Non-Parametric Sign Test with Example
1. What is the Sign Test?
The Sign Test is a non-parametric test used to test hypotheses about the median of a population when
normality cannot be assumed.
Uses: - Single-sample or paired data. - Test whether the median equals a hypothesized value (M0).
Idea: Count the number of observations above and below the hypothesized median.
2. Hypotheses
For a single sample:
H0 : Median = M0 vs H1 : Median M0 =
• Count + = number of observations > M0
• Count - = number of observations < M0
• Observations equal to M0 are ignored.
3. Test Statistic
• n = number of non-zero signs (excluding ties)
•S = min(number of +, number of -)
Exact p-value: Binomial distribution B(n, 0.5)
Large sample (normal approximation):
(n+ − n/2)
Z=
n/4
4. Example
Data (weights of 10 packages in kg): 4.8, 5.1, 4.9, 5.2, 5.0, 4.7, 5.3, 5.1, 4.9, 5.0
Test: Is the median weight M0 = 5 kg?
Step 1: Count signs relative to 5
1
Value Sign
4.8 –
5.1 +
4.9 –
5.2 +
5.0 0
4.7 –
5.3 +
5.1 +
4.9 –
5.0 0
• n+ = 4, n- = 4, n = 8 (excluding ties)
Step 2: Test statistic
S = min(n+ , n− ) = 4
Step 3: p-value (two-tailed)
P = 2 × P (Binomial(n = 8, p = 0.5) ≤ 4)
• Counts are equal → p > 0.05 → fail to reject H0
Conclusion: No evidence that median differs from 5 kg.
Bloom’s Taxonomy Mapping
1. Understanding: Concept of signs and median
2. Applying: Count + and – signs from data
3. Analyzing: Compute S, n, and p-value
4. Evaluating: Interpret results to accept/reject H0
