Difference of Proportions

Definition

A proportion is the fraction or percentage of a group that has a certain characteristic. The difference of proportions is the numerical difference between two such proportions, usually written as:

$$p_1 - p_2$$

where:

• $p_1$ = proportion in the first group
• $p_2$ = proportion in the second group

In statistical inference, the difference of proportions is often analyzed using a hypothesis test or a confidence interval to determine whether the two population proportions are truly different.

Example:

• Group 1: 45 out of 100 people prefer tea → $p_1 = 0.45$
• Group 2: 30 out of 100 people prefer tea → $p_2 = 0.30$

So, the difference of proportions is:

$$0.45 - 0.30 = 0.15$$

This means Group 1 has a 15 percentage point higher proportion than Group 2.

Main Content

1. Meaning of Difference of Proportions

• The difference of proportions measures how much one group’s proportion differs from another group’s proportion.
• It is used to compare categorical outcomes such as:
• pass/fail
• yes/no
• success/failure
• defective/non-defective
• It is important to distinguish between:
• proportion: the part of a whole
• difference in proportions: the gap between two proportions

Example:

If 80% of customers in store A are satisfied and 65% in store B are satisfied, the difference is:

$$0.80 - 0.65 = 0.15$$

So store A has a 15% higher satisfaction rate.

2. Sampling Distribution and Statistical Inference

• In real-world studies, proportions are usually taken from samples, not entire populations.
• Because samples vary, the observed difference may not exactly match the true population difference.
• Statistical inference helps determine whether the observed difference is likely due to:
• true population difference
• random sampling error
• A confidence interval for $p_1 - p_2$ estimates the range of plausible values for the true difference.
• A hypothesis test checks whether the difference is statistically significant.

Example:

Suppose:

• Sample 1: 52 successes out of 100 → $p_1 = 0.52$
• Sample 2: 40 successes out of 100 → $p_2 = 0.40$

Observed difference:

$$0.52 - 0.40 = 0.12$$

A statistical test may show whether this 12% difference is significant enough to conclude a real population difference.

3. Interpretation and Practical Meaning

• The difference of proportions is interpreted in percentage points, not as a percent increase.
• A difference of 0.10 means a 10 percentage point difference, not “10% more” in a relative sense.
• Interpretation must consider:
• sample size
• variability
• context of the problem
• direction of the difference

Example:

If the proportion of voters supporting candidate X is 0.48 in one region and 0.39 in another, then:

$$0.48 - 0.39 = 0.09$$

This means there is a 9 percentage point higher support in the first region.

Simple visual idea:

Group 1: ██████████ 52%
Group 2: ████████ 40%
Difference = 12 percentage points

Working / Process

1. Find the proportions in each group

• Count the number of people/items with the outcome of interest in each group.
• Divide by the total number in each group.
• Use: $$p = \frac{x}{n}$$ where $x$ is the number of successes and $n$ is the sample size.

2. Compute the difference

• Subtract the second proportion from the first: $$p_1 - p_2$$

• Interpret the sign:

  • positive result: first group has a higher proportion
  • negative result: second group has a higher proportion

3. Assess statistical significance

• If needed, use:
  • confidence interval
  • z-test for two proportions
• Compare the result with a null value of 0.
• If 0 is not in the confidence interval, the difference may be statistically significant.

Example process:

• Group 1: 75 successes out of 150 $$p_1 = \frac{75}{150} = 0.50$$

• Group 2: 54 successes out of 150 $$p_2 = \frac{54}{150} = 0.36$$

• Difference: $$0.50 - 0.36 = 0.14$$

So the first group’s proportion is 14 percentage points higher.

Advantages / Applications

• Helps compare outcomes between two populations or groups in a clear and simple way.
• Useful in many fields such as:
• public health
• market research
• education
• politics
• manufacturing
• Supports decision-making by showing whether one group performs better, worse, or similarly to another.
• Commonly used to evaluate:
• treatment effectiveness in medicine
• customer satisfaction in business
• pass rates in education
• defect rates in production
• survey responses in social research

Summary

• Difference of proportions compares two group proportions.
• It is calculated as $p_1 - p_2$.

• It is used to study whether the gap between two percentages is meaningful.

• Key point 1: It measures the gap between two proportions.

• Key point 2: It is often tested using confidence intervals or hypothesis tests.
• Key point 3: It is widely used in real-life comparisons of categorical data.
• Important terms to remember: proportion, sample proportion, population proportion, difference of proportions, confidence interval, hypothesis test, statistical significance