GoalExplain how scientists estimate very large numbers by counting a sample and scaling up, and judge how reliable an estimate is.
New words
estimate · a careful guess based on measurementssample · a small part that is counted or measured to find out about the wholemean · the total of the values divided by how many there arescale up · to multiply the result for the sample to get the answer for the whole
Explanation
There are far too many galaxies and stars to count one by one, so scientists estimate. They point a telescope at a small patch of the sky, count the galaxies in it, and multiply by the number of patches that would cover the whole sky. This works only if the patches counted are typical, so scientists use several, spread around, and find the mean. The estimate is more reliable if more patches are counted, if they are chosen without bias, and if faint galaxies can be seen as well as bright ones. Some galaxies hide behind others and some are too faint, so the estimate is always uncertain and the true number may be higher. You can try the same method at home: count the words on a few lines of a page and scale up to the whole page.
count a small patch→find the mean of several patches→multiply by the number of patches→estimate for the whole sky
Worked examples
Square
Galaxies counted
1
11
2
14
3
12
4
15
5
13
6
12
mean
12.8
A model sky is divided into 60 equal squares. A student counts the galaxies in 6 squares: 11, 14, 12, 15, 13 and 12. Total = 77. Mean = 77 ÷ 6 ≈ 12.8 per square. Estimate for the whole sky = 12.8 × 60 ≈ 770 (exactly 770 with the unrounded mean).
Problem
Improvement
only busy squares counted
choose squares at random
few squares counted
count more squares
faint galaxies missed
use a more powerful telescope
Ways to improve an estimate: count more squares, choose the squares at random (not just the ones that look busy), count each square twice, and use a better telescope to see faint galaxies.
Square
Galaxies counted
1
2
2
3
3
2
4
41 (a cluster)
5
3
6
2
A limit of the method: it assumes that every square is like the squares counted. In this sky one square holds a cluster. With all six squares the mean is 53 ÷ 6 ≈ 8.8, which scales to about 530 for 60 squares. Without the cluster the mean is 12 ÷ 5 = 2.4, which scales to 144. So scientists give an estimate as a range, not as an exact value.
Class activity
Estimate by sampling. Take a newspaper or book page with an adult. Choose five lines by closing your eyes and pointing, and count the words in each. Calculate the mean. Count the lines on the page, or measure the height of the text and divide by the height of one line. Multiply to estimate the number of words on the page. Then count the words in a whole column exactly. How close was the estimate? Which lines would have biased the result?
Practice
1
A model sky has 40 squares. A student counts 8, 10, 9, 12 and 11 galaxies in five squares. Estimate the total number of galaxies.
Square
1
2
3
4
5
galaxies counted
8
10
9
12
11
Total = 50, mean = 10 per square, estimate = 10 × 40 = 400 galaxies.
2
Why is it better to choose the squares at random than to choose the ones with the most galaxies?
The busiest squares are not typical. Scaling up from them would give an estimate that is too high. Random squares represent the whole sky more fairly.
3
A patch of sky is found to hold 3000 galaxies. Give two reasons why multiplying this up may not give the true number for the whole sky.
The patch may not be typical of the whole sky, and the telescope may miss faint or hidden galaxies. Counting more patches would help.
4
Nodira says: ‘The stars in the Milky Way would fly apart, because nothing joins them.’ What is wrong with that idea?
Gravity holds the stars, gas and dust together in the galaxy. The galaxy has a very large mass, so the pull is strong.