SPEAKER: |
|
TITLE: |
"Perception of 3D Graphs" |
DAY: |
Wednesday, February 24, 1999 |
TIME: |
4:30 p.m. [No coffee this week] - Note change of time! |
PLACE: |
KTH-103 - Note change of room! |
Three-dimensional scatterplots have been promoted as geometrical tools for understanding statistical concepts (e.g., by Monette, 1990), and as tools for analyzing data (e.g., Cook and Weisberg, 1989; Huber, 1987). At least in principle, three-dimensional scatterplots can reveal certain features of data, such as interaction, that cannot be apprehended in two-dimensional displays. Nevertheless, it is our experience that three-dimensional scatterplots can be difficult to decode, and it is not at all obvious that their theoretical promise is realized in their application. Originally the province of experimental graphical systems (such as those described by Becker, Cleveland, and Wilks, 1990, and Monette, 1990), three-dimensional dynamic scatterplots are now found in some standard and readily available statistical software packages (such as SAS/Insight), and in widely used statistical programming environments (such as S and Lisp-Stat).
There is a substantial experimental literature (much of it reviewed in Spence and Lewandowsky, 1990), dealing with the perception of one and two-variable displays. There is, however, to our knowledge, no analogous body of work dealing with the specific design decisions involved in constructing three-dimensional statistical displays.
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 conveniently generating datasets that can incorporate features such as nonlinearity and outliers. The system is designed to present stimuli to subjects automatically, to interact with the subjects, and to collect information about their responses and response-latencies.
The talk will describe joint work with Robert Stine, Department of Statistics, University of Pennsylvania.
|
John Fox is Professor of Sociology at McMaster University. He was previously Professor of Sociology and of Mathematics and Statistics at York University in Toronto, where he also directed the Statistical Consulting Service at the Institute for Social Research. Professor Fox earned a Ph.D. in Sociology from the University of Michigan in 1972. He has taught many workshops on statistical topics at such places as the summer program of the Inter-University Consortium for Political and Social Research and the annual meetings of the American Sociological Association. He has a long-standing interest in statistical graphics. John Fox's most recent book, Applied Regression Analysis, Linear Models, and Related Methods, was published by Sage in 1997.
|
The following references have been suggested by Dr. Fox to be used as supporting material for his talk. The references Becker, Cleveland and Wilks (1987), Cook and Weisberg (1989), and Fox and Stine (1998) have been placed on reserve at Thode Library (STATS 770: Statistics Seminar).
[1] Becker, R.A., W.S. Cleveland, and A.R. Wilks (1987), "Dynamic Graphics for Data Analysis (with discussion)," STATISTICAL SCIENCE 2, pp. 355--395.
[2] Cook, R. D., and S. Weisberg (1989), "Regression Diagnostics With Dynamic Graphics (with discussion)." TECHNOMETRICS 31, pp. 277--311.
[3] Fox, J. D., and R. A. Stine (1998), "A Lisp-Stat-Based System FOR Experimentation in 3D Statistical Graphics," Paper read at the Conference on the Interface, Minneapolis MN.
[4] Huber, P.J. (1987), "Experiences With Three-Dimensional Scatterplots," JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION 82, pp. 448--453.
[5] Monette, G. (1990), "Geometry of Multiple Regression and Interactive 3-D Graphics," pp. 209--256 in J. Fox and J. S. Long, eds. MODERN METHODS OF DATA ANALYSIS, Newbury Park CA: Sage.
[6] Spence, I., and S. Lewandowsky, (1990), "Graphical Perception," pp. 13--57 in J. Fox and J. S. Long, eds. MODERN METHODS OF DATA ANALYSIS, Newbury Park CA: Sage.