Categorizing data

Using the keyboard versus the mouse right button to enter categorical data When to request and how to complete the Missed Visit form When to request and how to complete the Missed Forms form How to complete all other forms... 

J.P. van de Geer, Multivariate Analysis of Categorical Data Applications. Sage Publications, Newbury Park, CA, 1993. 

A. Agresti, Categorical Data Analysis. Wiley, New York, 1990. 

Clean the Data If They Are Needed for Analysis 20 Categorize Data If Necessary 21 Avoid Hardcoding Data 24... 

The problem for the statistical programmer in categorizing data comes from text variables, or more specifically, free-text variables. A free-text variable is one that may contain any characters and is typically limited only in length. As an example, let s say you need to summarize the adverse events for a set of patients in a trial. The following SAS code shows the data and a quick summarization of the adverse events. 

Concomitant medications and prior medications are collected in one of two forms a list-type free-text format where the medications get coded later by data management, or a pre-categorized data format. Here is the free-text CRF format ... 

Categorical Data and Why Zero and Missing Results Differ Greatly 102 Performing Many-to-Many Comparisons/Joins 106 Using Medical Dictionaries 108 Other Tricks and Traps in Data Manipulation 112 Common Analysis Data Sets 118 Critical Variables Data Set 118 Change-from-Baseline Data Set 118 Time-to-Event Data Set 121... 

Categorical Data and Why Zero and Missing Results Differ Greatly... 

A laboratory shift table is a tabular display that can show you how a population s laboratory data change, or shift, over time. Often you want to see what happens to the patients lab values after therapeutic intervention. Did certain lab parameters drop below or above normal range Are there laboratory tests that have become of clinical concern A shift table can provide this information at a glance. Although the example that follows is focused on laboratory data, a shift table can be used to show the movement of any categorical data over time. 

PROC FREQ can be used to export frequencies of categorical data simply by specifying the OUT= option in the TABLES statement like this ... 

Obtaining Inferential Statistics from Categorical Data Analysis... 

Categorical Data Analysis Using the SAS0 System, Second Edition... 

Yoshida and Topliss compiled a dataset of 232 structurally diverse drugs (Table 16.4) and evaluated the possibility of constructing a predictive model for human oral bioavailability on categorical data [28]. The bioavailability data were classified into four categories ... 

LCA and CCK, on the other hand, appear to be strikingly dissimilar. All CCK procedures require at least one quasi-continuous indicator, and if there are none, the investigator has to create such an indicator (e.g., SSMAXCOV procedure). In contrast, LCA does not require continuous indicators and only deals with categorical data. In the case of categorical data, the patterns of interest are usually apparent, so there is no need to summarize the data with correlations. Therefore, LCA evaluates cross-tabulations and compares the number of cases across cells. This shift in representation of the data necessitates other basic changes. For example, LCA operates with proportions instead of covariances and yields tables rather than plots. These differences aside, the two approaches share a lot in common. LCA, like CCK, starts with a set of correlated indicators. It also makes the assumption of zero nuisance covariance-—in the LCA literature this is called the assumption of local independence, and it means that the indicators are presumed to be independent (i.e., uncorrelated) within latent classes. Moreover, LCA and CCK (MAXCOV in particular) use similar procedures for group assignment and both of them involve Bayes s theorem. 

On the other hand, LCA has a number of important advantages. One of them is the ability to estimate a full set of latent parameters from categorical data. SSMAXCOV can analyze categorical data, but it cannot estimate any latent parameters other than the taxon base rate. A second advantage of LCA is its ability to simultaneously evaluate the existence of multiple taxa. This feature of LCA can be invaluable in situations when more than one... 

Possible solutions modify existing diagnostic assessment instruments so criteria are assessed more dimensionally (no sxs, mild sxs, mod sxs, clinically signif sxs, severe sxs), plus assess for every diagnostic criterion use those taxometric procedures that can adequately deal with categorical data. 

Kass, G.V (1980). An exploratory technique for investigating large quantities of categorical data. Appl. Stat. 29 119-127. 

They work also for categorical data, and they can be used in the two- and multiple-group case. 

Agresti, Alan. (1996). An introduction to categorical data analysis. New York Wiley. 

The mean uses the most information from the sample, relying as it does on the actual numerical values. It is the most commonly used measure of location and therefore can be misused. For example, we have noted that it is inappropriate to use it for ordered categorical data such as 0 = absent, 1 = mild, 2 = moderate, 3 = severe since in taking an average the implicit assumption is being made that a change from absent to mild is identical to a... 

Moving up from binary is categorical data where there are more than two categories that form the basis of the measurement . The following are examples of categorical variables ... 

As we shall see later the data type to a large extent determines the class of statistical tests that we undertake. Commonly for continuous data we use the t-tests and their extensions analysis of variance and analysis of covariance. For binary, categorical and ordinal data we use the class of chi-square tests (Pearson chi-square for categorical data and the Mantel-Haenszel chi-square for ordinal data) and their extension, logistic regression. 

Consider four outcome categories labelled A, B, C and D and the comparison of two treatments in terms of the distribution across these categories. Taking the example of categorical data from Chapter 1, we might have ... 

As with binary and categorical data, is there an issue with small sample sizes Well, in fact, no, there is not. The MH test is a different kind of chi-square test and is not built around expected frequencies. As a consequence it is not affected by small expected frequencies and can be used in all cases for ordinal data. There are some pathological cases where it will break down but these should not concern us in practical settings. 

Measures such as the difference in event rates, OR, RR, RRR and NNT do not easily translate into the categorical data context. If we want to construct such measures in these cases we would collapse the outcome categories to two, the binary case, and proceed as before. In the categorical example covered earlier this could involve collapsing categories A, B and C to produce a binary outcome death/survival. 

Smith GD and Eggar M (1997) Letter to the Editor The Lancet, 350, 1182 Stokes ME, Davis CS and Koch GG (1995) Categorical Data Analysis Using the SAS System Cary, NC SAS Institute Inc. 

---