Visualizing Categorical Data: distplot
$Version: 1.3 (10 Jan 2001)
Michael Friendly
York University
Plots for discrete distributions
The DISTPLOT macro constructs plots of a discrete distribution designed to
diagnose whether the data follows one of the standard distributions: the
Poisson, Binomial, Negative Binomial, Geometric, or Log Series, specified
by the DIST= parameter. The usual (PLOT=DIST
) plot is constructed so that the points lie along a straight line when the
data follows that distribution. An influence plot (PLOT=INFL
) shows the influence of each observation on the choice of the distribution
parameter(s).
The DISTPLOT macro is called with keyword parameters. You must specify the
distribution to be fit (DIST=). and the
COUNT= and FREQ= variables. The arguments may be listed within parentheses in any order,
separated by commas. For example:
%distplot(data=queues, count=women, freq=queues, dist=binomial,
parm=0.435);
 DATA=

The name of the input data set [Default:
DATA=_LAST_
]
 COUNT=

Basic count variable
 FREQ=

Number of occurrences of count
 LABEL=

Horizontal (count) label
 DIST=

Name of distribution, one of
POISSON
, BINOMIAL
,
GEOMETRIC
, or LOGSERIES
,
 PARM=

Trial value of the distribution
parameter(s)
to level the
plot. For the Binomial distribution, PARM=p, the binomial probability of
success; for the Poisson, PARM=lambda, the Poisson mean
 Z=

Multiplier for error bars in the
PLOT=DIST
plot. [Default: Z=1.96
]
 PLOT=

What to plot: DIST and/or INFL [Default:
PLOT=DIST
]
 HTEXT=

Height of text labels in the plots [Default:
HTEXT=1.4
]
 OUT=

The name of the output data set [Default:
OUT=DISTPLOT
]
 NAME=

Name of the graphics catalog entry [Default:
NAME=DISTPLT
]
Example
%include vcd(distplot); * or include in an autocall library;
%distplot();
See also
goodfit Goodness of fit tests for discrete distributions
ordplot Diagnose form of discrete frequency distribution
poisplot Poissonness plot for discrete distributions
rootgram Hanging rootograms for discrete distributions