# John FoxMcMaster University | McMaster · Department of Sociology

John Fox

Doctor of Philosophy

## About

96

Publications

29,033

Reads

**How we measure 'reads'**

A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more

32,838

Citations

## Publications

Publications (96)

Predictor effect displays, introduced in this article, visualize the response surface of complex regression models by averaging and conditioning, producing a sequence of 2D line graphs, one graph or set of graphs for each predictor in the regression problem. Partial residual plots visualize lack of fit, traditionally in relatively simple additive r...

The Journal of Statistical Software was founded by Jan de Leeuw in 1996, the year before the Comprehensive R Archive Network (CRAN) first made R and contributed R packages widely available on the Internet. Within a few years, R came increasingly to dominate contributions to JSS. We trace the continuing development of R and CRAN, and the representat...

The multivariate linear model is Y(n×m) = X (n×p) B (p×m) + E (n×m) The multivariate linear model can be fit with the lm function in R, where the left-hand side of the model comprises a matrix of response variables, and the right-hand side is specified exactly as for a univariate linear model (i.e., with a single response variable). This paper expl...

Visual insights into a wide variety of statistical methods, for both didactic
and data analytic purposes, can often be achieved through geometric diagrams
and geometrically based statistical graphs. This paper extols and illustrates
the virtues of the ellipse and her higher-dimensional cousins for both these
purposes in a variety of contexts, inclu...

The R Commander graphical user interface to R is extensible via plug-in packages, which integrate seamlessly with the R Commander's menu structure, data, and model handling. The paper describes the RcmdrPlugin.survival package, which makes many of the facilities of the survival package for R available through the R Commander, including Cox and para...

Since its introduction, APL has frequently been touted as an ideal programming language for statistical applications. Among the attractive features of APL for statistics are its extensibility, the presence of primitives for operations such as sorting, matrix inversion, and arranging data, and powerful facilities for handling matrices and other arra...

This special volume of the Journal of Statistical Software on political methodology includes 14 papers, with wide-ranging software contributions of political scientists to their own field, and more generally to statistical data analysis in the the social sciences and beyond. Special emphasis is given to software that is written in or can cooperate...

OpenMx is free, full-featured, open source, structural equation modeling (SEM) software. OpenMx runs within the R statistical programming environment on Windows, Mac OS–X, and Linux computers. The rationale for developing OpenMx is discussed
along with the philosophy behind the user interface. The OpenMx data structures are introduced—these novel s...

Although seasonality might once have been a successful energy conservation strategy for people living in temperate regions, this physiological phenomenon may now foster accumulating annual weight gain and thereby feature in the risk profile for obesity. We tested the hypothesis that seasonality relates to BMI, and that this association is mediated...

Based partly on interviews with mem- bers of the R Core team, this paper considers the development of the R Project in the context of open-source software development and, more generally, voluntary activities. The paper de- scribes aspects of the social organization of the R Project, including the organization of the R Core team; describes the traj...

Based on recent work by Fox and Andersen (2006), this paper describes substantial extensions to the effects package for R to construct effect displays for multinomial and proportional-odds logit models. The package previously was limited to linear and generalized linear models. Effect displays are tabular and graphical representations of terms - ty...

In the era of data-centric-science, a large number of visualization tools have been created to help researchers understand increasingly rich business databases. Information visualization is a process of constructing a visual presentation of business quantitative data, especially prepared for managerial use. Interactive information visualization pro...

The Iowa Gambling Task (IGT) is the major plank of behavioral support for the Somatic Marker Hypothesis--a prominent theory of emotionally-based decision making. Despite its widespread use, some have questioned the ecological and discriminative validity of the IGT because a substantial proportion of neurologically-normal adults display a response p...

Les données des séries chronologiques sur le taux de condamnation des femmes au Canada sont étudiées par l'analyse de régression multiple pour évaluer l'hypothèse selon laquelle le taux de criminalité des femmes varie inversement par rapport à l'activité des femmes dans le rôle familial et directement avec leur participation dans des rôles en dehor...

Nous presentom ici une Ptude des changements au sein du march6 du travail en ce qui a trait la segregation sexuelle. L'article debute avec des considerations sur la nature de la segregation sexuelle, puis un survol de la recherche canadienne et amPricaine i ce sujet. Notre analyse utilise les categories ocupationnelles du recensement, que nous pair...

Les théoriciens de l'instrumentalisme affirment qu'un grand nombre des preneurs de décisions de l'état sont issus des cadres supérieurs de grandes entreprises et que ceci explique du moins en partie les politiques gouvernementales favorisant le processus de l'accumulation. D'autre part, les structuralistes soutiennent que ces liens ne sont pas esse...

The role of brain reward mechanisms, and associated personality traits, are of growing interest to researchers in the field of eating behaviours and obesity. However, the direction of causal influence has produced some debate and inconsistency in the literature. Some have argued that a reward deficiency syndrome increases the risk for obesity while...

An “effect display” is a graphical or tabular summary of a statistical model based on high-order terms in the model. Effect displays have previously been defined by Fox (1987, 2003) for generalized linear models (including linear models). Such displays are especially compelling for complicated models—for example, those including interactions or pol...

1 Starting the R Commander Once R is running, simply loading the Rcmdr package by typing the command library(Rcmdr) into the R Console starts the R Commander graphical user interface ("GUI"). To function properly under Windows, the R Commander requires the single-document interface (SDI) to R. 1 After loading the package, R Console and R Commander...

R is free, open-source, cooperatively developed software that implements the S statistical programming language and computing environment. The current capabilities of R are extensive, and it is in wide use, especially among statisticians. The sem package provides basic structural equation modeling facilities in R, including the ability to fit struc...

by John Fox. This document may be freely copied and distributed subject to the following conditions: The document may not be altered, nor may it be incorporated in whole or in part into any other work. Except with the direct written permission of the author, the document must be distributed in its entirety, including this title page.

Nonparametric regression analysis traces the dependence of a response variable on one or several predictors without specifying in advance the function that relates the predictors to the response. This article discusses several common methods of nonparametric regression, including kernel estimation, local polynomial regression, and smoothing splines...

Unlike S-PLUS, R does not incorporate a statistical graphical user interface (GUI), but it does include tools for building GUIs. Based on the tcltk package (which furnishes an interface to the Tcl/Tk GUI toolkit), the Rcmdr package provides a basic-statistics graphical user interface to R called the "R Commander." The design objectives of the R Com...

At least since Wirth, social scientists have debated the effects of urbanization on social life. More recently, research has focused on the relationship between residential density and social interaction. We argue that the effects of external density on neighbor socializing are likely to depend upon other aspects of the local neighbor hood environm...

Few studies of physical activity in patients with anorexia nervosa (AN) have included a suitable control group. Nor has such research considered the influence of parents' activity on that of their children. Our first prediction was that adolescents with AN would be significantly more active than healthy controls both prior to, and during, the progr...

Evidence suggests that eating-disordered families are overly concerned with social appearance and physical attractiveness. However, some argue that parental values are not sufficient to produce disordered eating in their offspring unless combined with certain third-factor effects of the child such as a psychological or biologic vulnerability. We te...

Three-dimensional dynamic scatterplots can reveal certain features of data that cannot be apprehended in marginal two-dimensional displays. Using graduate students as subjects, we sought to establish whether the detection of clusters and nonlinearity in 3-D plots varies by easily characterized properties of the data and the design of the display. W...

This paper describes the implementation in R of a method for tabular or graphical display of terms in a complex generalised linear model. By complex, I mean a model that contains terms related by marginality or hierarchy, such as polynomial terms, or main effects and interactions. I call these tables or graphs effect displays. Effect displays are c...

Bootstrapping is a general approach to statistical inference based on building a sampling distribution for a statistic by resampling from the data at hand. The term ‘bootstrapping, ’ due to Efron (1979), is an allusion to the expression ‘pulling oneself up by one’s bootstraps ’ – in this case, using the sample data as a population from which repeat...

Structural equation models (SEMs), also called simultaneous equation models, are multivariate (i.e., multi- equation) regression models. Unlike the more traditional multivariate linear model, however, the response variable in one regression equation in an SEM may appear as a predictor in another equation; indeed, variables in an SEM may in‡uence on...

Survival analysis examines and models the time it takes for events to occur. The prototypical such event is death, from which the name'survival analysis'and much of its terminology derives, but the ambit of application of survival analysis is much broader. Essentially the same methods are employed in a variety of disciplines under various rubrics-f...

We report a meta-analysis of all polls published in the five-month period preceding the June 2, 1997 Canadian federal election. The perception that the Liberals were hurt by the early election call receives support from the polling data. Similarly, media claims that the Reform party ran a strong campaign are supported by a gradual increase in the p...

Three-dimensional dynamic scatterplots can reveal characteristics of data -- such as certain kinds of clustering and nonlinearity -- that are not apparent in marginal two-dimensional views of the data. The experiments reported in this paper address the detection of nonlinearity in 3D dynamic scatterplots. Employing graduate-student subjects, we des...

At least in principle, three-dimensional dynamic scatterplots can reveal certain features of data that cannot be apprehended in marginal two-dimensional displays of the data. Using graduate students in statistics as subjects, we seek to establish whether the probability and rapidity of detection of clusters in 3D plots vary by easily characterized...

Considerable research has examined the correlates and consequences of both objective and subjective ratings of physical attractiveness. Numerous studies have found, for example, that subjective physical attractiveness is inversely related to weight and diet concerns. Surprisingly, however, no research has examined the relationship between objective...

We perform a meta-analysis of 23 voter polls reported in the news media during the two-month period preceding the 1995 Quebec sovereignty referendum. Contrary to common interpretations of the course of the referendum campaign, we find that there was a smooth and general increase in support for the sovereignty option during this period, and that Luc...

We describe a system, written in Lisp-Stat, for designing and conducting experiments in the perception of three-dimensional dynamic scatterplots. The system includes tools for designing stimuli that systematically vary aspects of the display (e.g., the presence of graphical elements such as axes, regression surfaces, and residuals) and for convenie...

The purpose of this study was to investigate the simultaneous influence of age, gender, physical-activity level, and personality factors on the reasons that individuals participate in leisure-time physical activity. In previous research only a few studies have considered potentially influential factors besides gender. Preliminary factor analytic pr...

As has often been recognized, several characteristics of APL make it especially well suited for programming statistical applications: APL treats arrays-vectors, matrices, and higher dimensional arrays-as data structures that can be processed without reference to their elements; it contains many power primitive functions and operators for manipulati...

A relationship between feminist attitudes and body satisfaction has been suggested in the literature, but has received no empirical support. Multiple regression analyses were performed on a sample of 200 primarily Caucasian women. Results indicated that after controlling for certain known predictors of body dissatisfaction, the Physical Attractiven...

A group of adult women (n = 351) were assessed on a number of psychological, behavioral, and body composition measures including weight preoccupation and leisure-time physical activity participation. The purpose of the study was twofold. It was intended (a) to investigate whether excessive exercises have specific characteristics in common other tha...

Working in the context of the linear model y = Xβ + ε, we generalize the concept of variance inflation as a measure of collinearity to a subset of parameters in β (denoted by β1, with the associated columns of X given by X1). The essential idea underlying this generalization is to examine the impact on the precision of estimation—in particular, the...

The functional role of exercise in the development of weight preoccupation and body dissatisfaction was examined. One hundred and twelve females, identified as regular participants in exercise programmes, completed an inventory that provided information on attitudes, behaviours and psychological characteristics related to dieting and weight-control...

To investigate age, sex, and situational differences in choice of consultant, 192 students were interviewed-24 males and 24 females at ages 8, 11, 14, and 17. All subjects were presented with 3 hypothetical problems-an impersonal problem, an interpersonal problem with a peer, and an interpersonal problem with a parent. Subjects were asked to indica...

Hope (1984); and (b) general criticism of the application of statistical models to nonexperimental data in the social sciences. The two parts of Freedman's argument are only weakly connected. I find myself largely in agreement with the specific criticisms that he levels against common practice, while at the same time disagreeing with many of the ge...

Fox (1980) describes a simplified and general method for calculating path-analytic effect decompositions. This article shows how that method may be extended to the calculation of specific indirect effects, that is, effects mediated by specific intervening variables. The method employs only the information contained in the structural coefficients of...

This article examines the following situation: A variable is measured for a sample of subjects on several occasions before and after an experimental intervention thought to affect the variable. This situation had broad application in the social sciences and elsewhere. The article introduces MANOVA tests for changes in level and slope that arise as...

Examining 1971 Canadian Census data on detailed occupations we find evidence that women's labor-force participation is negatively related to wages -- both for men and, to a lesser extent, for women. The negative effect of women's labor-force participation on wages occurs within occupational sectors and controlling for the educational level of occup...

One of the great virtues of structural equation models is that they permit the quantification of causal and noncausal sources of statistical relationship. The present article discusses efficient matrix methods of computation for effect decomposition and extends these methods to models with unstandardized variables and to nonrecursive models. An app...

In their most recent (1977) theoretical statement dealing with status characteristics and expectation states. Berger and his associates propose a linear function for predicting numerical values on a measure of relative power and prestige in task-oriented groups. In this paper we briefly outline the theory and the manner in which this function is em...

Four-person groups played repeated trials of an n-Person Prisoner's Dilemma game under two information conditions. In one condition, each player's choice was made known to the others following each trial. In the other information condition each subject was informed only of the number of others who cooperated or competed on each trial. The first con...

The effects of group size and cooperativeness of others upon an individual's propensity to cooperate is studied using the Take-Some format of the Prisoner's Dilemma game. By holding utility effects constant, this format permits comparisons to be made between the performances of subjects in different size groups. A 2 × 2 experimental design was empl...

To compare the effects of group size on cooperation, we introduce a class of simple multiperson games. These games can be regarded as n-person generalizations of the two-person prisoner's dilemma with expected value payoffs. In order to ensure that identical expected-value monetary alternatives are available to players in different-sized groups, we...

Three different presentation formats of the same Prisoner's Dilemma game were investigated. Fifteen pairs of male subjects were assigned to each of three experimental conditions. One group played a conventional matrix representation of the game, one group played an expected value representation of the game, and one group played a simplified probabi...

Classes of two-person zero-sum games termed "equivalent games" are defined. These are games with identical value and identical optimal mixed-strategies but with different matrix entries and thus different opportunities for exploiting a nonrational opponent. An experiment was conducted to investigate the strategy-choice behavior of subjects playing...

Subjects play a 2 × 2 zero-sum game without saddlepoint against a computer program opponent; the computer program either follows its minimax mixed strategy or adopts a (pre-defined) nonrational mixed strategy. It is found that there is a significant trend in the strategy choice behavior of subjects playing against a rational opponent such that thes...

Frank Duan recently posted a question to the r-help mailing list asking how to translate numbers into words. The program described in this column is a cleaned-up and slightly enhanced version of my response to his question. I found the problem to be an interesting puzzle, and the solution uses several programming techniques that demonstrate the fle...