How to Write the Perfect Conclusion & Evaluation for IB Physics (The Complete Guide)

Spread the love

How to Write the Perfect Conclusion & Evaluation for IB Physics

You have designed the experiment. You have collected the data. You have even wrestled with Excel to create a graph with error bars. Now, you just need to wrap it up.

Many students treat the Conclusion and Evaluation as a quick summary to rush through at the end of the night. This is a massive mistake.

In the IB Physics rubric, the Analysis and Evaluation criteria account for a huge chunk of your grade. A weak conclusion can drag a perfect experiment down to a mediocre score.

In this guide, we will break down the exact formula for writing a Level 7 Conclusion and Evaluation, including real examples of what to do (and what not to do).

Contents hide

1 Part 1: The Conclusion (The 3-Step Formula)

1.1 1. Directly Answer Your Research Question

1.2 2. Interpret the Gradient (Slope)

1.3 3. Compare to the Accepted Literature Value

1.4 Real Example: The Simple Pendulum

1.4.1 ❌ The “Level 4” Conclusion (Weak)

1.4.2 ✅ The “Level 7” Conclusion (Strong)

2 Part 2: The Evaluation (Limitations & Improvements)

2.1 Rule #1: Never Use the Phrase “Human Error”

2.2 The Two Types of Errors

2.3 The “Limitation + Improvement” Strategy

2.4 Final Tips for a 7/7 Evaluation

3 Summary Checklist

Part 1: The Conclusion (The 3-Step Formula)

Your conclusion needs to do three specific things. If you miss one, you lose marks.

1. Directly Answer Your Research Question

Don’t be vague. Look at your graph and state the relationship clearly.

Don’t say: “The data shows that length affects the period.”
Do say: “The graph of $T^2$ against $L$ is a straight line through the origin, indicating that the square of the period is directly proportional to the length ( $T^2 \propto L$ ).”

2. Interpret the Gradient (Slope)

If you linearized your data (which you should have!), the gradient of your line usually represents a physical constant. You must calculate this value and explain what it means.

Example: “The gradient of the line is $4.02 \, \text{s}^2\text{m}^{-1}$ . According to the formula $T^2 = (\frac{4\pi^2}{g})L$ , the gradient equals $\frac{4\pi^2}{g}$ .”

3. Compare to the Accepted Literature Value

This is your “sanity check.” Calculate the value of the constant (e.g., gravity, $g$ ) using your gradient, and compare it to the textbook value.

Percentage Error Formula:

$\text{\% Error} = \left| \frac{\text{Theoretical Value} - \text{Experimental Value}}{\text{Theoretical Value}} \right| \times 100$

If your error is low (e.g., under 5%), your experiment is accurate. If it is high, you must acknowledge it!

Real Example: The Simple Pendulum

Let’s look at how a student scoring a 4 writes a conclusion versus how a student scoring a 7 writes one.

❌ The “Level 4” Conclusion (Weak)

“In conclusion, my hypothesis was correct. As the length of the string got longer, the time period got longer too. The graph went up, which shows they are related. My value for gravity was $9.5 \, \text{m/s}^2$ , which is close to $9.81 \, \text{m/s}^2$ . Therefore, the experiment was a success.”

Why this fails: “Related” is too vague, there is no discussion of uncertainties, and “close to” is not scientific analysis.

✅ The “Level 7” Conclusion (Strong)

“The graph of $T^2$ against $L$ clearly shows a linear relationship with a positive gradient, intersecting the origin (within the uncertainty of the y-intercept). This confirms the theoretical relationship $T \propto \sqrt{L}$ .

From the gradient of the graph ( $4.15 \pm 0.20 \, \text{s}^2\text{m}^{-1}$ ), the experimental value for the acceleration due to gravity ( $g$ ) was calculated to be $9.51 \, \text{m/s}^2$ .

The accepted literature value for $g$ is $9.81 \, \text{m/s}^2$ . This gives a percentage error of 3.06%. Since the percentage error is smaller than the percentage uncertainty in my raw data (approx. 5%), the random errors alone can explain the discrepancy.”

Part 2: The Evaluation (Limitations & Improvements)

If the Conclusion is where you state what happened, the Evaluation is where you explain why it wasn’t perfect. This is where the difference between a 5 and a 7 usually happens.

Rule #1: Never Use the Phrase “Human Error”

If you write “human error” in your IA, you are essentially telling the examiner: “I was lazy and didn’t measure properly.”

Bad: “I might have counted the oscillations wrong.”
Good: “Reaction time introduced a random error into the timing measurements.” (This frames it as a limitation of the method, not the person.)

The Two Types of Errors

Random Errors (The “Scatter”): Unpredictable fluctuations causing points to scatter off the line.
Fix: Take more repeat readings and average them.
Systematic Errors (The “Shift”): Errors that shift all data in one direction (e.g., a zero-error on a ruler).
Fix: Recalibrate instruments or adjust the method.

The “Limitation + Improvement” Strategy

The best way to structure this section is to link every limitation directly to a realistic improvement.

Limitation (The Problem)	Improvement (The Fix)
Human Reaction Time: “Using a stopwatch introduced uncertainty.”	Use a Photogate: “Connect a light gate to a data logger to measure the period automatically.”
Parallax Error: “It was hard to read the ruler from an angle.”	Use a Mirror: “Place a mirror behind the object. When the reflection aligns with the object, the eye is perpendicular.”
Heat Loss: “The beaker lost heat to the room.”	Insulation: “Wrap the beaker in bubble wrap and use a lid to prevent convection.”

Final Tips for a 7/7 Evaluation

Be Specific: Don’t say “use better equipment.” Say “Use a Vernier caliper ( $\pm 0.05mm$ ) instead of a ruler.”
Be Realistic: Don’t suggest doing the experiment in a vacuum unless your school has one.
Analyze Outliers: If one point is way off, explain why specifically for that point.

Summary Checklist

Before you submit your IA, check these four boxes:

[ ] Did I state whether the graph is linear?
[ ] Did I calculate the gradient with units?
[ ] Did I calculate the % error vs. literature value?
[ ] Did I suggest a specific improvement for every error listed?