Changed reference models in proportionality analysis

Enrique Ordaz Romay

DOI: https://doi.org/10.48550/arXiv.physics/0602064

2006-02-10

Abstract:In the statistical analysis of objects, samples and populations with quantitative variables, in many occasions we are interested in knowing the proportions that exist between the different variables from a same object; if these proportions have relations of normality among them, that is to say, if in the sample are these same proportions, and finally if the deviations of concrete physical or statistical normality have meaning. The present study suggests the strategy of to use reference models for the analysis of the proportionality, using like example the studies of human proportionality from the method Phantom of Ross and Wilson and its extension towards scalable models.

Data Analysis, Statistics and Probability,Medical Physics

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is how to accurately evaluate and interpret the proportional relationships between different variables in statistical analysis, especially in the study of human body shape proportions. Specifically, the author focuses on the following aspects of problems:

1. **Proportional relationships between variables**: How do the proportional relationships between different variables within the research object compare with those in the sample or the population?
2. **Normality of proportional relationships**: Do these proportional relationships show consistency in the sample, and are the deviations statistically significant?
3. **Selection of reference models**: What kind of reference model is more appropriate for analyzing proportional relationships, especially how to handle the differences that occur when switching from one reference model to another?

To answer these questions, the author proposes a strategy of using reference models (such as the "Phantom" model proposed by Ross and Wilson) for proportional analysis and further extends this method to adapt to scalable models. This includes:

- **Defining and using reference models**: Establish a standard measurement model through statistical means and other mathematical models.
- **Calculation and interpretation of Z - scores**: Use Z - scores to measure the proportional differences between individuals and the reference model and interpret their statistical significance.
- **Conversion between different reference models**: Propose a compatible method that enables conversion between different types of reference models, thereby ensuring the consistency and comparability of results.

### Formula summary

The key formulas involved in the paper are as follows:

1. **Basic formula for proportional relationships**:
\[
\frac{A_{\text{variable1}}}{A_{\text{variable2}}} = \frac{B_{\text{variable1}}}{B_{\text{variable2}}}
\]

2. **Generalized formula for proportional relationships** (when the variable dimensions are different):
\[
\left( \frac{A_{\text{variable1}}}{A_{\text{variable2}}} \right)^n = \left( \frac{B_{\text{variable1}}}{B_{\text{variable2}}} \right)^n
\]

3. **Formula for calculating the required adjustment value \(\Delta x\)**:
\[
\Delta x = \frac{\left( \frac{B_{\text{variable1}}}{B_{\text{variable2}}} \right) - \left( \frac{A_{\text{variable1}}}{A_{\text{variable2}}} \right)}{\left( \frac{A_{\text{variable1}}}{A_{\text{variable2}}} \right)}
\]

4. **Formula for calculating Z - scores**:
\[
Z_i = \frac{x_i - \mu}{\sigma}
\]

5. **Z - score conversion formula between different reference models**:
\[
Z_{\text{variable}; P}(i) \approx Z_{\text{variable}; P}(P') + Z_{\text{variable}; P'}(i)
\]

Through these formulas and methods, the author aims to provide a systematic framework for more accurate analysis and interpretation of the proportional relationships between different objects, especially in the study of human body shape proportions.