Analyze and interpret
You've got a pile of data. Now for the exciting part: figuring out what it's actually telling you. This step has two jobs — analyze (find the patterns) and interpret (say what they mean for your question).
Raw data doesn't speak for itself. Analyzing means organizing it and pulling out the patterns — with averages and charts, or by finding recurring themes. Interpreting means stepping back and asking what those patterns actually mean for your question, honestly and without overclaiming. Analysis tells you what happened; interpretation tells you what it means.
Analyze, then interpret
Keep these separate in your head. First find out what the data shows; only then decide what it means.
Analyze
Organize the data and surface the patterns: averages, totals, percentages, charts, or recurring themes.
The question: “What does the data show?”
Interpret
Explain what those patterns mean for your research question — and be honest about how sure you can be.
The question: “So what — and how confident am I?”
Clean and organize before you analyze
Garbage in, garbage out. Spend a few minutes tidying the data so your analysis means something.
- Work on a copy. Keep your original raw data untouched — always.
- Hunt for errors. Typos, impossible values (a height of 400 cm), duplicate rows.
- Decide about missing data. Note what's missing and why; don't just quietly invent it.
- Get it into a workable shape. One row per observation, one column per thing you measured.
Ways to find the pattern
You don't need fancy statistics to do real analysis. Start with these — tap each to see how.
The basics go a long way: the average (mean or median), the range (highest to lowest), counts, and percentages. “Plants under red light grew an average of 6.2 cm vs. 4.1 cm under blue” is a real finding.
A good chart shows a pattern faster than a table ever could — and often reveals one you'd have missed. Graph your data early; it's the quickest way to actually see what happened.
If you compared two groups, look at the difference — then ask the key question: is it big enough to be real, or could it just be chance? Small samples and big overlap mean “be cautious.” (Comparison tests can answer this more formally once you're ready.)
For interviews or open-ended answers, read through everything and tag recurring ideas. Then count how often each theme shows up. “Seven of ten students mentioned stress” turns words into evidence.
Pick the right chart
The chart type depends on what you're showing. Match it and your finding jumps off the page.
Bar chart
Comparing categories or groups — like average growth under red vs. blue light.
Line chart
Change over time — like plant height measured each day for two weeks.
Scatter plot
The relationship between two numbers — like screen time vs. hours of sleep.
Pie chart
Parts of a whole — like the share of students who chose each option. Use sparingly.
Is your interpretation honest?
Analysis is only half the job — interpreting fairly is what makes it trustworthy. Run your conclusion through these six checks.
The trap almost everyone falls into
If you remember one thing from this step, make it this.
Correlation is not causation
Two things moving together doesn't mean one causes the other — something else might drive both, or it could be coincidence. Only a controlled experiment lets you claim cause. From a survey or observation, say things are “linked” or “associated,” not that one “causes” the other. This single habit will make your work far more credible.
Traps beginners fall into
- Cherry-picking the numbers that fit your hypothesis and ignoring the rest.
- Claiming cause from a correlation.
- Overclaiming from a small sample — a hint is not proof.
- Burying results that contradict what you expected.
- Treating “no clear effect” as failure — a null result is still a real, publishable finding.
- Using fancy statistics you can't explain. A simple analysis you understand beats an impressive one you don't.
Pro tip · the “so what?” test
For every number or chart, ask “so what does this mean for my question?” If you can't answer, it's just decoration. And don't fear a result that goes against your hypothesis — “I expected X but found Y” is real science, and often the most interesting story you can tell.