Skip to content
IGCSE·Tuition
Physics · Lesson

Calculate a repeated-measurement mean

Your timings do not agree with each other, and one of them looks completely out of place.

On this page
  1. How do I calculate a mean properly?
  2. Worked example
  3. The mistake to watch for
  4. Check yourself
  5. Where this leads next

A mean is the sum of the readings divided by how many readings you used. In measurement and quantities you meet it whenever an experiment is repeated to give a more reliable value.

The skill has two parts: deciding which readings belong in the average, and dividing by the correct number.

How do I calculate a mean properly?

  1. Look at all the readings first. Do they cluster? Is one far from the others?
  2. Decide about any anomalous result and say why it is excluded.
  3. Add the readings you are keeping.
  4. Divide by the number you kept, not the number you started with.
  5. Give the mean to the same number of decimal places as the readings, with the unit.

Worked example

A student times a trolley rolling down a ramp five times: 2.31 s, 2.28 s, 2.35 s, 3.05 s, 2.30 s. (Invented example data.)

Step 1, look at the set: four readings are between 2.28 s and 2.35 s. The 3.05 s reading is about 0.7 s away from all of them.

Step 2, decide: 3.05 s is anomalous, possibly a late stop. It is left out, and the reason is stated.

Step 3, add the rest: 2.31 + 2.28 + 2.35 + 2.30 = 9.24 s.

Step 4, divide by four readings: 9.24 ÷ 4 = 2.31 s.

Answer: mean time = 2.31 s (anomalous 3.05 s omitted).

Another everyday use is finding a period. If 10 oscillations of a pendulum take 18.5 s (itself a mean of repeated timings), one oscillation takes 18.5 ÷ 10 = 1.85 s.

The mistake to watch for

There are two common slips. The first is to keep the anomalous reading in the sum and average everything.

Mistaken answer: (9.24 + 3.05) ÷ 5 = 12.29 ÷ 5 = 2.458 s, about 2.46 s

This value is higher than four of the five readings and close to none of them.

The second is to remove the anomalous reading but still divide by five:

Mistaken answer: 9.24 ÷ 5 = 1.848 s

The correction is to divide by the number of readings actually used. Always ask whether your mean lies inside the cluster of readings. A mean far from every reading is a warning sign.

Check yourself

1. Find the mean of 15.2 cm, 15.6 cm and 15.4 cm.

Show answer

15.2 + 15.6 + 15.4 = 46.2. 46.2 ÷ 3 = 15.4 cm.

2. Twenty oscillations of a pendulum take 34.0 s. Find the period.

Show answer

34.0 ÷ 20 = 1.70 s.

3. A wire’s diameter is measured as 4.8 mm, 4.9 mm, 4.7 mm, 6.1 mm and 4.8 mm. Identify the anomalous reading and find the mean.

Show answer

6.1 mm is far from the cluster, so it is anomalous. Remaining: 4.8 + 4.9 + 4.7 + 4.8 = 19.2. 19.2 ÷ 4 = 4.8 mm.

Where this leads next

After calculating a mean, the natural question is how much to trust it, which is the focus of stating a realistic uncertainty from supplied instrument data. You may also want to revisit reading an analogue scale, and then try the measurement and quantities practice set. The bounds and rounding explainer shows how many decimal places a mean can honestly carry.

If experiment questions feel like guessing which readings to keep, our teachers can work through examples with you in online one-to-one Physics tuition.

Questions people ask

Why do we repeat measurements and take a mean?

Each single reading has small random errors, such as slightly early or late stopwatch clicks. Averaging several readings reduces the effect of these random errors, so the mean is usually closer to the true value than any one reading.

What is an anomalous result?

An anomalous result is a reading that clearly does not fit the pattern of the others, often from a slip in the method such as a missed start. Identify it by comparing with the rest, say why you are leaving it out, and calculate the mean from the remaining readings.

Why time 10 or 20 oscillations instead of one?

One swing is too quick to time accurately, and the reaction-time error is a large fraction of it. Timing many swings spreads that same error over a bigger total, so dividing by the number of swings gives a more reliable period.

Updated:

Your next step

If averaging and anomalous results still leave you unsure what to write in an experiment answer, a one-to-one teacher can rehearse the reasoning with you on fresh data sets.

Paid one-hour trial at your assigned teacher’s confirmed rate, starting from RM80.

Tuition is arranged with a parent or guardian. Send them this page on WhatsApp and they can enquire for you.

Parents: enquire here

  • 9,000+ students helped through our service
  • 9+ years helping IGCSE students