistical significance testing and peer review — but the fundamental logic remains the same. Does the evidence support the claim? What could have gone wrong? How can we do better next time? Mastering these questions at the IB level gives you a solid foundation for any scientific career.

SECTION 9

Practice Problems

Question 1: CONCEPTUAL

A student writes in their evaluation: 'The experiment had human error which affected the results.' Explain why this is an inadequate evaluation statement, and rewrite it as a strong evaluation statement for an experiment measuring the speed of sound using a resonance tube.

Show answer

The statement is inadequate because 'human error' is vague and does not identify a specific source of uncertainty, its direction of effect, or how it could be improved. A strong version would be: 'The position of the resonance node was determined by ear, introducing a random uncertainty of approximately ± 0.5 cm in locating the loudest point. This uncertainty could be reduced by using a microphone connected to a data logger to detect the maximum sound intensity objectively, improving precision to ± 0.1 cm.' The key is specificity: name the error, quantify it, explain its effect, and propose a concrete improvement.

Question 2: BASIC CALCULATION

A student measures the specific heat capacity of aluminium and obtains c = 950 J kg⁻¹ K⁻¹. The accepted value is 897 J kg⁻¹ K⁻¹. Calculate the percentage error.

Show answer

Percentage Error = |950 − 897| ÷ 897 × 100% = 53 ÷ 897 × 100% ≈ 5.9%. This tells us the experimental value is about 5.9% higher than the accepted value. The fact that the result is too high suggests that not all the thermal energy from the heater was transferred to the aluminium — some may have been lost to the surroundings, meaning the student overestimated the energy absorbed by the metal. This systematic error could be reduced by using better insulation around the calorimeter.

Question 3: INTERMEDIATE

In a spring constant experiment, a student's percentage error is 2.5% and their percentage uncertainty from the gradient of the F vs. x graph is 4.0%. What can the student conclude about whether their result agrees with the accepted value? What type of error is most significant in this experiment?

Show answer

Since the percentage error (2.5%) is less than the percentage uncertainty (4.0%), the student's experimental value agrees with the accepted value within the limits of experimental uncertainty. The discrepancy can be explained by random error alone, and there is no strong evidence of systematic error. The dominant source of uncertainty is random error, likely from difficulty in reading the extension precisely or slight variations in how the force was applied. To improve, the student could use a digital caliper for more precise extension measurements and increase the number of data points to better define the gradient.

Question 4: APPLIED

You investigated how the angle of incidence affects the angle of refraction for light passing from air into a glass block, and used your data to determine the refractive index. Your graph of sin(θᵢ) vs. sin(θᵣ) gives a gradient of 1.42, but the accepted refractive index of the glass is 1.52. Your percentage uncertainty is ± 3%. Write a complete evaluation paragraph identifying the most likely source of this discrepancy and suggesting a specific improvement.

Show answer

The percentage error is |1.42 − 1.52| ÷ 1.52 × 100% ≈ 6.6%, which is larger than the 3% uncertainty, indicating a systematic error. The measured refractive index is too low, meaning the gradient of the sin(θᵢ) vs. sin(θᵣ) graph was too small. This suggests the refracted angles were systematically measured too large. The most likely cause is parallax error: when viewing the refracted ray through the glass block, the pins marking the emergent ray may not have been placed exactly along the true path. Additionally, if the glass block was not perfectly rectangular, the normal line drawn at the surface would be slightly off, skewing all angle measurements. To improve, the student should use a narrow laser beam instead of pins, which removes parallax entirely and produces a clearly defined ray path. The student should also verify the glass block's geometry with a protractor before beginning.

Question 5: CRITICAL THINKING

Two students both measure the acceleration due to gravity using the same pendulum apparatus. Student A obtains g = 9.60 m s⁻² with a percentage uncertainty of ± 2%. Student B obtains g = 10.10 m s⁻² with a percentage uncertainty of ± 5%. Which student's result is more accurate? Which is more precise? Could both results be considered 'valid' within their uncertainties? Justify your answers.

Show answer

Student A's result is more accurate because her percentage error is |9.60 − 9.81| ÷ 9.81 × 100% ≈ 2.1%, compared to Student B's percentage error of |10.10 − 9.81| ÷ 9.81 × 100% ≈ 3.0%. Student A is also more precise because her uncertainty is ± 2% compared to B's ± 5%. However, when we check validity: Student A's percentage error (2.1%) is slightly larger than her uncertainty (2.0%), suggesting a small systematic error may be present — her result barely falls outside the accepted range. Student B's percentage error (3.0%) is well within his uncertainty (5.0%), so his result is consistent with the accepted value within experimental limits. Paradoxically, B's less precise result is statistically valid, while A's more precise result narrowly fails the validity check. This illustrates an important lesson: high precision does not guarantee agreement with the true value, and a result with larger uncertainty can still be scientifically valid if the accepted value falls within the uncertainty range.

SUMMARY

Lesson Summary

Summary

Concluding means stating a clear answer to your research question, justified by your processed data (graphs, calculations), and comparing your result to accepted theoretical values using percentage error. Evaluating requires you to identify specific random and systematic errors, assess the quality of your procedure and data range, and propose realistic, specific improvements that are each linked to an identified weakness.

Remember the decision rule: if your percentage error is within your percentage uncertainty, random error accounts for the discrepancy. If the percentage error exceeds the uncertainty, a systematic error is likely present. Always be specific — avoid vague phrases like 'human error' and instead name the exact cause, explain its direction of effect, and describe a concrete fix. These skills are essential for the IB Internal Assessment and for any future scientific work.