2x2Tafel

2x2Tafel, usually called a 2x2 contingency table, is a two-by-two matrix used in statistics to summarize the relationship between two binary variables. It contains four cells with counts a, b, c, and d. Typically, one variable is arranged along the rows (e.g., disease: present vs absent) and the other along the columns (e.g., test result: positive vs negative).

From these counts, researchers derive measures of association and accuracy. Sensitivity a/(a+c) and specificity d/(b+d) describe test performance. Positive predictive value a/(a+b) and negative predictive value d/(c+d) describe the probability of actual states given a test result. The risk ratio (a/(a+b)) / (c/(c+d)) and the odds ratio ad/bc quantify the strength of association. Overall accuracy is (a+d)/(a+b+c+d).

Statistical tests commonly used with a 2x2 table include the chi-squared test for independence, Fisher's exact test for small samples, and McNemar's test for paired data. The table serves as a basis for logistic regression and other modeling techniques that handle binary outcomes.

Applications include epidemiology and clinical research for evaluating diagnostic tests, treatment effects in cohort studies, and quality-control assessments. Limitations include its restriction to binary variables, potential distortion from small expected cell counts, and sensitivity to table orientation in interpretation.

The term 2x2-Tafel is used in German-language literature to denote such a table; in English-language contexts it is usually called a 2x2 table or a 2-by-2 contingency table.