# Author

Distinguished Researcher in Computational Statistics

Rick Wicklin, PhD, is a distinguished researcher in computational statistics at SAS and is a principal developer of SAS/IML software. His areas of expertise include computational statistics, simulation, statistical graphics, and modern methods in statistical data analysis. Rick is author of the books Statistical Programming with SAS/IML Software and Simulating Data with SAS.

0
Bivariate normal probability in SAS: Rectangular regions

This article shows how to use SAS to compute the probabilities for two correlated normal variables. Specifically, this article shows how to compute the probabilities for rectangular regions in the plane. A second article discusses the computation over infinite regions such as quadrants. If (X,Y) are random variables that are

0
Data visualization tip: Plot rates, not counts

Plot rates, not counts. This maxim is often stated by data visualization experts, but often ignored by practitioners. You might also hear the related phrases "plot proportions" or "plot percentages," which mean the same thing but expresses the idea alliteratively. An example in a previous article about avoiding alphabetical ordering

0
On resizing an array when an index is out of bounds

Converting a program from one language to another can be a challenge. Even if the languages share many features, there is often syntax that is valid in one language that is not valid in another. Recently, a SAS programmer was converting a program from R to SAS IML. He reached

0
Tip: Avoid alphabetical order for a categorical axis in a graph

Howard Wainer, who used to write the "Visual Revelations" column in Chance magazine, often reminded his readers that "we are almost never interested in seeing Alabama first" (2005, Graphic Discovery, p. 72). His comment is a reminder that when we plot data for a large number of categories (states, countries,

0
Standard errors for maximum likelihood estimation

In several previous articles, I've shown how to use SAS to fit models to data by using maximum likelihood estimation (MLE). However, I have not previously shown how to obtain standard errors for the estimates. This article combines two previous articles to show how to obtain MLE estimates and the

0
The distribution of the sample median for normal data

A previous article shows how to use Monte Carlo simulation to approximate the sampling distribution of the sample mean and sample median. When x ~ N(0,1) are normal data, the sample mean is also normal, and there are simple formulas for the expected value and the standard error of the

0
The distribution of the sample median

An elementary course in statistics often includes a discussion of the sampling distribution of a statistic. The canonical example is the sampling distribution of the sample mean. For samples of size n that are drawn from a normally distribution (X ~ N(μ, σ)), the sample mean is normally distributed as

0
Quantiles of the generalized birthday problem

A previous article discusses the birthday problem and its generalizations. The classic birthday problem asks, "In a room that contains N people, what is the probability that two or more people share a birthday?" The probability is much higher than you might think. For example, in a room that contains

0
The generalized birthday problem

The birthday-matching problem (also called the birthday paradox or simply the birthday problem), is a classic problem in probability. Simply stated, the birthday-matching problem asks, "If there are N people in a room, what is the chance that two of them have the same birthday?" The problem is sometimes called

0
On the evaluation of the function exp(x) - 1

Recently I wrote about numerical analysis problem: the accurate computation of log(1+x) when x is close to 0. A naive computation of log(1+x) loses accuracy if you call the LOG function, which is why the SAS language provides the built-in LOG1PX for this computation. In addition, I showed that you

0
Approximate functions by using Taylor series and rational functions

SAS supports a special function for the accurate evaluation of log(1+x) when x is near 0. The LOG1PX function is useful because a naive computation of log(1+x) loses accuracy when x is near 0. This article demonstrates two general approximation techniques that are often used in numerical analysis: the Taylor

0
Functions for continuous probability distributions in SAS

The documentation for Python's SciPy package provides a table that concisely summarizes functions that are associated with continuous probability distributions. This article provides a similar table for SAS functions. For more information on the CDF, PDF, quantile, and random-variate functions, see "Four essential functions for statistical programmers." SAS functions for

0
Use design matrices to analyze subgroups in SAS IML

A previous article shows ways to perform efficient BY-group processing in the SAS IML language. BY-group processing is a SAS-ism for what other languages call group processing or subgroup processing. The main idea is that the data set contains several discrete variables such as sex, race, education level, and so

0
On computing Kendall's tau statistic in SAS

One thing I have learned about rank-based statistics over the years is "Be careful of tied values!" On multiple occasions, I have been asked, "Why doesn't the SAS result for [NAME] statistic agree with my hand calculation?" The answer is sometimes because of the way that tied values are handled.

Programming Tips
0
Define or extract the diagonals of a matrix

Many useful matrices in applied math and statistics have a banded structure. Examples include diagonal matrices, tridiagonal matrices, banded matrices, and Toeplitz matrices. An example of an unsymmetric Toeplitz matrix is shown to the right. Notice that the matrix is constant along each diagonal, including sub- and superdiagonals. Recently, I

0
Avoid domain errors by using Taylor series

The other day I was trying to numerically integrate the function f(x) = sin(x)/x on the domain [0,∞). The graph of this function is shown to the right. In SAS, you can use the QUAD subroutine in SAS IML software to perform numerical integration. Some numerical integrators have difficulty computing

0
Use PROC SGPLOT to embed a graph inside another graph

Did you know that you can embed one graph inside another by using PROC SGPLOT in SAS? A typical example is shown to the right. The large graph shows kernel density estimates for the distribution of the Cholesterol variable among male and female patients in a heart study. The small

0
The SG annotation macros in SAS

I don't often use the SG annotation facility in SAS for adding annotations to statistical graphics, but when I do, I enjoy the convenience of the SG annotation macros. I can never remember the details of the SG annotation commands, but I know that the SG annotation macros will create

0
On the performance of BY-group processing in SAS IML

Many SAS procedures support a BY statement that enables you to perform an analysis for each unique value of a BY-group variable. The SAS IML language does not support a BY statement, but you can program a loop that iterates over all BY groups. You can emulate BY-group processing by

0
Model data from published summary statistics

There are many ways to model a set of raw data by using a continuous probability distribution. It can be challenging, however, to choose the distribution that best models the data. Are the data normal? Lognormal? Is there a theoretical reason to prefer one distribution over another? The SAS has

0
Simulate the use of personal checks in the US

Does anyone write paper checks anymore? According to researchers at the Federal Reserve Bank of Atlanta (Greene, et al., 2020), the use of paper checks has declined 63% among US consumers since the year 2000. The researchers surveyed more than 3,000 consumers in 2017-2018 and discovered that only 7% of

0
Generate random uniform points in an ellipse

I have previously written about how to efficiently generate points uniformly at random inside a sphere (often called a ball by mathematicians). The method uses a mathematical fact from multivariate statistics: If X is drawn from the uncorrelated multivariate normal distribution in dimensiond, then S = r*X / ||X|| has

0
Bootstrap predicted means by using PROC GLMSELECT

A previous article shows how to use the MODELAVERAGE statement in PROC GLMSELECT in SAS to perform a basic bootstrap analysis of the regression coefficients and fit statistics. A colleague asked whether PROC GLMSELECT can construct bootstrap confidence intervals for the predicted mean in a regression model, as described in

0
A simple way to bootstrap linear regression models in SAS

I've written many articles about bootstrapping in SAS, including several about bootstrapping in regression models. Many of the articles use a very general bootstrap method that can bootstrap almost any statistic that SAS can compute. The method uses PROC SURVEYSELECT to generate B bootstrap samples from the data, uses the

0
Add Unicode symbols and format text labels in SAS

It has been more than a decade since SAS 9.3 changed the default ODS destination from the old LISTING destination to more modern destinations such as HTML. One of the advantages of modern output destinations is support for Unicode symbols, superscripts, subscripts, and for formatting text by using boldface, italics,

0
Plot confidence intervals for a regression model in SAS

A SAS programmer wanted to use PROC SGPLOT in SAS to visualize a regression model. The programmer wanted to visualize confidence limits for the predicted mean at certain values of the explanatory variable. This article shows two options for adding confidence limits to a scatter plot. You can use a

0
Construct an envelope function for the acceptance-rejection method

The acceptance-rejection method (sometimes called rejection sampling) is a method that enables you to generate a random sample from an arbitrary distribution by using only the probability density function (PDF). This is in contrast to the inverse CDF method, which uses the cumulative distribution function (CDF) to generate a random

1 2 3 50