Reading a bar graph: what length is standing for

0:00You have seen these before. Bars, side by side, some tall and some short. They are in newspapers, on weather apps, in the back of every report you get handed. So this is not a new drawing to learn. It is one you already half read — and the job is to find out what you have been trusting. Because a bar graph asks you to believe something, and it is worth knowing what.

0:29Start with data we have already drawn once. Eight classes, and the number of students absent from each. Three, five, four, two, none, one, five, seven. Last time we drew that as tiles. One tile per student, counted out. Now here it is again as bars. The difference is not decoration — it is what your eye is asked to do. On the pictograph you counted objects. On the bar graph you are measuring a length.

1:00That swap is the whole invention. It sounds like nothing. It changes what the drawing can carry. And a length means nothing on its own, for exactly the reason a tile did not. So there is a sentence printed above the graph, doing the same job the key did. One unit of length equals one student. Read it carefully: it has two halves, and they live in different places. One unit of length — that is on the paper, measurable with a ruler.

1:35One student — that is out in the world, in a classroom. The sentence is a bridge between them. Cover it up and the bars are just rectangles. Now, the graph has lines running across it — and here the book stops to make a point of something that looks like housekeeping. Those lines have to be equally spaced. Zero to one is the same distance as one to two, and as two to three, all the way up.

2:05And you should ask why that needs saying at all. Here is why. Suppose the gaps grew as you went up. A bar reaching the fourth line would look four times one reaching the first — when the numbers might be nothing like that. Your eye compares lengths. It cannot help it. So the drawing has to be built so that comparing lengths gives the right answer. Equal steps are not neatness. They are the promise that makes the picture honest.

2:38Right. Reading a value off — there is a method worth doing slowly at first. How many were absent from Class Two? Find Class Two along the bottom. Go up its bar to the top. Then go across to the scale and read the number. Five. Which class had the most absent? You measure nothing for that one. You look for the tallest bar. Class Eight. Notice how much faster the second question was.

3:07Reading one value takes work; finding the biggest takes a glance. A bar graph is built for the second kind. Now Class Five, where something looks wrong. There is no bar there at all. Just a gap where one should be. And the temptation is to read that as missing — as though somebody forgot that class, or the data never came in. It is not missing. Nobody was absent from Class Five. The frequency is zero, and a bar of length zero is exactly what zero looks like.

3:45This is the same point the empty pictograph row made, and it is worth making twice — a blank really is ambiguous on a page. Zero is a measurement. It is an answer. Draw it as nothing, and label the class anyway. Two more rules, and they are about the bars rather than the scale. Every bar has the same width. And the gaps between them are equal too. Why insist? Because if widths varied, width would become a second signal — and nobody ever told the reader what it means.

4:22So the width is held constant on purpose, to carry no information. Only length is allowed to speak. One thing is free, though: bars may run upwards or across the page. That is a design choice, and it changes nothing. Now change the scale — because this is where a bar graph pulls away from a pictograph completely. Vehicles crossing a busy junction, counted for each of the six hours from six in the morning until midday.

4:56And these bars run across instead of up. The declaration this time: one unit of length equals a hundred vehicles. So the scale beside them is marked a hundred, two hundred, three hundred, and on up to twelve hundred. Ask what a pictograph would do here. At one symbol per vehicle it needs thousands. Even at a hundred a symbol you are counting twelve tiny pictures and squinting at halves.

5:26The bar graph does not care. Twelve hundred is just a longer line. Now read them, and notice the readings are not all the same kind. The busiest hour is seven to eight, and its bar ends exactly on the twelve hundred mark. That reading is exact. Twelve hundred vehicles. The quietest is six to seven, and its bar stops somewhere between the hundred mark and the two hundred. So what do you report? About one hundred and fifty.

6:00And that word about is not weakness or laziness. It is precision about your own precision. The bar genuinely does not tell you whether it is a hundred and forty or a hundred and sixty. Saying a hundred and fifty flatly claims something the drawing does not support. So learn the hedge. When a bar ends between two marks, say about, and mean it. Here is a question that needs two bars instead of one. How many vehicles crossed between eight in the morning and ten?

6:36Two hours, so two bars. Eight to nine, and nine to ten. Eight to nine reads a thousand. Nine to ten reads eight hundred. Add them. Eighteen hundred vehicles. And picture what you just did — you laid one bar on the end of the other. Adding quantities is joining lengths, and the graph makes that literal. Now the whole morning, all six hours. And here is something the book does not tell you.

7:05It states five of the six values in words. The ten to eleven bar it simply draws, at about seven hundred, and never mentions. So to total the morning you have to read a bar nobody read for you. One fifty, twelve hundred, a thousand, eight hundred, seven hundred, six hundred. Four thousand four hundred and fifty vehicles. One last graph, and it shows what the scale is really for.

7:37A country's population, once every ten years, from nineteen fifty-one to two thousand and one. Three hundred and sixty million, four hundred and forty, five hundred and forty, six hundred and eighty, eight hundred and forty, and one thousand and twenty million. You cannot count that. Not with tiles, not with anything. So the scale does the work: one unit of length stands for a hundred million people. Which makes the first bar three and a bit units long, and the last one just over ten.

8:13Same drawing as the absentees. Same rules. All that changed is one sentence above the graph. And look what the picture hands you for free. Over those fifty years it grew by six hundred and sixty million — nearly three times over. Better than that: work out each decade's growth. Eighty million, then a hundred, then a hundred and forty, then a hundred and sixty, then a hundred and eighty.

8:43Every decade added more than the decade before. That is not one fact, it is a pattern — and you got it by subtracting neighbouring bars. Which brings us to the thing worth taking away. Go back to the traffic graph. Six to seven is nearly empty. Seven to eight is the busiest hour of the morning. And after that, every hour is quieter than the last. You have now read all three of those off the picture. And not one of them is interesting yet.

9:19The interesting question is why. Why is seven to eight the peak? School and work starting, presumably — but that is a guess you would have to check. Why does it fall away steadily? Perhaps everyone who needed to travel already has. The graph cannot answer either. What it does is make the questions unavoidable — a shape that obvious demands a reason. So the reading is not the end of the work. It is the beginning.

9:50Next time, we stop reading other people's bar graphs and draw one ourselves — which means choosing the scale, and that turns out to be the whole difficulty.