AP Statistics (AP Stats): Finding Proportions/Percentiles with Normalcdf and InvNorm

AP Statistics – Finding Proportions/Percentiles with Normalcdf and InvNorm

AP Statistics Study Guide: Finding Proportions/Percentiles with Normalcdf and InvNorm

What This Is

This topic focuses on using the normal distribution to find probabilities (proportions) and percentiles (cutoff values) for normally distributed data. On the AP exam, you’ll use normalcdf to find the proportion of data below, above, or between values, and invNorm to find the value corresponding to a given percentile. These skills are essential for confidence intervals, hypothesis tests, and interpreting real-world data—like determining what percentage of SAT scores fall above a certain threshold or finding the cutoff score for the top 10% of test-takers.

Key Terms & Formulas

• Normal Distribution (?, ?): A symmetric, bell-shaped distribution defined by its mean (?) and standard deviation (?).
• Standard Normal Distribution (Z): A normal distribution with-= 0 and-= 1.
• z-score: z = (x – ?) / ?-Measures how many standard deviations a value is from the mean.
• normalcdf(lower, upper, ?, ?) (TI-84): Returns the proportion of data between lower and upper in a normal distribution with mean ? and standard deviation ?.
• If no ?/? are given, assumes standard normal (?=0, ?=1).
• For "less than x," use -1E99 as lower.
• For "greater than x," use 1E99 as upper.
• invNorm(area, ?, ?) (TI-84): Returns the value x such that P(X < x) = area in a normal distribution with mean ? and standard deviation ?.
• If no ?/? are given, assumes standard normal.
• Empirical Rule (68-95-99.7): In a normal distribution:
• ~68% of data falls within 1? of ?.
• ~95% within 2?.
• ~99.7% within 3?.
• Percentile: The value below which a given percentage of data falls (e.g., the 90th percentile is the value where 90% of data is below it).
• Sampling Distribution of p?: If np-10 and n(1–p)-10, the sampling distribution of the sample proportion p? is approximately normal with ? = p and ? = ?(p(1–p)/n).

Step-by-Step / Process Flow

Finding a Proportion (Probability) with normalcdf

Example: SAT scores are normally distributed with-= 1050 and-= 200. What proportion of students score above 1300?

1. Identify the distribution: Normal with-= 1050,-= 200.
2. Determine the bounds:
3. We want P(X > 1300), so lower = 1300, upper = 1E99.
4. Check conditions (if sampling):
5. For proportions: np-10 and n(1–p)-10 (not needed here since we’re given a population).
6. Calculate using normalcdf:
7. normalcdf(1300, 1E99, 1050, 200)-0.1056 (or 10.56%).
8. Interpret in context: About 10.56% of students score above 1300.

Finding a Percentile (Cutoff) with invNorm

Example: What score separates the top 5% of SAT test-takers (? = 1050,-= 200)?

1. Identify the distribution: Normal with-= 1050,-= 200.
2. Determine the area:
3. Top 5% means P(X > x) = 0.05, so P(X < x) = 0.95.
4. Calculate using invNorm:
5. invNorm(0.95, 1050, 200)-1380.4.
6. Interpret in context: A score of 1381 (round up) separates the top 5%.

Finding a Proportion for a Sample Proportion (p?)

Example: A factory claims 5% of its lightbulbs are defective. In a sample of 400 bulbs, what’s the probability that more than 7% are defective?

1. Check conditions:
2. np = 400(0.05) = 20-10 and n(1–p) = 400(0.95) = 380-10-Normal approximation is valid.
3. Find ? and ?:
4. ? = p = 0.05
5. ? = ?(p(1–p)/n) = ?(0.05(0.95)/400)-0.0109.
6. Calculate using normalcdf:
7. P(p? > 0.07) = normalcdf(0.07, 1E99, 0.05, 0.0109)-0.0344.
8. Interpret in context: There’s a 3.44% chance that more than 7% of the sample is defective.

Common Mistakes

• Mistake: Using normalcdf for non-normal data.

• Correction: Only use normalcdf if the data is approximately normal (check with a histogram or given info). For skewed data, the normal approximation fails.

• Mistake: Forgetting to convert between "greater than" and "less than" for invNorm.

• Correction: invNorm gives the value for P(X < x). For "top 10%," use invNorm(0.90), not invNorm(0.10).

• Mistake: Mixing up normalcdf bounds (e.g., using 1E99 as lower instead of upper).

• Correction: For P(X > a), use normalcdf(a, 1E99, ?, ?). For P(X < a), use normalcdf(-1E99, a, ?, ?).

• Mistake: Ignoring the 10% condition when sampling without replacement.

• Correction: If sampling without replacement, check that the sample size n-0.10N (where N is the population size).

• Mistake: Using invNorm with the wrong mean/standard deviation.

• Correction: Always include-and-in invNorm unless working with the standard normal (Z) distribution.

AP Exam Insights

• FRQs often ask for:
• The proportion of data above/below/between values (use normalcdf).
• The cutoff value for a given percentile (use invNorm).
• Sampling distributions of p? (check np-10 and n(1–p)-10).
• Tricky distinctions:
• normalcdf vs. invNorm: normalcdf gives a probability, invNorm gives a value.
• Population vs. sample: If given a sample, you may need to calculate p? and its standard deviation.
• Calculator pitfalls:
• Forgetting to use -1E99 or 1E99 for tails.
• Mixing up the order of lower and upper in normalcdf.
• Not rounding invNorm results appropriately (e.g., SAT scores are whole numbers).

Quick Check Questions

1. (Multiple Choice)

Heights of adult men are normally distributed with-= 70 inches and-= 3 inches. What proportion of men are shorter than 65 inches? (A) 0.0475 (B) 0.0500 (C) 0.9525 (D) 0.9500

Answer: (A) 0.0475 Explanation: normalcdf(-1E99, 65, 70, 3)-0.0475.

2. (FRQ Part)

A college entrance exam has scores that are normally distributed with-= 500 and-= 100. (a) What score corresponds to the 80th percentile? (b) What proportion of students score between 450 and 600?

Answer: (a) invNorm(0.80, 500, 100)-584.16-584 (round to nearest whole number). (b) normalcdf(450, 600, 500, 100)-0.5328-53.28%.

Last-Minute Cram Sheet

1. normalcdf(lower, upper, ?, ?)-Proportion between two values.
2. invNorm(area, ?, ?)-Value for a given percentile.
3. For sampling distributions of p?: Check np-10 and n(1–p)-10.
4. Empirical Rule: 68% (1?), 95% (2?), 99.7% (3?).
5. For "less than x": normalcdf(-1E99, x, ?, ?).
6. For "greater than x": normalcdf(x, 1E99, ?, ?).
7. Always check normality before using normalcdf/invNorm.
8. For invNorm, "top 10%" = invNorm(0.90), not invNorm(0.10).
9. Round invNorm results to match the context (e.g., whole numbers for test scores).
10. Sampling without replacement? Check n-0.10N.