Understanding the IB Chemistry IA
The internal assessment is worth twenty percent of your final grade and it does not get graded by the teacher alone. The IB sends a sample to external examiners, so the work has to survive scrutiny from someone who has read thousands of these over the years. I watched students waste weeks on investigations that looked impressive but scored poorly because they lacked clarity in the research question. That is usually the first place things go wrong. You need a question that is specific enough to answer with data, but not so narrow that you cannot do meaningful error analysis.
What Ib Chemistry Internal Assessment Examples Actually Look Like
Top-scoring IAs share a pattern, though it is not obvious if you have never seen one. They do not read like lab reports from high school. They read like mini research papers with a clear thread from question through conclusion. A typical high-scoring IA covers something like the effect of temperature on the rate of reaction between sodium thiosulfate and hydrochloric acid, or the concentration of vitamin C in different fruit juices using redox titration. The topic itself matters less than how the investigation is structured. I have seen students score Level 6 with a relatively simple titration because their evaluation was sharp and their methodology was well-justified. I have also seen complicated calorimetry experiments score Level 3 because the student could not identify why their results were inconsistent. Here is what the structure generally looks like in practice:
Personal Engagement section — this does not mean writing a personal story about why you like chemistry. It means showing that you thought about the question yourself. Maybe you modified the procedure because the standard method had a problem you noticed. Maybe you sourced materials differently. That counts. Exploration section — this is where most students lose marks. Your research question must be clearly stated. The background theory needs to be relevant, not copied from a textbook. Your method needs to include enough detail that someone else could repeat it, including control variables, range of values, and number of trials. Analysis section — raw data goes into tables with units and uncertainty. Processed data includes averages, standard deviations if appropriate, and any graphical analysis. Error bars on graphs are not optional if you want full marks. A graph without error bars is a missed opportunity.
Get the Full Details
Evaluation section — this is where the grade is won or lost. You need to comment on the quality of your data, identify specific sources of error, and suggest realistic improvements. Generic suggestions like "use more precise equipment" will not score well. "The thermometer had an uncertainty of ±0.5 degrees Celsius, which contributed approximately X percent error to the enthalpy calculation" is what examiners look for.
Common Pitfalls That Sink IAs
The biggest mistake I see is students treating the IA as a demonstration of technique rather than an investigation. You can perform a perfect titration and still get a mediocre score if you did not frame it around a meaningful question and did not critically evaluate the results. Another issue is the data. Many students collect six data points and call it a dataset. That is not enough for any meaningful trend analysis. You need at least five to seven values across a reasonable range, with repeats at each point. Three trials at each concentration is the minimum I would recommend before you start looking at patterns. Uncertainty handling is another weak spot. Students often calculate percentage uncertainty for individual measurements but then fail to propagate those uncertainties through to their final results. If your final enthalpy change is 57.3 kJ/mol, you need to show how the uncertainty in mass, temperature change, and specific heat capacity combines to give you an uncertainty in that value. Without that, your conclusion has no basis for being taken seriously.
I ran into a specific problem once with a student doing a kinetics experiment measuring the decomposition of hydrogen peroxide using manganese dioxide as a catalyst. They collected volume of oxygen data every ten seconds for three minutes. The data was noisy. Their R-squared value on the best-fit line was 0.87, which looked bad. What they did not realize was that the gas syringe they were using had a friction issue that caused intermittent sticking, creating artificial jumps in the data. Instead of accepting the low R-squared and moving on, they spent an hour testing the syringe at different angles and found that tilting it slightly reduced the friction. They adjusted their method, collected new data, and got an R-squared of 0.96. That kind of hands-on troubleshooting is exactly what earns marks in personal engagement and evaluation.
How to Approach Your IA Step by Step
Start with the research question. If you cannot state it in one clear sentence, you do not have a question yet. "How does temperature affect reaction rate?" is too vague. "How does temperature between 25 and 55 degrees Celsius in 10-degree increments affect the initial rate of reaction between 0.5 M sodium thiosulfate and 1.0 M hydrochloric acid, measured by the disappearance of a marked cross?" is something you can actually investigate. Next, do a literature search. Not to copy sources, but to find what other people have done with similar questions. This helps you justify your method choices and gives you equations and constants you can reference. Examiners expect you to cite relevant sources, not just state facts. Then plan your method with enough detail. I usually tell students to write their procedure in past tense as if they have already done the experiment. This forces them to think through every step beforehand. If you find yourself writing "then add more acid if needed," you need to specify exactly how much and how concentrated that acid is.
Run a trial before collecting your real data. A trial takes twenty minutes and can save you three weeks of rewriting. I did this with an iodine clock experiment where the timing was off because the solution concentrations were slightly different than planned. The trial revealed that I needed to prepare a dilute stock solution to get the reaction time into a measurable range. Without that trial, I would have wasted hours on data that was essentially useless. When you analyze, use proper graphing. Linearize your data if there is a theoretical relationship that allows it. If you are studying reaction order, plotting log rate against log concentration gives you a straight line whose slope is the order with respect to that reactant. That is far more informative than a curved graph and it gives you a quantitative answer directly from the gradient. For uncertainty, calculate absolute uncertainty for each measurement, then propagate. The rule is simple: for addition and subtraction, add absolute uncertainties. For multiplication and division, add percentage uncertainties. When you raise a value to a power, multiply the percentage uncertainty by that power. Apply these consistently and your error analysis will be defensible.
Where Students Lose Unnecessary Marks
Word count is a real constraint. The maximum is 6,500 words. Many students blow past this with excessive background theory or detailed descriptions of equipment that everyone already knows how to use. Be concise. The examiner does not need a paragraph explaining what a burette is. They need to know why you chose a 50 mL burette over a 25 mL one and how that choice affected your uncertainty. Referencing is another area where students are sloppy. Use a consistent format throughout. The IB does not specify a particular style, but you need to cite every source you use. If you copied a value or an equation from somewhere, cite it. If you used a method adapted from a textbook, cite it. Missing references are an easy way to lose marks in the personal engagement criterion. Conclusion quality matters more than students think. Your conclusion should directly answer the research question using your data. Do not introduce new information in the conclusion. If you calculated an activation energy from your Arrhenius plot, state it clearly with its uncertainty and compare it to the accepted value. Discuss whether the difference is statistically significant given your error margins.
A Note on What Works and What Does Not
Some topics are easier to execute well than others. Titration-based investigations tend to be cleaner because the endpoints are sharp and the calculations are straightforward. Kinetics experiments are more forgiving of method variations because you are looking at trends rather than a single value. Calorimetry is the riskiest topic I would recommend avoiding unless you have good insulation and careful temperature measurement. Heat loss is nearly impossible to eliminate completely, and it is hard to argue convincingly that your results are reliable when you know there is a significant systematic error you cannot fix. If you are struggling with your IA topic, talk to your teacher early. Not at the last minute when you already have data and realize the analysis does not work. Early feedback on your research question and method can prevent you from going down a dead end. The IA is not designed to be easy. It is designed to show that you can think like a scientist, which means asking a good question, collecting decent data, analyzing it honestly, and being clear about what you do not know. That is it. Nothing flashy required.
