Skip to content
Tutorials Center
My learning

Topics

Lesson 10 of 45 ยท Measurement, Error and Statistics

Chapter 2 review: measurement, error and statistics

Review accuracy, precision, error propagation, confidence intervals and significance tests with a one-page cheat sheet, recap table and ten-question quiz.

This chapter gave you the machinery for saying how much to trust a number. Before moving to sampling and calibration, check that the pieces fit together.

Four cards: error types, statistics of replicates, confidence interval formula, and the three significance tests.
Caption: chapter 2 in one picture: error types, replicate statistics, the confidence interval and the three tests.

Recap table

Lesson Key ideas Formula
Accuracy and propagation bias, scatter, blunders; significant figures sums: absolute in quadrature; products: relative in quadrature
Statistics of replicates mean, s, RSD, bell curve xฬ„ ยฑ t s / โˆšn
Significance tests t, F, Q t = abs(xฬ„ โˆ’ ฮผ) โˆšn / s, F = sโ‚ยฒ/sโ‚‚ยฒ, Q = gap/range

Try it

Python One data set, the whole chapter

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Check your understanding

Quick check ยท 10 questions

of 10 answered You scored / 10 ยท perfect!

  1. 1. Which error is revealed by analyzing a certified reference material?

    A CRM has a known true value, so a consistent offset shows as bias.
  2. 2. The relative uncertainties of three independent factors are 1 %, 2 % and 2 %. The relative uncertainty of their product is

    โˆš(1 + 4 + 4) = 3 %.
  3. 3. 0.0250 has how many significant figures?

    The leading zeros do not count; 2, 5 and the trailing 0 after the decimal point do.
  4. 4. The standard deviation of the mean of n replicates is

    Averaging reduces the uncertainty by โˆšn.
  5. 5. For n = 6 and 95 % confidence, which t value is used (2.571, 2.447, 2.365)?

    Degrees of freedom are n โˆ’ 1 = 5, and t(95 %, 5) = 2.571. The value 2.447 belongs to 6 degrees of freedom.
  6. 6. Which test compares the precision of two methods?

    The F-test compares two variances.
  7. 7. Results: 3.10, 3.12, 3.11, 3.13, 3.30. Q = 0.85 and the 90 % critical value for n = 5 is 0.642. The suspect value is

    Q exceeds the critical value, so rejection is supported, but only with a documented reason in practice.
  8. 8. A narrow confidence interval means

    The interval reflects random scatter and n; bias is invisible in it.
  9. 9. A t-test gives p = 0.30. At ฮฑ = 0.05 you conclude

    p > ฮฑ, so Hโ‚€ is not rejected.
  10. 10. The best way to halve a confidence interval, without changing s, is to

    The width scales as 1/โˆšn.

Summary

  • Distinguish bias from scatter, and propagate uncertainty in quadrature.
  • Describe replicates by mean, s, RSD and a t-based interval.
  • Use t, F and Q tests to make decisions, and discard data only with cause.

Next: Chapter 3: sampling and calibration