StatNote 38 : The sign test
Richard A. Armstrong
Abstract
Richard A. Armstrong
Abstract
The sign test is a simple non-parametric test which can be used on paired data, i.e., two related samples, matched samples, or repeated measurements on the same sample. It was developed by Wilcoxon before the more powerful and familiar ‘Wilcoxon signed-rank test’ described in a previous statnote . This statnote describes the use of the sign test with reference to two scenarios: (1) to compare the cleanliness of two hospital wards as assessed by a sample of observers and (2) to compare bacterial contamination on cloths and sponges from a domestic kitchen.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
The sign test is a simple non-parametric test which can be used on paired data, i.e., two related samples, matched samples, or repeated measurements on the same sample. It was developed by Wilcoxon before the more powerful and familiar ‘Wilcoxon signed-rank test’ described in a previous statnote . This statnote describes the use of the sign test with reference to two scenarios: (1) to compare the cleanliness of two hospital wards as assessed by a sample of observers and (2) to compare bacterial contamination on cloths and sponges from a domestic kitchen.
Key concepts: Sign test, Wilcoxon signed-rank test, Sign (mathematics), Statistics, Test (biology), Mathematics, Sample (material), Mann–Whitney U test