Assessing Research Question Measurability

Operationalization: From Concept to Data

As a statistician, you are an expert in analyzing variables. The focus of this lesson is the preceding step: defining those variables in the first place. Operationalization is the process of defining an abstract concept in a way that allows it to be measured. This involves specifying:

• The Variable: The characteristic being measured (e.g., comfort, satisfaction).
• The Data Type: How the variable is categorized. This can be:
  • Categorical: Represents groups or categories.
    • Nominal: Unordered categories (e.g., wire type: steel vs. plastic).
    • Ordinal: Ordered categories (e.g., satisfaction: low, medium, high).
  • Numerical: Represents quantities.
    • Discrete: Countable integers (e.g., number of partners).
    • Continuous: Can take any value within a range (e.g., bust circumference in cm).
• The Metric: The specific unit or scale used for measurement (e.g., inches, count per month, a score on a 1-to-7 Likert scale).

Let's see how this is done in practice by examining a formal research paper related to your interest in bra design.

Application 1: Quantifying Bra Fit and Well-being

We'll start with our bra design question. It contains several concepts that need to be measured: "omega breast shape," "incidence of underarm wire-poking," and the comparison between "BRT method" and "standard sizing."

To see how researchers tackle similar challenges, we'll look at a study that investigates the relationship between breast size and health.

The study you just reviewed demonstrates several key measurement techniques:

• Creating a Custom Metric: The "Breast size score (BSS)" is a brilliant example of operationalization. Researchers converted standard alpha-numeric bra sizes (like 12D or 14C) into a single, continuous ordinal score. This allows for more powerful statistical analysis than using size categories alone.
• Using Standardized Instruments: For complex concepts like "quality of life," they didn't invent their own measure. They used validated, pre-existing questionnaires like the SF-36 and BREAST-Q. This ensures the measurement is reliable and comparable to other studies.
• Simplifying for Clarity: For "upper back pain," they used a simple binary (Yes/No) categorical variable. For "body satisfaction," they used a Numerical Rating Scale (NRS) from 0-10. These choices represent a trade-off between nuance and ease of collection/analysis.

Now, let's apply these ideas to our own specific question. How would we measure its components?

• Independent Variable: Sizing Method. This is a nominal categorical variable with two levels: (1) BRT method, (2) Standard formula method. Participants would be assigned to one of these two groups.
• Dependent Variable: Incidence of underarm wire-poking. We have several options, each with different properties:
  • Categorical (Binary): A 'yes/no' question: "In the past week, did you experience poking from your underwire?" This is simple but lacks detail.
  • Ordinal (Likert Scale): A frequency question: "Rate the frequency of underarm wire-poking on a 5-point scale, from 1 (never) to 5 (constantly)." This captures more information about the extent of the problem.
  • Numerical (Discrete): A count: "How many times yesterday did you consciously adjust your bra because the underwire was poking you?" This is very specific but may be difficult for participants to recall accurately.

The choice of metric is a strategic one. For your students, it's crucial to understand that this choice dictates the statistical test they can use later (e.g., chi-squared for categorical data, t-test or ANOVA for continuous data).

This simple bra size chart is another example of measurement in action. It operationalizes a continuous physical attribute ('Under Bust Measurement') into a discrete commercial size ('Easy Fit Size' or 'Bra Size').

Application 2: Measuring Relationship Dynamics

Now, let's turn to our question on polyamorous relationships. It contains the concepts of "relationship structure," "frequency of date nights," and "satisfaction."

Quantifying subjective experiences like satisfaction or commitment is a core challenge in social research. The following study on polyamory provides excellent examples of how to do this.

This paper shows that even the most personal and subjective aspects of a relationship can be made measurable. Key takeaways include:

• Multi-Item Scales: To measure "satisfaction" or "commitment," the researchers didn't rely on a single question. They used multi-item scales adapted from the Investment Model Scale (IMS). Averaging responses across several related questions provides a more reliable and valid metric than any single item could.
• Quantifying Behavior: "Relationship communication" was measured by asking for the frequency of discussion on various topics, using an ordinal scale from "never" to "daily." This turns a qualitative concept into quantifiable data.
• The Challenge of Metrics: The discussion around "Percentage of time spent on sexual activity" is particularly insightful. The authors acknowledge the limitations of their metric—it's a proportion, not an absolute amount, and "sexual activity" is undefined. This highlights a critical point for your students: every measurement choice has trade-offs and potential flaws that must be considered.

Let's apply this to our specific question:

• Independent Variable: Monthly frequency of scheduled, one-on-one 'date nights'. This is a numerical, discrete variable. The metric is a simple integer count (0, 1, 2, ...).
• Dependent Variable: Satisfaction scores on 'communication' and 'intimacy'. As established in the last lesson and reinforced here, this would be measured using an ordinal scale, likely a Likert-type scale for each item (e.g., 1-7). The score on each item would be the metric.

The image below presents an attempt to place relationship orientation on a numerical scale. While perhaps not a rigorously validated scientific instrument, it illustrates the drive to quantify even deeply personal identities. For a social researcher, it prompts interesting questions about the validity and utility of such scales. Do the ordered points make sense? What is lost and what is gained by assigning a number to one's orientation?

Conclusion

In this lesson, we explored the 'M' in SMART: Measurability. We saw that making a question measurable is a practical process of operationalization, where abstract concepts are translated into concrete variables with defined data types and metrics.

Key Takeaways:

• Operationalization is Key: A question is only measurable if you can define how you will capture its core concepts as data.
• Choose Your Tools: Measurement involves selecting or designing tools, such as binary questions, Likert scales, or validated multi-item questionnaires.
• Every Metric is a Choice: There is no single perfect way to measure a complex concept. The choice of a metric (e.g., binary vs. scale, direct count vs. self-report) is a trade-off between simplicity, nuance, and validity. This choice has direct consequences for the statistical methods you can employ.

Preview of the Next Lesson:

We have now formulated questions that are both Specific and Measurable. But can we actually collect this data? In our next lesson, we will examine the 'A' in SMART: Achievable. We will discuss how to assess the feasibility of a research plan given real-world constraints on time, resources, ethics, and access to participants.