Can negative evaluation lower my grade?

Not if it’s evidence-based. In IB Design Technology, examiners reward justified judgement, not positivity. If you admit a limitation and show the test data that revealed it, you’re doing exactly what evaluation is meant to do. The key is to avoid emotional language (“it was terrible”) and use analytical language (“requirement 3 was partly met because…”) instead. Also explain why the limitation occurred, because that shows design understanding rather than regret. Finally, add a realistic improvement linked to that limitation, so your evaluation reads like learning, not failure.

Do I need to evaluate every single requirement?

You should evaluate the requirements that genuinely define success. In IB Design Technology, shallow coverage of ten minor requirements often scores worse than deep coverage of five important ones. Pick requirements that are measurable and tied directly to the user need, because those create meaningful testing evidence. If a requirement can’t be tested, it will be hard to evaluate convincingly, so either rewrite it earlier or don’t centre your evaluation on it. Your goal is to make the marking easy: clear requirement, clear evidence, clear judgement. If you want to tighten your overall IA direction, Common IB Design Technology IA Mistakes and How to Fix Them is a strong reset.

Should I include improvements in the evaluation section?

Yes, but only improvements that are realistic and directly supported by evidence. In IB Design Technology, improvements are strongest when they sound like the next iteration you would genuinely build, not a brand-new concept. For each limitation, propose one improvement, then briefly explain how you would test it (so it stays grounded). Avoid introducing untested features that don’t relate to any requirement, because that can make your evaluation feel unfocused. If time is short, prioritise improvements that fix the requirements that were “partly met,” because that’s where marks are usually hiding. A good improvement reads like a controlled experiment: change one variable, then test again.