In the 1960s and '70s, before nonparametric regression methods became widely available, it was common to apply a nonlinear transformation to the dependent variable before fitting a linear regression model. This is still done today, with the most common transformation being a logarithmic transformation of the dependent variable, which fits
John Tukey was an influential statistician who proposed many statistical concepts. In the 1960s and 70s, he was fundamental in the discovery and exposition of robust statistical methods, and he was an ardent proponent of exploratory data analysis (EDA). In his 1977 book, Exploratory Data Analysis, he discussed a small
On Twitter, I saw a tweet from @DataSciFact that read, "The sum of (x_i - x)^2 over a set of data points x_i is minimized when x is the sample mean." I (@RickWicklin) immediately tweeted out a reply: "And the sum of |x_i - x| is minimized by the sample
In categorical data analysis, it is common to analyze tables of counts. For example, a researcher might gather data for 18 boys and 12 girls who apply for a summer enrichment program. The researcher might be interested in whether the proportion of boys that are admitted is different from the
In The Essential Guide to Bootstrapping in SAS, I note that there are many SAS procedures that support bootstrap estimates without requiring the analyst to write a program. I have previously written about using bootstrap options in the TTEST procedure. This article discusses the NLIN procedure, which can fit nonlinear
When you have many correlated variables, principal component analysis (PCA) is a classical technique to reduce the dimensionality of the problem. The PCA finds a smaller dimensional linear subspace that explains most of the variability in the data. There are many statistical tools that help you decide how many principal
Recently, I showed how to use a heat map to visualize measurements over time for a set of patients in a longitudinal study. The visualization is sometimes called a lasagna plot because it presents an alternative to the usual spaghetti plot. A reader asked whether a similar visualization can be
What is McNemar's test? How do you run the McNemar test in SAS? Why might other statistical software report a value for McNemar's test that is different from the SAS value? SAS supports an exact version of the McNemar test, but when should you use it? This article answers these
Longitudinal data are measurements for a set of subjects at multiple points in time. Also called "panel data" or "repeated measures data," this kind of data is common in clinical trials in which patients are tracked over time. Recently, a SAS programmer asked how to visualize missing values in a
This article implements Passing-Bablok regression in SAS. Passing-Bablok regression is a one-variable regression technique that is used to compare measurements from different instruments or medical devices. The measurements of the two variables (X and Y) are both measured with errors. Consequently, you cannot use ordinary linear regression, which assumes that
Sometimes it is useful to know the extreme values in data. You might need to know the Top 5 or the Top 10 smallest data values. Or, the Top 5 or Top 10 largest data values. There are many ways to do this in SAS, but this article shows examples
How can you estimate percentiles in SAS Viya? This article shows how to call the percentile action from PROC CAS to estimate percentiles of variables in a CAS data table. Percentiles and quantiles are essentially the same (the pth quantile is the 100*pth percentile for p in [0, 1]), so
Last year, I wrote almost 100 posts for The DO Loop blog. My most popular articles were about data visualization, statistics and data analysis, and simulation and bootstrapping. If you missed any of these gems when they were first published, here are some of the most popular articles from 2021:
Did you know that the loess regression algorithm is not well-defined when you have repeated values among the explanatory variables, and you request a very small smoothing parameter? This is because loess regression at the point x0 is based on using the k nearest neighbors to x0. If x0 has
I was recently asked how to create a frequency polygon in SAS. A frequency polygon is an alternative to a histogram that shows similar information about the distribution of univariate data. It is the piecewise linear curve formed by connecting the midpoints of the tops of the bins. The graph
A previous article discusses how to use SAS regression procedures to fit a two-parameter Weibull distribution in SAS. The article shows how to convert the regression output into the more familiar scale and shape parameters for the Weibull probability distribution, which are fit by using PROC UNIVARIATE. Although PROC UNIVARIATE
It can be frustrating when the same probability distribution has two different parameterizations, but such is the life of a statistical programmer. I previously wrote an article about the gamma distribution, which has two common parameterizations: one that uses a scale parameter (β) and another that uses a rate parameter
This article shows how to create a "sliced survival plot" for proportional-hazards models that are created by using PROC PHREG in SAS. Graphing the result of a statistical regression model is a valuable way to communicate the predictions of the model. Many SAS procedures use ODS graphics to produce graphs
A previous article discusses the geometry of weighted averages and shows how choosing different weights can lead to different rankings of the subjects. As an example, I showed how college programs might rank applicants by using a weighted average of factors such as test scores. "The best" applicant is determined
People love rankings. You've probably seen articles about the best places to live, the best colleges to attend, the best pizza to order, and so on. Each of these is an example of a ranking that is based on multiple characteristics. For example, a list of the best places to
Following on from my introductory blog series, Data Science in the Wild, we’re going to start delving into how you can scale up and industrialise your Analytics with SAS Viya. In future blogs we will look at how you can augment your R & Python code to leverage SAS Viya
A recent article about how to estimate a two-dimensional distribution function in SAS inspired me to think about a related computation: a 2-D cumulative sum. Suppose you have numbers in a matrix, X. A 2-D cumulative sum is a second matrix, C, such that the C[p,q] gives the sum of
Many nonparametric statistical methods use the ranks of observations to compute distribution-free statistics. In SAS, two procedures that use ranks are PROC NPAR1WAY and PROC CORR. Whereas the SPEARMAN option in PROC CORR (which computes rank correlation) uses only the "raw" tied ranks, PROC NPAR1WAY uses transformations of the ranks,
For many univariate statistics (mean, median, standard deviation, etc.), the order of the data is unimportant. If you sort univariate data, the mean and standard deviation do not change. However, you cannot sort an individual variable (independently) if you want to preserve its relationship with other variables. This statement is
It is well known that classical estimates of location and scale (for example, the mean and standard deviation) are influenced by outliers. In the 1960s, '70s, and '80s, researchers such as Tukey, Huber, Hampel, and Rousseeuw advocated analyzing data by using robust statistical estimates such as the median and the
When data contain outliers, medians estimate the center of the data better than means do. In general, robust estimates of location and sale are preferred over classical moment-based estimates when the data contain outliers or are from a heavy-tailed distribution. Thus, instead of using the mean and standard deviation of
You can standardize a numerical variable by subtracting a location parameter from each observation and then dividing by a scale parameter. Often, the parameters depend on the data that you are standardizing. For example, the most common way to standardize a variable is to subtract the sample mean and divide
A previous article discusses the definition of the Hoeffding D statistic and how to compute it in SAS. The letter D stands for "dependence." Unlike the Pearson correlation, which measures linear relationships, the Hoeffding D statistic tests whether two random variables are independent. Dependent variables have a Hoeffding D statistic
There are many statistics that measure whether two continuous random variables are independent or whether they are related to each other in some way. The most well-known statistic is Pearson's correlation, which is a parametric measure of the linear relationship between two variables. A related measure is Spearman's rank correlation,
Ranking is a fundamental concept in statistics. Ranks of univariate data are used by statisticians to estimate statistics such as percentiles (quantiles) and empirical distributions. A more advanced use is to compute various rank-based measures of correlation or association between pairs of variables. For example, ranks are used to compute
