PSYC2009 Study Guide - Final Guide: Level Of Measurement, Bar Chart, Cumulative Frequency Analysis
Measurement, Variables and Data
• Construct - a concept, usually a characteristic or property, that underlies measurement.
• Indicator - constructs are indicators of another construct if knowing about one provides
information about the other.
• Indicators are operationalisations of their respective constructs if they give us access to some
aspect of their real-world manifestation.
• If an operationalisation captures the totality of a construct's real-world manifestation, then we
claim that it measures that construct.
• Valid - an indicator is valid if it is not contaminated by some other influence or error.
• Reliable - an indicator is reliable if it produces the same result every time under identical
conditions.
• Variable - an operationalisation of a construct that can change.
What kinds of indicators should we use?
There are at least 5 ways to find out things about people's behavioural or attitudinal proclivities.
1. Self-reports
2. Direct observations
3. Ask people about someone else whom they know well.
4. Ask people about others 'in general' or 'on average'.
5. Find indirect observational evidence, eg: examining people's rubbish for evidence of
consumer habits and preferences.
Eg: Stress measurements:
1. Direct measurement of physiological indicators, eg: blood pressure.
2. Self-reports of a checklist of symptoms, and self-reports of life events/occurrences.
3. Self-evaluation of feelings and thoughts.
Discrete and Continuous Variables
• Discrete - can only take certain values and no others, eg: gender.
• Continuous - those which could take any value over their "range".
Types of Scales and Measurement
• Scales and measurement determine what we can meaningfully say about that variable, what
questions we can answer with it and what hypotheses we can test.
• Nominal - a variable has categories or states that differ in kind but not in any order or degree,
eg: religion or gender.
• Ordinal - a variable has categories or ranks that are ordered but whose differences are not
quantifiable, eg: disagree --> strongly agree.
• Interval - a variable has values on a scale whose differences are quantifiable but which lacks a
true zero-point, eg: IQ.
• Ratio - a variable has a scale with quantifiable differences a true zero.
How to tell ordinal from interval:
• Are the differences between values on a scale truly comparable?
• Eg: Scale of life stress - count the number of major changes, eg: marriage, that have occurred
in the past 6 months. Someone with a score of 20 is presumed to have undergone more stress
than someone with 15, and likewise than someone scoring 10, but are the differences
between 20 v 15 and 15 v 10 equal? We don't know, unless we are willing to make some
assumptions.
How to tell interval from ratio:
• Even if differences between values on a scale are truly comparable, is the zero-point absolute?
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
Measurement, variables and data: construct - a concept, usually a characteristic or property, that underlies measurement. Indicator - constructs are indicators of another construct if knowing about one provides information about the other. Indicators are operationalisations of their respective constructs if they give us access to some aspect of their real-world manifestation. Eg: stress measurements: direct measurement of physiological indicators, eg: blood pressure, self-reports of a checklist of symptoms, and self-reports of life events/occurrences, self-evaluation of feelings and thoughts. Discrete and continuous variables: discrete - can only take certain values and no others, eg: gender, continuous - those which could take any value over their range. Interval - a variable has values on a scale whose differences are quantifiable but which lacks a true zero-point, eg: iq: ratio - a variable has a scale with quantifiable differences a true zero.
