The sign test

What type of data is needed for the Sign Test?

Nominal data (categorical, e.g., +/–).

What kind of design must the Sign Test be used with?

Repeated measures or matched pairs (related data).

Is the Sign Test used for a difference or a correlation?

A difference.

What is the first step of the sign test?

convert the data to nominal data. You do this by subtracting score A from score B. If the answer is negative, we simply record a minus sign. If the answer is positive, we record a plus sign

What is the second step of the sign test?

calculate the value of S (the calculated value). We work this out by adding up all of the pluses and minuses. We then take the less frequent sign and call this S

What is the third step of the sign test?

locate the critical value and compare it to the calculated value

What is S in the Sign Test?

The smaller of the number of + or – signs.

How do you handle ties (no change) in the Sign Test?

Exclude them from the total N.

What does it mean if the result is significant?

There is a real difference → reject null hypothesis.

What does it mean if the result is not significant?

Insufficient evidence → fail to reject null hypothesis.

What should you always do before using a statistical test?

Decide on your significance level and whether it’s one- or two-tailed.

Why is the Sign Test non-parametric?

It doesn’t rely on assumptions about data distribution (uses categorical signs, not means).