The sign test is used for hypothesis testing by concerning the value of the median for one population, and comparing the median differences for two variables for two dependent samples.

Assumptions of the sign test.

  • Random sampling
  • Independent measurements
  • Data should be measurement, counts or scores.

Sign test for one population in Excel

Step 1: Put your two example data sets into a table to prepare for the test, set the null and alternative hypothesis

Step 2: Prepare a sign column calculated the sign values

Step 3: calculated the number of positive signs and number of negative signs

Step 4: Calculate the number of positive signs and number of negative signs

Step 5: Calculate the smaller sign value

Step 6: Then calculated the P-value (probability value)

Step 7: According to the P-value, check whether the null hypothesis is rejected or not.

Figure1: Sign test for one sample

Sign test for two paired samples in Excel

Step 1: Calculated the sign value

Step 2: Calculate the abs value

Step 3: Calculate the sign rank value

Step 4: Calculate the positive and negative sum values

Step 5: Calculate the test-statistics (calculated value) – lowest sum value is taken as the test statistics

Step 6: find the critical value from the critical value table for given significant value

Step 7: If critical value is greater than the calculated value null hypothesis will rejected.

Figure1: Sign test for two paired sample

Sign Test in R

#FOR ONE SAMPLE SIGN TEST
install.packages("BSDA")
library(BSDA)
SIGN.test(data, md, conf.level, alternative)

#FOR TWO PAIRED SAMPLES
data = read.csv(‘‘data_file.csv’’, header = TRUE)
SIGN.test (data$A, data$B, alternative, conf.level)

Leave a Reply

Your email address will not be published.