Study Notes on Levels of Measurement

Chapter 2: Levels of Measurement and Their Importance

  • Source: Salkind, Tests and Measurement 3e. SAGE Publishing (2018) unless otherwise noted.

Introduction to Variables and Measurement

  • Key Questions: What is a variable? What is measurement?
  • Source: Salkind, Tests and Measurement 3e. SAGE Publishing, 2018.

Origin of Measurement Levels

  • S.S. Stevens (1946): The foundational work by Stevens focused on how different types of variables are measured.
    • Key Concerns:
    • How does the precision of measurement help in classifying variables?
    • Is it possible to create a system to categorize outcomes based on the characteristics of variables?
    • Source: Salkind, Tests and Measurement 3e. SAGE Publishing, 2018.

Importance of Measurement Methods

  • Core Concept: The methodology used to measure a variable significantly impacts the level of measurement.
    • Example: Height Measurement
    • Group classification based on height can simply rank individuals or provide detailed numeric data which is more informative.
    • Source: Salkind, Tests and Measurement 3e. SAGE Publishing, 2018.

Levels of Measurement

  • Definition: The levels of measurement indicate how much information is conveyed by the outcome measure.
    • There are four levels of measurement: Nominal, Ordinal, Interval, and Ratio.
    • Source: Salkind, Tests and Measurement 3e. SAGE Publishing, 2018.

Nominal Level of Measurement

  • Characteristics:
    • Focus on differences in quality rather than quantity.
    • Categorical or discrete in nature.
    • Qualitative - involves labels without inherent order.
    • Mutually Exclusive: A variable can only belong to one category at a time.
    • Example:
    • Categorization such as nurse vs. non-nurse (Verhaeghe et al., 2003).
    • Room assignments such as blue room vs. red room.
    • Source: Salkind, Tests and Measurement 3e. SAGE Publishing, 2018.

Ordinal Level of Measurement

  • Characteristics:
    • Allows for variables to be ordered or ranked.
    • Groups are placed in categories with an implied order, indicating “better than” or “worse than” relationships.
    • Limitations: Does not indicate how much “better” one category is over another.
    • Example: Ranking common childhood fears (Burkhardt, Loxton, & Muris, 2003).
    • Applications:
    • Pop music charts, college football standings.
    • Note: There exists some quantitative difference, but the exact magnitude between ranks is unknown.
    • Source: Salkind, Tests and Measurement 3e. SAGE Publishing, 2018.

Interval Level of Measurement

  • Characteristics:
    • Assigns values based on an underlying continuum with equal intervals.
    • Provides a significant amount of information, revealing what the score truly means.
    • Facilitates judgment about performance or levels (how much “better” or “worse”).
    • Example: Scores on the Wechsler Intelligence Scale for Children with low birth weight participating in early intervention programs (Hill, Gunn, Waldfogel, 2003).
    • Equal Intervals: Each incremental increase reflects an equal amount of the variable being measured. For example, the interval between scores of 1 and 4 is equivalent to that between 6 and 9.
    • Source: Salkind, Tests and Measurement 3e. SAGE Publishing, 2018.

Ratio Level of Measurement

  • Characteristics:
    • Assumes the existence of an absolute zero, denoting a complete absence of the trait being measured.
    • Commonly utilized in physical and biological sciences.
    • Examples in behavioral or social sciences are rare.
    • Interpretations of Zero:
    • A zero score on an intelligence test does not indicate a lack of intelligence, while a zero on a finger-tapping task represents absence of a specific ability.
    • Examples:
    • Measurement categories such as rainfall, height, weight, and temperature.
    • Nutritional Values:
    • The difference in calorie content between food items exemplifies true zero – the absence of that nutrient is meaningfully represented.
    • Source: Salkind, Tests and Measurement 3e. SAGE Publishing, 2018.

Summary of Characteristics Defining Levels of Measurement

  • Table 2.1: Summary of Levels of Measurement and Definitions:
    • Nominal: Least information, categorical.
    • Ordinal: More information, order can be assigned, but lacks quantitative details.
    • Interval: Most information available, allows ranking and can measure quantitative differences.
    • Ratio: Most precise, allows for absolute zero, can measure true quantities.
    • Source: Salkind, Tests and Measurement 3e. SAGE Publishing, 2018.

Key Takeaways for Measurement Techniques

  • Lesson: When selecting a technique for measuring a variable, opt for the highest level of measurement.
    • Most Preferred: Interval level is often sought for access to the most detailed information.
    • Example Application: Assessing strength in senior citizens can differ when describing them simply as 'weak or strong', ranking them, or specifying weights they can lift.
    • Source: Salkind, Tests and Measurement 3e. SAGE Publishing, 2018.

Final Insights on Measurement

  • The level of measurement influences accuracy and precision in research findings.
    • Important Note: Accurate measurement processes are linked to higher precision levels of measurement.
    • Caution: Researchers must be careful when treating ordinal and interval data as these are inherently different. Misinterpretations can arise.
    • Example of Misinterpretation: Treating ordinal scores such as personality tests as if they were interval-level observations can lead to incorrect conclusions.
    • Complexity in Instruments: Differences in intelligence test scores (e.g., 100 vs. 105 versus 125 vs. 130) imply varying implications based on their interval nature.
    • Source: Salkind, Tests and Measurement 3e. SAGE Publishing, 2018.