pt 2: Normal Distribution and Percentiles
Chapter 6: Normal Distribution - Review and Examples
Review of Formulas:
- The formula sheet contains expected value formulas for binomial distributions and formulas from Chapter 6.
- Key formula: Transformation formula: , where:
- = observation
- = mean
- = standard deviation
Normal Distribution Table:
- Use either the table or calculator for z-values.
- Calculators have a built-in z-function.
Probability Calculations:
- Problems involve finding P(X < value).
- Given: X has a normal distribution, with and .
Transformation:
- Apply the transformation formula: .
- Convert X to Z, then use the calculator.
Area Representation:
- Plot the mean in the middle.
- The given value (e.g., 18.6) is located relative to the mean.
- The inequality sign indicates the area of interest.
Complement Rule:
- If calculating probabilities to the right, use the complement rule: 1 - P(X < value).
- Transform to Z, then calculate.
Conditional Probability Example:
Given download time is more than 17.3 seconds, find the probability it's less than 19.2 seconds.
Apply the conditional probability rule:
Where:
- : X < 19.2
- : X > 17.3
Overlapping Events:
- Determine the overlap of events.
- If no overlap, events are mutually exclusive.
- If overlap exists, calculate the intersection.
Calculation:
- P(17.3 < X < 19.2) = P(X < 19.2) - P(X < 17.3)
- Transform X values to Z values using .
- Use the calculator to find probabilities.
Calculator Usage:
- Input Z-values into the calculator to find probabilities.
- Example: For z = 0.24, probability = 0.5948.
Table Usage:
- The table contains negative and positive values.
- Find the Z-value in the table to read off the probability.
Marginal and Joint Probabilities:
- Reference to marginal and joint probabilities is made.
Important Note:
- Calculating probabilities where results in zero because there is no area.
Expected Values and Variances:
- For normal distribution, mean and standard deviation are typically given.
Importance and Applications
*P-values
* These concepts are important for p-values, hypothesis tests, and confidence intervals in later courses.
*Regression
* Normal distribution underlies regression models.
Percentiles
Percentiles are the inverse of probability calculations.
Formula:
- To calculate percentiles, use:
- Where represents a given probability.
Alpha:
- corresponds to the median.
- corresponds to the first quartile.
- corresponds to the third quartile.
Percentile Calculation:
- Given a probability, find the corresponding X value.
Examples of Percentile Calculations
Definitions:
- Percentile: General term.
- Quartiles: Specific percentiles (25th, 50th, 75th).
- (80th percentile),
Example 1: Finding the 80th Percentile of Download Time.
Given: X = download time, normally distributed with and .
The 80th percentile divides the distribution such that 80% is to the left.
Calculation:
- Find using a calculator (inverse normal function).
Interpretation: 80% of downloads are less than 22.208 seconds.
Example 2: Finding X Value for a Given Probability.
Find X such that 5% of download times are less than X.
Percentile: .
Calculation:
- Find using a calculator (inverse normal function).
5% of download times will be less than 9.7755 seconds.
Example 3: Quartiles.
Quartiles divide data into four parts.
Calculations:
- (25th percentile): , The value for ,
- (50th percentile/median): It's equal to the mean which is 18.
- (75th percentile): , The value for ,
Interquartile range
* Calculate the interquartile range (IQR):
* It means that 50% of the download time will be between 14.63 and 21.37 seconds.
Example 4: Symmetry and Probability.
Find A and B such that probability between A and B is 90%, symmetrically around the mean.
Setup:
P(A < X < B) = 0.90, is set symmetrically around .
The area to the left of A is 0.05, and to the right of B is 0.05.
A is the 5th percentile, and B is the 95th percentile.
Formula
- , and