Dictionary Review


Author: Everitt, Brian

Translators:Duca, Piegiorgio; Bellini, Aldo; Maistrello, Mauro

Publisher: Il Pensiero Scientifico Editore

Publication date: Italian version: July 1998 (English version: 1995, Cambridge University Press)

Price: €41.00

ISBN: 88-7002-820-8

Available from: Il Pensiero Scientifico Editore

via Bradano 3/c, I-00199 Roma, Italy


Languages:Italian and English

Number of pages: 436

Number of entries: 1600+


For the last 20 years, Dieter Rasch's Biometrische Wörterbuch (originally D-E-F-S-I-R-H-Cz-Pl*. Berlin: VEB Deutscher Landwirtschaftsverlag, 1987; now reissued by Elsevier) has been the statistical reference of choice.

Rasch's work towers above the other two known contributions in this field: The EU unindexed Glossaire de Statistique (Da-D-E-F-I-Du*. Luxembourg: CCE, 1975, internal document) and my own equally forgettable English-Italian Glossary of Statistics (Sci-Tech Translation Journal, Vol. X, No. 3, July 1994).


Truth be told, statistics is not a glamorous field. Statisticians suffer all the indignities of mathematicians, but without their prestige and obvious genius. Probably because of that stigma, our Dizionario was well hidden and I landed it thanks only to the exceptional efforts of Libreria Goldoni (calle dei Fabbri 4742, I-30125 Venice). Still, it was well worth the wait.


It is a 12" x 20" hardcover, solidly sewn, with boldfaced headwords and cross-references in italics. The recueil consists of about 1600 terms and definitions, 70 figures and zero typos. The lemmata cover investigational, genetic and epidemiological research, and the figures illustrate all types of diagrams and graphic representations.


Ignore the preface. Brian Everitt, Institute of Psychology, London, seems to indicate that he is absolutely in disagreement with the custom of asking the readers to write and correct the author's mistakes. Still, he is a statistician and therefore too conventional and optimist not to follow tradition on this matter. Therefore, corrections are welcome and said contributions shall be recognized in future editions.

I suspect that roundabout request to be an excuse, used by Everitt to quote and chastise a Prof. Sutherland, allegedly guilty of having published a not too perfect Dictionary of Psychology. The possible relevance to the reader is beyond my understanding.


Instead, it's important to pay attention to the translators' note. With the purpose of enriching Everitt's work, Duca, Bellini and Maistrello [DBM] inform us that they have listed in both directions the English terms included in the original text and their Italian translated equivalents. What a brilliant decision! It transforms a good monolingual reference into a solid bilingual dictionary, something we can really use.


Furthermore, the definitions originally prepared by Everitt and translated by DBM are useful twice, per se and as ready-to-use Italian phraseology. This is exceptional: Although this approach is the focus of Vincenzo Marino's Vocabolario Medico Fraseologico (Padua: Piccin, 1985) and represents a strong point in favor of Taber's Cyclopedic Medical Dictionary (Thomas, Clayton ed. Taber- Dizionario Enciclopedico di Scienze Mediche. Milan: McGraw-Hill Libri Italia, 1994), in general it has been ignored, last but not least by Gould, Chiampo et al. (Dizionario Enciclopedico di Medicina. I-E*. Bologna: Zanichelli/McGraw-Hill, 1988).


DBM language is very current. Some might deplore the passing of σ2, scarto quadratico medio and errore standard, but nowadays everybody agrees with DMB and equates standard deviation with deviazione standard, confining Rasch's scostamento quadratico medio with the 19th century classics from which it was derived.

I personally prefer a simpler voce to the borrowed noun item (as in gli item di un questionario), but there is little doubt of the latter's diffusion in the literature. The same can be said of another barbarism, scaling instead of messa in scala. Conversely, trimming is creepily rendered as trimmatura and trimmed as trimmato, abandoning their more prosaic pairings ritocco e ritoccato.


Still, I find it difficult to criticize a dictionary because it's too updated and even more difficult to raise my voice against DMBs' success in injecting a glimpse of excitement in a field not known for its extreme ebullience.


In real terms, the translators' potential debt with EDMB [Everitt + DMB troika] could become quite significant.

The list of experimental designs was the first to attract my attention. It includes 33 headwords in all, from balanced incomplete block design (o disegno a blocchi bilanciati incompleti) to disegno standard or Fibonacci dose design. In the same vein, there are seven estimate/estimations, from minimum chi-squared to robust (stima del minimo chi quadrato; stima robusta); and 47 rates, from age-specific incidence rates or tassi di incidenza età-specifici to relative standardized mortality rate or tasso relativo standardizzato di mortalità.


My random quantification continued with consistently impressive results. For instance, EDMB include 9 tables, from fourfold to doubly ordered contingency tables (tabella a quattro celle; tabelle di contingenza biordinali), 61 distributions, from bell-shaped (distribuzione a campana o gaussiana) to distribuzioni non centrali or non-central distributions, 15 regressions, from attraverso l'origine to verso la media (regression through the origin; regression to the mean) and 30 coefficients, from beta to Fourier.


Using a list derived from past translations, I also attempted to randomly match the terms of both Rasch and EDMB, as summarized below:


English / Italian Headwords




additive effect

effetto additivo

Instead: Additive gene effect or Effetto genetico additivo


age heaping

arrotondamento dell'età

þ as Rilevazione arrotondata dell'età





audit trail

traccia di audit

þ as Audit trail

bar chart

istogramma a barre/colonne/canne d'organo


þ as Grafico a barre


valore di riferimento

þ (3 strings)

box & whisker plot

diagramma a caselle e segmenti

þ as Grafico a scatola e baffi [?]

class mark

valore centrale della classe


compartment model

modello a scompartimenti


þ as Modello compartimentale [!]

contour plot

istogramma di contorno

distance matrix

matrice di distanze


þ as Misura di distanze [!]

draughtsman's plot

diagramma a scacchiera

þ as Grafico a scacchiera


speranza matematica


adattamento empirico


Instead: Eyeball test

grouped binary data

dati binari

Instead: Grouped data or Dati raggruppati


harmonic analysis

analisi armonica


Hausdorf dimension

dimensione di Hausdorf


hidden association

associazione latente


idempotent matrix

matrice idempotente


indicator variable

variabile indice





kernel function

funzione kernel


Latin square

quadrato latino

þ + 5 more entries


manifest variable

variabile manifesta




þ as Accoppiamento






open label study

studio in aperto



caso estremo; valore erratico


þ as Outlier

path analysis

analisi dei coefficienti di direzione


þ as Path analisi [?]

regression spline

regressione spline


Instead: Funzione spline

two-sided test

test bilaterale; test a due code


þ as Test a due code

variate variable

variabile aleatoria, variabile stocastica


My sample was strongly biased against Rasch (from the start, many strings were included in my list exactly because Rasch did not). Nevertheless, the German benchmark scores an impressive 15 hits out of 32 items or 46%, thus confirming its continuing value. The real surprise is EDMB: scoring 25 hits, they obtain an incredible 78% success rate.


The conclusion is simple: when relevance is tested, our group of statisticians averages far better than most lexicographers.