Topic 1.6 · teacher page · AI, SL and HL

Small errors do not stay small

Why a 4% error on each side becomes 12% on the area, and the bound notation examiners insist on.

The one thing to do with the animation

Ask whether rounding two measurements to the nearest unit is a big deal.

They will say no, and for a single measurement they are right: 4.17% at worst. Then reveal the area, which can be anywhere from 74.75 to 93.75 against a nominal 84.

Multiply quantities and the percentage errors roughly add. That is the one sentence to leave on the board, and it is the whole reason this sub-topic exists rather than being a rounding exercise.

The inequality signs are not decoration

4.05 ≤ x < 4.15. Closed below, open above, because 4.15 would round up to 4.2. Students write both as ≤ and it is marked wrong.

And the pairing flips. Greatest area is upper times upper, but greatest difference is upper minus lower, and greatest quotient is upper over lower. Make them think about which bound makes the answer big rather than copying a pattern.

The answers

1. Lower bound of 350 to 2 sfThe unit is 10, so 345.
2. Percentage error of 3.14 for π0.0507%.
3. Greatest area12.5 × 7.5 = 93.75.

Where the marks go

1 markBounds with the correct inequality signs.

1 markThe right pairing of bounds for the quantity asked for.

1 markDividing by the exact value in a percentage error, and rounding only at the end.

What each wrong answer tells you

They giveWhat it means
349.5 (Q1)Treated 350 as exact to the nearest unit. To 2 sf the rounding unit is 10.
355 (Q1)Gave the upper bound.
0.0405 (Q2)That is the error for 22/7, which is the better approximation.
0.00159 (Q2)Has the fraction, has not multiplied by 100.
84 (Q3)Gave the nominal area rather than the greatest possible.
84.5 (Q3)Added half a unit to the area instead of to each side first.

Other things they will say

"Why not just round at the end?" You should, and that is the point. Rounding mid-calculation is a second source of error on top of the measurement error.

"Is 22/7 better than 3.14?" Yes, 0.0402% against 0.0507%, which surprises people. Worth doing because it shows a percentage error is a real comparison and not a formality.

"How many significant figures should I give?" Match the data. Three is the default in these courses unless the question says otherwise, and an answer to seven figures from data given to two is its own error.

A possible order

 What is happening
1Is rounding a big deal? Reveal the area range.
2Bounds, with the inequality signs laboured.
3Percentage error, including the pi comparison.
4Compound quantities: area, then a difference, then a quotient.
5Choosing a sensible accuracy for an answer.

Two things not to say

Do not write both bounds with ≤. They will copy it and lose the mark.

Do not say percentage errors "add" as if it were exact. They roughly add for products; the exact answer comes from the bounds.

Questions to set

Three tiers, ramping the way practice should: the method on its own, then the method inside something real, then a challenge. Set the tier the class in front of you needs rather than one undifferentiated sheet. Answers are given so these can go straight onto a board.

1Fluency

The method on its own, with friendly numbers. Set these first and move on quickly once they are secure.

  1. BoundsA mass is 7.4 kg to 1 decimal place. Give the bounds.
    7.35 ≤ m < 7.45
  2. Percentage error3.2 is used for a true value of 3.15. Find the percentage error to 3 sf.
    1.59%

2In context

The same skill inside a real situation, where the first job is working out what is being asked.

  1. Bounds in contextA rectangular room is measured as 5 m by 4 m, each to the nearest metre. Give the least and greatest possible floor area.
    Least 4.5 × 3.5 = 15.75 m²; greatest 5.5 × 4.5 = 24.75 m²
  2. Choose an accuracyA journey of 12.4 km takes 0.37 hours. A student gives the speed as 33.513 513 51 km/h. What should they write, and why?
    33.5 km/h, to 3 sf. The data has 3 and 2 significant figures, so an answer to ten figures claims precision that is not there.

3Challenge

Reasoning, working backwards, or spotting an error. These are where the top grades are decided.

  1. Bounds for a quotientSpeed = distance / time, with distance 100 m to the nearest metre and time 12 s to the nearest second. Find the greatest possible speed and explain the pairing.
    100.5 / 11.5 = 8.74 m/s. Greatest quotient needs the largest top and the smallest bottom, which is the opposite pairing to a product.
  2. Reason about compounding errorWhy can a volume computed from three measurements each accurate to 2% be out by more than 6%?
    The errors multiply rather than add. 1.02³ = 1.0612, so just over 6%, and the extremes push it further. The mark is for saying the errors compound.

Practicalities

Works on a phone. Nothing is loaded from any other site, so it runs behind a school firewall, and nothing a student does is saved or sent anywhere.