ASSIGNMENT
Course: Introduction to BIOSTATISTICS [BBT203]
BIOSTATISTICS [BBT230]
Topic: Different Types of Statistical Tests
1. Introduction
Statistical tests are used in hypothesis testing to determine whether there is a
significant difference or relationship between variables. The choice of statistical
test depends on the type of variables, number of groups, and whether the data
are paired or independent. Different tests are used for continuous outcomes and
categorical outcomes.
2. Tests for Continuous Outcomes
2.1 Independent t-test
Purpose:
Used to compare the means of two independent groups when the data are
normally distributed.
Example:
Comparing the average blood pressure between male and female patients.
Hypothesis Example:
H₀: Mean blood pressure of males = Mean blood pressure of females
H₁: Mean blood pressure of males ≠ Mean blood pressure of females
2.2 Wilcoxon Rank-Sum Test (Mann–Whitney Test)
Purpose:
Used when comparing two independent groups but the data are not normally
distributed.
Example:
Comparing pain scores between two treatment groups when the data are skewed.
2.3 Paired t-test
Purpose:
Used to compare two related or paired measurements when the data are
normally distributed.
Example:
Comparing patient weight before and after a diet program.
Hypothesis Example:
H₀: Mean difference between before and after = 0
H₁: Mean difference ≠ 0
2.4 Wilcoxon Signed-Rank Test
Purpose:
Used for paired data when the distribution is not normal.
Example:
Comparing pain levels of the same patients before and after medication.
2.5 One-Way ANOVA
Purpose:
Used to compare the means of more than two independent groups when the data
are normally distributed.
Example:
Comparing average cholesterol levels among three different diet groups.
2.6 Kruskal–Wallis Test
Purpose:
Non-parametric alternative to ANOVA when data are not normally distributed.
Example:
Comparing median recovery time among three different treatments.
2.7 Repeated Measures ANOVA
Purpose:
Used when multiple measurements are taken from the same subjects over time.
Example:
Measuring blood pressure of patients at 3 different time points after treatment.
2.8 Friedman Test
Purpose:
Non-parametric alternative to repeated measures ANOVA.
Example:
Comparing pain levels of patients at multiple time points when data are not
normally distributed.
2.9 Pearson Correlation
Purpose:
Used to measure the relationship between two continuous variables when data
are normally distributed.
Example:
Relationship between height and weight.
2.10 Spearman Correlation
Purpose:
Used to measure the relationship between two variables when data are not
normally distributed.
Example:
Relationship between study time and exam ranking.
3. Tests for Categorical Outcomes
3.1 Chi-Square Test
Purpose:
Used to test the association between two categorical variables.
Example:
Association between smoking status and lung disease.
3.2 Fisher’s Exact Test
Purpose:
Used instead of Chi-square when sample size is small.
Example:
Testing relationship between drug use and recovery in a small sample study.
3.3 McNemar’s Test
Purpose:
Used for paired categorical data.
Example:
Comparing patient diagnosis before and after treatment.
3.4 Kappa Statistic
Purpose:
Measures agreement between two observers or raters.
Example:
Two doctors diagnosing the same disease from X-rays.
4. Conclusion
Different statistical tests are used depending on the research design and type of
data. Continuous outcomes typically use tests such as t-tests, ANOVA, or
correlation tests, while categorical outcomes use Chi-square, Fisher’s exact test,
McNemar’s test, and Kappa statistics. Selecting the correct statistical test is
important for accurate hypothesis testing and reliable research results.