Introduction
In our last two lessons, we began building the ROCCC framework for information validity. We first established that data must be Reliable, meaning it's collected using consistent and standardized methods. Then, we examined Originality, distinguishing between trustworthy primary data (which you collect yourself) and more suspect secondary data (collected by others), using the example of professionally-fitted vs. self-reported bra sizes.
Today, we address the first 'C' in ROCCC: Comprehensive. A dataset can be reliable and original, but if it only represents a narrow slice of the phenomenon you're studying, any conclusions you draw will be incomplete at best and misleading at worst. This lesson is designed to help you assess if a dataset is comprehensive enough for the research question, focusing on the specific challenge of covering sufficient body types in a bra sizing study.
Defining "Comprehensive" Data
In data analysis, "comprehensive" does not simply mean "large." A dataset with millions of records can still be non-comprehensive if all the records are of the same type.
A comprehensive dataset is one that captures the full spectrum of variation in the characteristics relevant to your research question. It should be representative of the population you wish to understand or make claims about.
For instance, if your goal is to design a new line of bras for the general market, a dataset consisting of measurements from 1,000 women who all wear a 34B is large but not comprehensive. It tells you nothing about the needs of women who wear a 30G or a 42C. Comprehensiveness is fundamentally about coverage and representativeness.
Case Study: The True Complexity of Bra Sizing
To understand why comprehensiveness is so critical in your field of interest, we first need to appreciate the underlying complexity of the measurement system itself. The familiar band-and-cup size notation can be deceptive.
Sister Sizing Bra Sizes and How Do Bras Sizes Work & Bras Sizing Explained | Features Tips & Benefit
To start, let's watch a brief segment from the video 'Sister Sizing Bra Sizes' by Nicola Crook Online. This will quickly demonstrate that a given cup letter does not represent a fixed volume, which is a crucial concept for understanding data needs.
Please watch from 05:07 to 06:18. Focus on the explanation of why a 34D cup and a 36D cup are not the same size, even though they share the same cup letter.
This reveals our first challenge: cup volume is relative to band size. This relationship implies that a truly comprehensive dataset for bra design cannot treat cup and band as simple, independent variables. It must capture a sufficient range of their combinations to reflect the actual variation in breast volume.
However, the complexity goes much deeper than just volume. Effective bra design must account for the vast diversity in breast morphology—shape, position, and symmetry.
Utilisation of 3D Body Scanning as a Research Tool When Establishing Adequate Bra Fit
The research paper 'Utilisation of 3D Body Scanning...' provides an excellent overview of the numerous factors that influence breast characteristics and the methods researchers use to capture this variation. This will help us define what a comprehensive dataset for bra design should encompass.
Please review the following sections of this paper: 'Review Topic 1 - Breast Shape and Profiling' (starts on page 7): Skim this section to appreciate the many variables that influence breast shape (genetics, ethnicity, age) and the terminology used to classify shapes (e.g., hemispherical, conical, ptotic). 'Differences Affecting Fit' (starts on page 11): Read this subsection to see how researchers try to classify body shapes using 3D scanning technology and statistical methods to address fit issues. 'Profiling Methodology' (starts on page 38): Look over this section, particularly the tables and figures. Notice how researchers create ratios and use percentiles to profile different aspects like bust height, bust spread, and breast drop. You don't need to memorize the formulas, but appreciate the statistical approach to quantifying shape. First paragraph of the 'Conclusions' (page 47): Read this paragraph to see how the study resulted in a five-point profiling tool to quantify differences and assess consumer characteristics.
This paper powerfully illustrates what comprehensiveness means in practice. A comprehensive dataset for bra design must go beyond simple band and cup measurements to include data on a wide array of morphological features. The development of a "five-point profiling tool" is a direct attempt to systematically capture and categorize this variation.
Acknowledging the Limits of Comprehensiveness
Achieving a perfectly comprehensive dataset is often an unachievable ideal. Practical constraints like time, budget, and participant access mean that most research datasets have limitations. The crucial skill is to recognize and articulate these limitations.
Comparative Morphological Evaluation of Young Women's Breast ...
Let's look at another study that explicitly acknowledges its own lack of comprehensiveness. This is a hallmark of good scientific practice.
Please read the paragraph under 'Figure 5' in the linked section (starts with 'However, there are several limitations to this study...'). Note how the authors clearly state the limitations of their sample and call for a more comprehensive information base.
The authors' honest self-assessment—that their sample of young Chinese women with B/C cups "could not reflect the whole Chinese young female population"—is exactly the kind of critical evaluation you must apply to any dataset. When you encounter a dataset, you should always ask: Who is missing?
Here, the missing groups are clear:
- Women outside the 18-25 age range.
- Women with cup sizes smaller than B or larger than C.
- Women of non-Chinese ethnicity.
A conclusion from this dataset, such as "Bra A provides the best lift," is only valid for the specific population segment studied. It cannot be generalized without further, more comprehensive data.
Application to Polyamorous Relationship Research
Let's translate this concept to your social research interests. Suppose you are tackling the research question: "What are the primary factors that contribute to long-term satisfaction in single-female, multiple-male polyamorous relationships?"
The term "long-term" immediately sets a high bar for comprehensiveness.
An incomprehensive dataset for this question might be one gathered from a single online forum for people new to polyamory. Such a dataset would likely be over-represented with:
- Relationship Duration: Relationships under three years old.
- Age: Individuals in their 20s and 30s.
- Relationship Structure: Dynamics that are still being negotiated and have not yet stabilized.
- Self-Selection Bias: Individuals who are actively seeking advice and may be experiencing more conflict than the average.
Drawing conclusions from this dataset would be perilous. You might find that "setting clear rules" is the dominant factor for satisfaction, but this could be an artifact of new relationships, where rule-setting is a primary activity. This finding might not hold for a 15-year-old polyamorous unit where dynamics are more fluid and intuitive.
A comprehensive dataset would need to be deliberately constructed to include a wide range of:
- Relationship Durations: From 1 year to 20+ years.
- Life Stages: Including partners who are students, have established careers, are raising children, or are nearing retirement.
- Living Arrangements: Covering units that cohabitate, live separately, or have a mix of both.
- Cultural Backgrounds: To see if satisfaction factors differ across various social contexts.
Only by ensuring this breadth of coverage can you have confidence that your findings are truly about long-term satisfaction in general, not just the experiences of a narrow, easily accessible subgroup.
Conclusion
In this lesson, we have focused on the 'Comprehensive' aspect of the ROCCC framework. We've seen that a dataset's utility is defined not just by its size, but by its ability to represent the full range of variation relevant to a research question.
Key Takeaways:
- Comprehensiveness is about representativeness, not volume. It means capturing the full spectrum of characteristics in your population of interest.
- Assessing comprehensiveness requires domain knowledge. To know if a bra-fitting dataset is comprehensive, you must first understand the complex morphology of the human body.
- Sampling is key. A dataset's comprehensiveness is heavily dependent on the sampling strategy used. Convenience samples are often not comprehensive.
- A lack of comprehensiveness limits generalizability. Findings from a narrow dataset cannot be safely applied to a wider population.
Preview of the Next Lesson:
We have now examined data for its Reliability, Originality, and Comprehensiveness. Next, we will address the second 'C' in our framework: Current. Is the data recent enough to be relevant? We will explore the concept of data decay and how to decide if a dataset is still timely enough for your analysis.