Writing up results honestly

A weather forecast that says "70% chance of rain" is more useful than one that says "it will rain". It tells you what the forecaster knows and how sure they are. A good project write-up does the same: it says what you found, how sure you are, and where your view stops.

Every project ends with results. For project D the write-up is part of the work and gets tutor feedback. For the others it is a short note beside your report, and it is the part you would show someone.

A five-part shape

1. Question: what you set out to do, in one sentence.
2. Method: what you did, in a few steps, so someone could repeat it.
3. Results: the numbers, with counts.
4. Limits: what your work cannot see or did not test.
5. Next: one or two things you would do with more time.

Short is fine. Half a page with all five parts beats three pages with no limits.

Numbers: always show the count

A percentage on its own hides how much evidence is behind it. "50% of sites had a tracker" could mean 1 of 2 or 500 of 1,000. Always give the count beside it: "12 of 40 practice sites (30%) set a third-party cookie".

Report the misses and false alarms too, not only the hits. If your blocker wrongly blocked 2 requests, say so, and say what broke. A reader trusts a result more when they can see you looked for your own mistakes.

Do not round in your own favour. 94.6% is "about 95%", not "nearly all". And say where each number came from: the hidden world, your practice world, or a single test run.

Limits: where your view stops

Some limits apply to almost every project in this course:

• It is a simulated world. The practice web follows the course's rules. The real web is messier, so a score here is evidence, not a promise about real sites.
• Your tool sees only what it looks at. A blocker that reads requests in the page cannot see data a site sends from its own server to a partner later. Say so.
• Browsers change. If you mention what a browser does, date it, as this course does with "(checked October 2026)". A fact that was true last year may not be now.
• Two things happening together is not proof that one causes the other. If sites with more trackers also had more ads, that does not show the trackers caused the ads.

Overclaiming, and how to fix it

Look at what the honest versions add: the denominator, where the number came from, and the edge of what was tested. None of them is longer than two sentences.

Keep the write-up itself private

A privacy project should practise what it preaches. Use only the practice world and synthetic data. Do not paste real people's data, real logs or real names into a write-up. Where the project compares a policy with what a site does, describe the practice site; do not name real companies.

A common misconception

"Listing limits makes my work look weak." It does the opposite. Every result has limits; the only question is whether you found them or a reader does. A clear limits section shows you understand your own work, which is the skill a reviewer is looking for.