Notes: Measurement Scales and Variable Types (Nominal, Ordinal, Interval, Ratio)

Measurement Scales, Populations, and Statistics: Detailed Study Notes

  • Overview: The transcript covers data types, how researchers classify variables, and why the scale of measurement matters for analysis. It also introduces populations vs. samples, descriptive vs. inferential statistics, and practical calculation tools (sums, means, rounding, etc.). The instructor emphasizes that many measurements are constructed (not purely observed) and that the choice of scale affects analysis strategy.

1) Key ideas about measurement scales

  • What is being counted vs. what is being measured

    • Some phenomena are countable (e.g., breaths per minute, reaction time) and appear to be objective phenomena.

    • Many variables in psychology are constructed by researchers (interval/ratio scales) rather than being observed directly; this gives researchers latitude in how to anchor and label scales.

  • Anchors on scales

    • Example with two scales where anchors are: very negative on top and very positive on bottom. One scale runs from 0 to 10; another from -5 to +5.

    • Even though the numeric representation differs, the statistical procedures applied to these scales often look the same, because the data are treated as numeric in analyses.

  • Meaning of zero

    • A zero value can have different meanings depending on the scale. In one scale, zero may indicate extreme dislike; in another, neutral or baseline. The meaning of the zero should guide interpretation and analysis.

  • Summary: How scales are represented changes interpretation, not necessarily the mechanics of many statistical procedures.

2) Types of variables (nominal, ordinal, interval, ratio) with examples

  • Nominal variables

    • Definition: Categorical data with no intrinsic order. Values are labels/categories.

    • Example: Major (Psychology, Biology, Other). Coding in data might be 1, 2, 3, but these numbers are just placeholders, not meaningful quantities.

    • Conversation in transcript: Major identified as nominal.

  • Ordinal variables

    • Definition: Categorical data with a meaningful order, but the differences between adjacent categories are not necessarily equal or numeric.

    • Example: T-shirt size (Small, Medium, Large) – ordered, but not numeric in a way that supports arithmetic.

    • Example discussed: Income categories (Under $20k, $20k–$40k, etc.) – ordered categories, not exact numeric amounts.

    • Also mentioned: Likert-type agreement scales (Strongly disagree to Strongly agree) are technically ordinal, though often treated as interval in practice.

  • Interval variables

    • Definition: Numeric, ordered with equal intervals between values, but no true zero point.