Opening your dashboard…

← Back to the 6 steps

5 Step 5 of 6

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.

Two jobs, in order

Analyze, then interpret

Keep these separate in your head. First find out what the data shows; only then decide what it means.

First

Analyze

Organize the data and surface the patterns: averages, totals, percentages, charts, or recurring themes.

The question: “What does the data show?”

Then

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?”

First things first

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.
Make sense of it

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.

Show it well

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.

Your turn

Is your interpretation honest?

Analysis is only half the job — interpreting fairly is what makes it trustworthy. Run your conclusion through these six checks.

HonestYou circle back and actually answer what you set out to ask.
Off trackYou report neat facts that never address your question.
Honest“Red light was linked to slightly more growth in my sample.”
Overclaim“Red light makes all plants grow much faster.”
Honest“Screen time and poor sleep were related” (unless you ran an experiment).
Leap“Screens cause bad sleep” from a survey alone.
Honest“The warmer windowsill could also explain the growth.”
Tunnel visionAssuming your first idea is the only possible cause.
Honest“Only 12 plants, one location, two weeks — a small first look.”
PretendingPresenting a tiny study as the final word.
Honest“My hypothesis wasn't supported” — reported plainly.
HidingQuietly dropping the findings that didn't fit your guess.
Your interpretation passes 0 of 6 checks
The big one

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.

Watch out for

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.

Do this now: summarize your data, make a chart that shows the pattern, then write your finding in one or two plain sentences tied back to your question — with your limitations named. If your claim says exactly what the data supports and no more, you're ready for Step 6.