a.Aditi / practice lab

SKILLS · QUANTITATIVE REASONING

Weighted percentages

Before calculating, ask: “A percentage of what?” The denominator tells you how much each group should count.

01 · THE IDEA

Count people before averaging percentages.

If 60% of a small group and 20% of a large group succeed, the overall percentage is closer to 20%. More people sit behind that rate.

Pooled percentage = total successes ÷ total people × 100

For separate, nonoverlapping groups measuring the same outcome: add the successes, add the denominators, then divide. If only rates are given, first use successes = decimal rate × group size.

MOVE THE GROUP SIZES

Same rates. Different combined result.

Group A stays at 60% success. Group B stays at 20%. Move people between the groups while keeping 100 people in total.

A: 12 successes / 20 peopleB: 16 successes / 80 people

Actual pooled rate28%(12 + 16) / 100
Equal-group average40%(60% + 20%) / 2

The pooled rate is closer to 20% because group B has more people. The equal-group average stays at 40% because its weights stay 50:50.

“Overall,” “pooled,” or “combined”

Use the actual group sizes. Each person, cell, or other denominator unit counts equally.

(n₁ × p₁ + n₂ × p₂) / (n₁ + n₂)

Here p is a decimal proportion. Multiply the result by 100 for a percentage.

“Standardized” or “equal weight”

Use the reference weights the question specifies. Those may differ from the observed group sizes.

w₁ × p₁ + w₂ × p₂

Weights sum to 1. Equal weights for two groups means ½ and ½; standardization does not always use equal weights.

A simple average is valid when the denominators are equal, or when the question explicitly asks for equal weight per group. It can also happen to match when rates are identical; that coincidence does not make it a general pooling rule.

02 · WORKED EXAMPLES

Same arithmetic, different questions

1. Pool the actual people

A task is completed by 60% of group A’s 20 participants and 20% of group B’s 80 participants. What percentage completes it overall?

  1. A contributes 0.60 × 20 = 12 completions.
  2. B contributes 0.20 × 80 = 16 completions.
  3. Combine: (12 + 16) / (20 + 80) = 28%.

Why not 40%? (60% + 20%) / 2 gives the small group the same influence as the large group. The actual weights are 20% and 80%.

2. Now the question changes the weights

Using those same rates, what would completion be if the two groups were represented equally?

  1. Use reference weights of 0.50 and 0.50.
  2. 0.50 × 60% + 0.50 × 20% = 40%.

This describes a population with equal group representation. It does not replace the 28% observed pooled rate.

If the reference population were 75% A and 25% B, the standardized rate would instead be 0.75 × 60% + 0.25 × 20% = 50%.

3. Read the comparison within each subgroup

Original practice data: task success
Task difficultyGroup X: successes / totalGroup Y: successes / total
Easier8 / 10 (80%)63 / 90 (70%)
Harder18 / 90 (20%)1 / 10 (10%)

X is 10 percentage points higher within each difficulty. Yet its pooled rate is (8 + 18) / 100 = 26%, while Y’s is (63 + 1) / 100 = 64%.

Why the reversal? X has mostly harder tasks; Y has mostly easier tasks. The group composition changes the pooled comparison. This is a form of Simpson’s paradox.

Give both difficulties equal weight and X is (80% + 20%) / 2 = 50%; Y is (70% + 10%) / 2 = 40%. That comparison answers an equal-composition question. Neither comparison by itself proves a causal effect.

03 · TRY IT IN DIFFERENT FORMS

Eight ways this can be asked

0 of 8 checked

For each question, name the denominator and write the setup before calculating. Use a hint if needed, then choose an answer to see the explanation.

Untimed learning practice. These responses stay on this page and reset when you leave or reload; they do not change your test results.

QUESTION 1 / 8 · P/S

Counts in a table

Two nonoverlapping groups complete the same memory task. What percentage of all participants recalled the target word?

Question data
GroupRecalled the wordParticipants
A1830
B1470
Choose one answer
1 of 8

TAKE INTO YOUR NEXT PASSAGE

A four-step setup check

  1. Name the whole. All participants? Responders only? Total cells? Final volume?
  2. Choose the weights. Actual sizes for pooling; specified reference weights for standardization.
  3. Build the numerator. Convert rates to counts or amounts when needed, then add.
  4. Check the range. With positive weights, a weighted mean lies between its component values and moves toward the more heavily weighted one.

Explain your setup in one sentence: “I’m dividing ___ by ___ because the question asks for ___.”

Concept references

All worked examples and practice questions here are original learning exercises. For the underlying definitions, see NIST’s weighted mean formula and CDC’s explanation of standardization using reference weights.