Logarithmic graphs
Studying this course? Keep your progress in one place.
Key facts
- Taking logs turns and into straight lines you can measure.
- Power law: plot against ; gradient , intercept .
- Exponential law: plot against ; gradient , intercept .
- The intercept is a logarithm: undo it with .
- Predictions outside the data (extrapolation) are less reliable.
| Plot | against | against |
|---|---|---|
| Gradient | ||
| Intercept |
- Plot:
- against
- Gradient:
- Intercept:
- Plot:
- against
- Gradient:
- Intercept:
Why straighten data
A straight line has a gradient and intercept you can measure, which a curve does not.
The two relationships in the specification are the power law and the exponential law . The graph you draw shows which model you are testing.
means , but any base works if you are consistent.
- Curve or take logs to get a line
Which graph is a straight line if ?
Power law
against is a straight line with gradient and intercept .
. Compare with where and . So .
If this graph is a straight line the data fit a power law.
Worked example
A graph of against is a straight line through and . Find .
- 1
Gradient: .
- 2
Intercept: .
- 3
Undo the log: .
A graph of against has gradient and intercept . What is ?
Exponential law
against is a straight line with gradient and intercept .
. The horizontal axis is itself, not . So and .
Worked example
A graph of against is a line through and . Find and in .
- 1
Gradient: , so .
- 2
Intercept: , so .
A graph of against has gradient and intercept . What are and ?
Estimating parameters
Decide which graph is straight, find the gradient and intercept, then undo the logarithm.
You may be given pairs, a table of values, or two points on a line of best fit.
Find the gradient from two points far apart, , then the intercept by substituting a point into .
- 1
Choose the straight graph
lg y vs lg x, or lg y vs x
- 2
Gradient
From two points far apart
- 3
Intercept
Substitute into Y = mX + c
- 4
Undo the log
Use 10 to the power
Worked example
A line of best fit for against passes through and . Find .
- 1
.
- 2
.
- 3
.
Why choose two points far apart on the line of best fit to find the gradient?
Using the model and its limits
Estimates are less reliable outside the range of the data.
Once the parameters are known, substitute into the original relationship to estimate , or into the line equation to find .
Extrapolation is unreliable, and exponential models cannot grow without limit in real situations.
Within the data
- Interpolation
- Reasonably reliable
Outside the data
- Extrapolation
- Unreliable: the pattern may not continue
Worked example
A model satisfies . Find when .
- 1
.
- 2
.
- 3
.
Why is a prediction far outside the range of the data unreliable?
Try an exam question
Data are modelled by . A graph of against is a straight line through and . Find and .
[4 marks]
- [1]Gradient , so .
- [1]Intercept: .
- [1].
- [1].
That's the notes covered.
Carry on to the next subtopic.
Try a question
Which equation links and ?
1 mark, marked against the mark scheme
Which equation links and ?
1 mark, marked against the mark scheme