REVIEW
Transformation of equations in analysis of proportionality through referent models
Data Analysis, Statistics and Probability · Submitted 22 Feb 2006 · 21 references
Pith's verdict: A useful recipe for kinanthropometry, but the general transformation claim overreaches: the derivation silently assumes homogeneity, and two of the paper's own examples violate it.
T0 · AI review
Why Pith says this
Highlight any text to comment on it.
Referee report
The paper
The paper claims
Any parametric equation can be transformed into a single simplified Z-score equation, valid for small Z scores.
The reading
What carries the argument
The central object is the SZSE, the 'simplified in Z score equation,' displayed as equation (7). It is the first-order McLaurin expansion of the Z-score version of a parametric equation, evaluated around the reference model. The machinery has three parts: the reference (Phantom) table of means and standard deviations; the Z-score substitution; and the factoring assumption that the parameter scales homogeneously with height, so the height-ratio term $(\mathrm{height}(i)/\mathrm{height}(\mathrm{Ph}))^n$ can be pulled out of every term. The coefficients $C_{f,x_j}$ are partial derivatives of $f$ with respect to each variable, evaluated at the reference point.
Extended reading notes
Core claim
The central claim is that any parametric equation $f(x_1,\dots,x_m)$ can be transformed into an equation in Z scores of the form $$f(Z) \approx \left[f(\mathrm{Ph}) + \sum_j C_{f,x_j} s_{x_j}(\mathrm{Ph}) Z_{x_j}\right] \left(\frac{\mathrm{height}(i)}{\mathrm{height}(\mathrm{Ph})}\right)^n.$$ The route is to substitute each variable's reference expression $x_j(i) = x_j(\mathrm{Ph}) + s_{x_j}(\mathrm{Ph}) Z_{x_j}(i)\left(\frac{\mathrm{height}(i)}{\mathrm{height}(\mathrm{Ph})}\right)^{n_j}$, expand in a McLaurin series around the reference point, and drop terms beyond first order. The approximation is expected to hold when the Z scores are small; on the ponderal index the paper verifies agreement for $|Z_{\mathrm{weight}}|<4$, and for the bone-mass equation it states that the difference from the parametric result stays below 1% inside $(-4,4)$.
Load-bearing premise
The derivation assumes the parameter scales with a single height exponent $n$, so the same height-ratio factor can be factored out of every term; for the bone-mass equation that is not actually true, and the paper chooses $n=2.924$ by hand as an average of the physical dimension 3 and the equation's own exponent 2.848, so the claimed accuracy depends on that choice.
Editorial extensions
If this is right
- If the simplified equation is correct, any linear body-composition equation (e.g., skin-fold fat-mass formulas) has a Z-score version built from the same coefficients, each weighted by the reference standard deviation of its variable.
- For the ponderal index, the parametric definition collapses to $PI \approx 42.41 - 1.88\,Z_{\mathrm{weight}}$, valid whenever $|Z_{\mathrm{weight}}| < 4$.
- Height enters equations that originally omit it through the height-ratio factor and through the Z scores themselves, so two individuals with identical skin-folds but different heights receive different predicted fat percentages.
- Equation (5) supplies a per-study threshold for how large a Z score may be before the referent model is no longer adequate, tying the method's validity to the study's significance level.
Reading between the lines
- Pith inference: the same derivation should apply to any referent model that supplies a table of means and standard deviations, not only the Phantom model, so the method is portable to other populations or types of objects.
- Pith inference: the hand-chosen exponent $n=2.924$ for the bone-mass equation could be tested and, if necessary, replaced by a value fitted to data, since the height-ratio exponent controls how rapidly errors grow as height moves away from the reference height.
- Pith inference: the $|Z|<4$ boundary is inferred from one worked example; a systematic comparison across several non-homogeneous equations would show whether it is a general property of the expansion or specific to the ponderal index.
Original abstract
In proportionality of objects, samples or populations, usually we work with Z score of proportionality calculated through referent models, instead directly with the variables of the objects in itself. In these studies we have the necessity to transform, the equations that use the variables of the object, in equations that directly use like variables Z score. In the present work a method is developed to transform the parametric equations, in equations in variables Z using like example the studies of human proportionality from the Phantom stratagem of Ross and Wilson.
The compiled record
The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
The abstract, sentence by sentence
In proportionality of objects, samples or populations, usually we work with Z score of proportionality calculated through referent models, instead directly with the variables of the objects in itself. In these studies we have the necessity to transform, the equations that use the variables of the object, in equations that directly use like variables Z score. In the present work a method is developed to transform the parametric equations, in equations in variables Z using like example the studies of human proportionality from the Phantom stratagem of Ross and Wilson.
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | In proportionality of objects, samples or populations, usually we work with Z score of proportionality calculated through referent models, instead directly with the variables of the objects in itself. | background 1.00 | 0.07 | 0.06 | asserted 0.98 |
| s1 | In these studies we have the necessity to transform, the equations that use the variables of the object, in equations that directly use like variables Z score. | background 0.94 | 0.16 | 0.08 | asserted 1.00 |
| s2 | In the present work a method is developed to transform the parametric equations, in equations in variables Z using like example the studies of human proportionality from the Phantom stratagem of Ross and Wilson. | contribution 0.87 | 0.92 | 0.56 | asserted 1.00 |
What was examined, and what was not
- arXiv listing abstract v1
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
References
Discussion
Disagree with the review, or know something it missed? Leave a comment, highlight any passage above to comment on it, or challenge this review formally.