GE04-04 Geography Watch
Scatter graphs: relationships, best-fit lines and predictions
In this lesson
Scatter graphs: how to plot points on a scatter graph, describe the relationship between two variables - positive, negative or none, strong or weak. AQA 8035, Edexcel 1GA0, Edexcel 1GB0, Eduqas C111, Eduqas C112, OCR J383 and OCR J384 GCSE. Graphs 4 of 5. More at https://scholafly.com
What it covers
- 0:00 Intro: scatter graphs: best-fit lines
- 0:48 What a scatter graph
- 1:43 Describing a relationship
- 5:20 The line of best fit
- 8:18 Scatter graphs: best-fit lines: interpolation
- 10:24 Perchley's population, invented
About this video
Scatter graphs: relationships, best-fit lines and predictions | GCSE Geography | GE04-04
Scatter graphs: how to plot points on a scatter graph, describe the relationship between two variables - positive, negative or none, strong or weak.
AQA 8035, Edexcel 1GA0, Edexcel 1GB0, Eduqas C111, Eduqas C112, OCR J383 and OCR J384 GCSE. Graphs 4 of 5.
More at https://scholafly.com
Watch first: GE04-01 - https://youtu.be/F3xjsSvKGSM
Video code: GE04-04 (search "ScholaFly GE04-04")
#GCSEGeography #Graphs #ScatterGraphsBestFitLines
For teachers
This GCSE Geography lesson teaches scatter graphs: relationships, best-fit lines and predictions. By the end, students should be able to plot points on a scatter graph, describe the relationship between two variables - positive, negative or none, strong or weak - with named examples, data and anomalies, draw a best-fit line (straight, through the middle of the points, about as many above as below, not forced through the origin), and use it to interpolate within the data and, only when asked, extrapolate beyond it, saying that the extension is less certain. It works through three worked examples and the mistakes examiners report.
Exam board specification references:
AQA 8035
- 3.4.2
- 3.4.4
Edexcel 1GA0
- MS:Statistical:3
Edexcel 1GB0
- MS-Stat.3
Eduqas C111
- AppA:2.3
- AppA:4.1
Eduqas C112
- AppA:2.3
- AppA:4.1
OCR J383
- 3.3(10)
- 3.3(11)
- 3.3(8)
- 3.3(9)
- 3[graphs]
- 5d[statistical]-b3
OCR J384
- 5d[Statistical skills][3]
- 9.3.10
- 9.3.11
- 9.3.8
- 9.3.9
- 9[Graphs and charts to be studied][4]
Read the transcript
How far can you trust a line drawn through a cloud of dots? Plot every country by how rich its people are against how long they live, and the dots lean together. On the whole, richer countries have longer lives. But a few dots sit far from the crowd, such as a poorer country whose people live as long as those in much richer ones. A line through the crowd gives the rule, and the dots off it are the ones that ask why. It's tempting to extend that line as far as you like and read off any answer you want. How far it holds depends on how the line is drawn, and where you read it.
A scatter graph shows whether two things are related, with one point for each case, such as a country or a river site. The thing thought to cause the change goes along the bottom axis. The thing that responds goes up the side. Rainfall and river flow: which one goes along the bottom axis? Rainfall. More rain is what drives the flow higher, so river flow belongs on the vertical axis. To plot a point, go along to its value on the bottom axis, then up to its value on the side. Count the small squares on both axes, the same scale check as on any graph, and mark the point with a small neat cross.
A relationship has a direction and a strength. Positive means that as one thing goes up, the other goes up too. Negative means that as one goes up, the other goes down. None means the points show no pattern. Strong means the points lie close to a line. Weak means they're scattered loosely around one. In the fictional land of Owlobia, eight districts are named after their main towns, with invented figures. A scatter graph plots the percentage of adults who can read and write against infant deaths, meaning babies who die before their first birthday, for every thousand born. Owlton has ninety-five per cent who can read and write, and five infant deaths per thousand. Hootley has ninety per cent and twelve. Perchley has eighty and twenty-five. Talonford has seventy and forty. Featherby has sixty per cent and fifty-five deaths per thousand. Marksby has fifty and seventy. Gradeley has forty and eighty-five. And Penbury has forty-five and twenty. As reading and writing rise, what do infant deaths do, and what's that relationship? They fall, so the relationship is negative. Most points sit close to a line, which makes it strong. A full description has three steps. First, the relationship, with its direction and strength. Second, evidence: named cases with their values. Third, the anomaly, named with its values. For the second step, two students write different sentences. A says: as reading and writing fall, infant deaths rise. B says: in Gradeley, where forty per cent can read and write, there are eighty-five infant deaths per thousand, while in Owlton, at ninety-five per cent, there are only five. Which sentence adds a new point, A or B? B. A is the same relationship turned round, so it says nothing new. B gives named places with their values, which is evidence. One examiner's report notes students who repeated the same point in reverse, giving the opposite, and it records that this earned no credit. So make your second point a new one: named places with their values. Now, which district breaks the pattern? Penbury breaks the pattern. Only forty-five per cent can read and write, yet it has just twenty infant deaths per thousand, far below the line. One examiner's report, on a graph of income against life expectancy, records that only twelve point four per cent of students scored full marks. Many named the relationship and added nothing else. The habit that fixes it is all three steps, every time: the relationship, named evidence with values, and the anomaly.
A line of best fit is one straight ruled line that follows the trend of the points. Now, before the rules: must the line start at the origin, where both axes are nought? No. Your data decide where it runs, and it begins at nought only if they lead it there. It goes through the middle of the points, with about as many above it as below. It doesn't have to pass through any point. Draw it across the spread of the data, from about the first point to the last. Students measured the width of the River Tawny at eight sites, at different distances from its source, with invented results. At one kilometre it was two metres wide. At three kilometres, three and a half metres. At five, five metres. At seven, five and a half. At nine kilometres it was eight metres. At eleven, eight and a half. At thirteen, ten metres. And at fifteen kilometres, eleven metres wide. Three lines are drawn on the river graph. Line A starts at the origin and runs up through the points. Line B joins each point to the next, dot to dot. Line C runs straight through the middle, from about two metres at one kilometre to about eleven metres at fifteen, with four points above it and four below. Which one is the line of best fit: A, B or C? C. It is the only one that follows the trend evenly, sharing the crosses out on both sides. B is a zigzag, and A is pinned at nought. Now, why does pinning line A to the origin spoil it here? Pinned at nought, it tilts away from the data. The early crosses all end up over it and the later ones under it, so it stops sharing them out evenly. One examiner's report calls a line drawn through the origin a common error, and adds that a line of best fit does not necessarily need to pass through it. Another report, on a different paper, records that about eighty per cent of students failed to secure the mark for drawing the best fit line correctly. A third report, on another paper again, says most students were able to draw a line of best fit. Before you commit to yours, count the points above it and below it.
Reading a value off your line, inside the range of your data, is called interpolation. Start on the axis the question gives you. Here that's distance, along the bottom. Go up from ten kilometres to your line, then across to the width axis on the side. What width does your line give at ten kilometres from the source? About eight metres. Ten kilometres sits inside the measured sites, so this reading is interpolation, and it's fairly reliable. Read it off the line you drew, never off the nearest point. Extrapolation means extending the line beyond the data, to predict a value nobody measured. Draw the extension dashed, and only when the question asks for a prediction beyond the data. Now, extend the dashed line to twenty kilometres. Roughly what width does it give? About fourteen to fifteen metres at twenty kilometres. Now, why is that twenty kilometre figure less reliable than the one at ten? It is less reliable because it lies beyond the last measured site. The river could change further downstream, for example where another river joins it, and the line can't know that. The further a line goes beyond its data, the less certain the prediction, because the relationship itself may change. One more limit, in a line: a relationship on a scatter graph doesn't prove that one thing causes the other. The fuller limits belong to the video on spotting weaknesses in how data is presented.
Some courses also ask you to extend a trend on an ordinary line graph. Perchley's population, in invented figures, was forty thousand in two thousand, forty-three thousand in two thousand and five, forty-six thousand in twenty ten, forty-nine thousand in twenty fifteen, and fifty-two thousand in twenty twenty. That's a rise of three thousand every five years, a straight-line trend. Continued with a dashed ruled line, it gives fifty-five thousand in twenty twenty-five, fifty-eight thousand in twenty thirty, and sixty-one thousand in twenty thirty-five. Now, carry the dashed line on. What does it estimate for twenty forty? About sixty-four thousand people, following the same steady climb. Now, what has to stay true for that estimate to hold? The trend has to carry on unchanged. New housing, new jobs, or people moving in or out could all bend it, so call the figure a projection.
To finish, the questions turn round to you.
On a scatter graph, which axis takes the cause, and which the response?
The cause goes along the bottom, and the response goes up the side.
Now, as height above sea level rises, temperature falls. What relationship is that?
A negative relationship: a rise in one comes with a fall in the other.
Now, what are the three steps of a full relationship answer?
The relationship with its direction and strength, then named evidence with values, then the anomaly.
Now, what is one thing a line of best fit doesn't have to do?
It doesn't have to pass through the origin, or through any point. It follows the middle of the trend, balanced above and below.
Now back to the cloud of countries: how far can you trust the line?
Inside the spread of the dots, fairly well, which is interpolation. Beyond them, less and less, which is extrapolation, because the relationship may change. And the dots far from the line are the anomalies to name.
If you can draw a balanced line and say how far to trust it, a thumbs-up says you could teach this one. If not, say the best-fit rules out loud once and they stick, then come back to it.
Next in the chapter is Spotting weaknesses in how data is presented. Hopefully that's next in your feed, but if not, search for ScholaFly G E oh four, oh five.
Every other video is on scholafly.com too, easy to search by video code or by keyword. Whatever you're revising, go there and have a look, or, watch the next video.
Related terms
For: AQA GCSE 8035, Edexcel GCSE 1GA0, Edexcel GCSE 1GB0, Eduqas GCSE C111, Eduqas GCSE C112, OCR GCSE J383, OCR GCSE J384
On the specification
| Board | Spec | Statement |
|---|---|---|
| AQA GCSE 8035 | 3.4.2 | |
| AQA GCSE 8035 | 3.4.4 | |
| Edexcel GCSE 1GA0 | MS:Statistical:3 | |
| Edexcel GCSE 1GB0 | MS-Stat.3 | |
| Eduqas GCSE C111 | AppA:2.3 | |
| Eduqas GCSE C111 | AppA:4.1 | |
| Eduqas GCSE C112 | AppA:2.3 | |
| Eduqas GCSE C112 | AppA:4.1 | |
| OCR GCSE J383 | 3.3(10) | |
| OCR GCSE J383 | 3.3(11) | |
| OCR GCSE J383 | 3.3(8) | |
| OCR GCSE J383 | 3.3(9) | |
| OCR GCSE J383 | 3[graphs] | |
| OCR GCSE J383 | 5d[statistical]-b3 | |
| OCR GCSE J384 | 5d[Statistical skills][3] | |
| OCR GCSE J384 | 9.3.10 | |
| OCR GCSE J384 | 9.3.11 | |
| OCR GCSE J384 | 9.3.8 | |
| OCR GCSE J384 | 9.3.9 | |
| OCR GCSE J384 | 9[Graphs and charts to be studied][4] |
For teachers
This GCSE Geography lesson teaches scatter graphs: relationships, best-fit lines and predictions. By the end, students should be able to plot points on a scatter graph, describe the relationship between two variables - positive, negative or none, strong or weak - with named examples, data and anomalies, draw a best-fit line (straight, through the middle of the points, about as many above as below, not forced through the origin), and use it to interpolate within the data and, only when asked, extrapolate beyond it, saying that the extension is less certain. It works through three worked examples and the mistakes examiners report.