Skip to content
Tutorials Center
My learning

Topics

Lesson 15 of 44 ยท Measurement and Sampling

Chapter 3 Review: Measurement and Sampling

Review levels of measurement, reliability, validity and sampling methods with a recap table and a ten-question mixed quiz covering all of chapter 3.

You now know how numbers attach to mental and behavioral variables, how to judge a measure, and how a sample can fairly represent a population. This review puts the chapter on one page and tests it.

A one-page summary of chapter 3: four levels of measurement, reliability and validity, and sampling methods.
Chapter 3 in one picture: pick the level of measurement, check reliability and validity, and sample in a way that avoids bias.

Recap table

Topic What to remember
Levels Nominal, ordinal, interval, ratio; each level adds a property and allows more statistics
Reliability Consistency: test-retest, internal consistency (alpha), inter-rater (kappa)
Validity of measures Face, content, criterion and construct; convergent and discriminant evidence
Relationship Reliability is necessary but not sufficient for validity
Sampling Random, stratified, cluster versus convenience and volunteer
Bias vs error Bias is systematic and not fixed by size; error is random and shrinks with size

Check your understanding

Quick check ยท 10 questions

of 10 answered You scored / 10 ยท perfect!

  1. 1. Which variable is measured at the ratio level?

    Words recalled has equal gaps and a true zero, so ratios make sense.
  2. 2. The median is most appropriate for which type of data?

    It uses order only. Nominal data call for the mode.
  3. 3. A questionnaire gives wildly different scores when the same people repeat it a week later. It lacks

    Inconsistent scores across time show low test-retest reliability.
  4. 4. What does Cronbach's alpha estimate?

    It summarizes how well items in a scale hang together.
  5. 5. A scale correlates strongly with another measure of the same construct and weakly with unrelated ones. This supports

    Convergent plus discriminant evidence supports construct validity.
  6. 6. A measure that is consistent but off target is

    Consistency does not guarantee that the measure captures the intended concept.
  7. 7. Which sampling method guarantees proportional representation of subgroups?

    Stratified sampling draws randomly within each subgroup in proportion to its size.
  8. 8. Why did a bigger sample not rescue the Literary Digest poll?

    A biased method gives a systematically wrong answer, and more of the same only makes it look precise.
  9. 9. How does the 95% margin of error change when n is multiplied by four?

    Error shrinks with the square root of n, so four times the people halves it.
  10. 10. Which statement is correct?

    They are different steps that protect external and internal validity respectively.

Summary

  • Choose the highest practical level of measurement and match statistics to it.
  • Check reliability first, then build the case for validity with several kinds of evidence.
  • Use probability sampling when you can; remember that size cannot cure bias.

Next: Chapter 4 Overview: Descriptive and Correlational Studies