Continuous Probability Distributions and the Normal Curve

Course Policies and Testing Procedures

• All cell phones, smartwatches, and electronic devices equipped with Wi-Fi connectivity must be powered completely off and securely stowed away in a zippered bag prior to testing.
• Testing etiquette requires chairs to be pushed back completely under the table at the conclusion of an exam. Failure to push a chair under the table will result in the chair compliance question being marked false, incurring a penalty of $5$ points deducted from the exam score.
• Recorded tutorial videos and supplementary instructional materials are accessible within the video and resource library on Canvas.
• The instructional material on continuous probability distributions begins content for Test $3$ and is not assessed on the upcoming exam.
• The exit ticket assessment corresponds to Section $6.1$.

Descriptive Statistics Project: Lick and Flavor Data

• The descriptive statistics project requires analyzing two distinct types of data: quantitative lick count data and categorical flavor data.
• The full master dataset includes student contributions from four $10:16$ class sections, compiling a total of $134$ observational data values. Demonstrations utilize a smaller representative dataset containing $39$ values.
• Analytical requirements for quantitative lick data:
  • Calculate standard descriptive summary statistics.
  • Explicitly calculate the numerical boundaries for the lower fence and upper fence using the mathematical outlier formulas.
  • Construct a horizontal boxplot in StatCrunch via the path: Graph > Boxplot > select lick data > select horizontal orientation.
  • Outliers plot automatically as individual dots. Placing the cursor over any dot reveals its exact numerical value.
  • Identified outliers must be recorded in the designated reporting boxes. If no outliers are present, the term "none" must be written in the first box.
• Analytical requirements for categorical flavor data:
  • Construct a frequency distribution table recording each flavor category alongside its corresponding raw frequency counts, relative frequencies, cumulative frequencies, and cumulative relative frequencies.
  • Generate a Pareto chart in StatCrunch via the path: Graph > Bar Chart > With Data > select flavor > order by count descending > select value above bar.
  • Selecting the option to display values above the bars prints the exact category counts directly onto the chart, eliminating manual counting.
• Formatting requirements for submitted graphs:
  • Column variables must be properly named prior to generating figures to ensure axis labels populate automatically.
  • Every generated graph must contain a descriptive title. If a title is omitted, the graph must be edited, re-copied, and pasted back into the final submission document.

Continuous Random Variables and Probability Density Functions

• Continuous random variables represent quantitative measurements across continuous scales rather than discrete counts.
• A probability density function (PDF) is an equation used to determine probabilities for continuous random variables.
• Two fundamental mathematical properties must be satisfied by every probability density function:
  • Property $1$: The total area under the graph of the equation across all possible values of the random variable must equal $1$.
  • Property $2$: The curve must satisfy $y \ge 0$ for all defined values of the random variable. Geometrically, the graph cannot extend below the horizontal $x$-axis.
• In continuous probability models, probability is mathematically identical to area ($\text{Area} = \text{Probability}$).
• Calculating area under arbitrary non-linear continuous curves formally requires integral calculus, which is evaluated computationally by statistical software (such as StatCrunch) or via pre-calculated probability tables.

Uniform Probability Distribution

• The uniform distribution is a continuous probability distribution where every sub-interval of equal length within the distribution's range is equally likely to occur.
• The geometric graph of a uniform probability density function forms a rectangle.
• Delivery window scenario:
  • A United Parcel Service (UPS) package is scheduled to arrive between $10:00\,\text{AM}$ and $11:00\,\text{AM}$.
  • Let the continuous random variable $X$ denote the delivery arrival time measured in minutes after $10:00\,\text{AM}$.
  • The domain of the distribution is defined over the interval $[0, 60]$, where $X = 0$ corresponds to $10:00\,\text{AM}$ and $X = 60$ corresponds to $11:00\,\text{AM}$.
  • Every $1\,\text{minute}$ interval within $[0, 60]$ has the exact same probability of occurrence (for example, the arrival window from $10:15\,\text{AM}$ to $10:16\,\text{AM}$ is equally likely as the window from $10:40\,\text{AM}$ to $10:41\,\text{AM}$).
• Calculating height and function definition:
  • Total area of the rectangle must equal $1$: $\text{Area} = \text{Length} \times \text{Height} = 1$.
  • Given a total length of $60 - 0 = 60\,\text{units}$:
  • $60 \times \text{Height} = 1$
  • $\text{Height} = \frac{1}{60}$
  • The probability density function is defined as $y = \frac{1}{60}$ for $0 \le X \le 60$, and $y = 0$ for all values where $X < 0$ or $X > 60$.
• Interval probability determination:
  • To calculate the probability that the delivery occurs between $15$ and $30\,\text{minutes}$ after the hour:
  • The interval width is $30 - 15 = 15\,\text{units}$.
  • $\text{Area} = 15 \times \frac{1}{60} = \frac{15}{60} = 0.25$
  • There is a $25\%$ probability ($0.25$) that the package arrives between $15$ and $30\,\text{minutes}$ past $10:00\,\text{AM}$.

The Normal Distribution: Definition, Parameters, and Geometry

• The normal distribution is a continuous probability distribution characterized by a symmetric, bell-shaped curve.
• Symmetry and central measures:
  • The distribution has a single peak located at its exact center.
  • The center represents the balance point (mean $\mu$), the division into two equal halves (median), and the highest frequency/density point (mode).
  • In a normal distribution, the measures of central tendency are equal: $\text{Mean} = \text{Median} = \text{Mode} = \mu$.
• Parameters across probability distributions:
  • Binomial distribution: Specified completely by number of trials $n$ and success probability $p$.
  • Poisson distribution: Specified completely by average rate $\lambda$ and time interval $t$.
  • Normal distribution: Specified completely by mean $\mu$ and standard deviation $\sigma$.
• Parameter functions in the normal distribution:
  • Mean ($\mu$): Serves as the location parameter. It establishes the central placement of the curve along the horizontal $x$-axis. A distribution with $\mu = 3$ and $\sigma = 1$ has the identical shape as one with $\mu = 0$ and $\sigma = 1$, shifted $3\,\text{units}$ to the right.
  • Standard deviation ($\sigma$): Serves as the shape parameter. Because the total enclosed area under the curve is fixed at $1$, changing $\sigma$ alters spread and height. A larger $\sigma$ flattens and broadens the curve; a smaller $\sigma$ compresses the spread, forcing the peak taller and narrower.
• Inflection points:
  • Defined as the points on a curve where the curvature changes direction (transitions between concave down and concave up).
  • Inflection points occur exactly one standard deviation away from the mean: at $\mu - \sigma$ and $\mu + \sigma$.
  • For $\mu = 0$ and $\sigma = 1$, inflection points sit at $-1$ and $1$. For $\mu = 0$ and $\sigma = 2$, inflection points sit at $-2$ and $2$.
• Historical reference tables:
  • Standard normal probability tables historically appeared in physical statistics textbooks in three distinct locations: inside the front hard cover, in the appendices, and on a perforated tear-out cardboard formula insert.

Properties of the Normal Distribution and the Empirical Rule

• Core properties of normal distributions:
  • Perfect bilateral symmetry about the vertical line passing through the mean $\mu$.
  • Equality of mean, median, and mode at the center peak.
  • Inflection points situated precisely at $\mu \pm \sigma$.
  • The curve approaches, but never touches, the horizontal axis as it extends toward $-\infty$ and $+\infty$.
  • Although the theoretical domain spans from $-\infty$ to $+\infty$, probabilities beyond $4$ or $5$ standard deviations are negligible, allowing the distribution to effectively model real-world variables with finite boundaries.
• Connection to the Empirical Rule:
  • The Empirical Rule is mathematically derived from the properties of the normal distribution.
  • Approximately $68\%$ of the data falls within one standard deviation of the mean ($[\mu - \sigma, \mu + \sigma]$).
  • Approximately $95\%$ of the data falls within two standard deviations of the mean ($[\mu - 2\sigma, \mu + 2\sigma]$).
  • Approximately $97\%$ to $99.7\%$ of the data falls within three standard deviations of the mean ($[\mu - 3\sigma, \mu + 3\sigma]$).
• Verifying normality from sample data:
  • Observational data should be evaluated via relative frequency histograms and boxplots to inspect for bell-shaped symmetry.
  • Real-world empirical data is rarely perfectly symmetric; moderate adherence to a bell shape is sufficient to apply normal models.

Application and Interpretation: Giraffe Weight Example

• Model specifications:
  • Giraffe weights are approximately normally distributed with a mean of $\mu = 2200\,\text{pounds}$ and a standard deviation of $\sigma = 200\,\text{pounds}$.
• Graphical construction procedure:
  • Draw a horizontal number line representing the $x$-axis labeled with weight units.
  • Sketch a symmetric bell-shaped curve resting on the axis.
  • Label the central peak with the mean: $\mu = 2200$.
  • Identify and mark the inflection points one standard deviation from the center at $\mu - \sigma = 2200 - 200 = 2000\,\text{pounds}$ and $\mu + \sigma = 2200 + 200 = 2400\,\text{pounds}$.
  • Plot the target cutoff value $X = 2100\,\text{pounds}$ to the left of the mean $2200$, between $2000$ and $2200$.
  • Shade the targeted region under the curve to the left of $X = 2100\,\text{pounds}$.
• Area and interpretation:
  • Given that the area under the normal curve to the left of $X = 2100\,\text{pounds}$ is $0.3085$, two distinct interpretations exist:
  • Proportion interpretation: The proportion of giraffes that weigh less than $2100\,\text{pounds}$ is $0.3085$ (or $30.85\%$).
  • Probability interpretation: The probability that a randomly selected giraffe weighs less than $2100\,\text{pounds}$ is $0.3085$ (or $30.85\%$).
• Analytical strategy:
  • In continuous probability distributions, calculations are restricted to intervals rather than individual isolated points ($P(X = c) = 0$ for any exact value $c$).
  • Hand-sketching curves, marking parameters, and shading designated target areas is essential to reinforce conceptual understanding of probability areas rather than executing software steps mechanically.

Questions & Classroom Discussion

• Question: What is the expected geometric shape of a uniform probability distribution where all one-minute time intervals are equally likely to occur?
  • Student Response: Bell-shaped.
  • Instructor Clarification: A bell-shaped distribution is tall in the center and shorter at the tails, which would imply central minutes are more likely to occur than extreme minutes. Because every one-minute interval has equal probability, the heights must be identical across the entire domain, producing a flat-topped rectangle (analogous to the equal-frequency histogram produced by rolling a fair six-sided die with outcomes $1$ through $6$).
• Question: In the uniform delivery window from $10:00\,\text{AM}$ to $11:00\,\text{AM}$, what is the interval length between $15$ and $30\,\text{minutes}$, and what is the resulting area?
  • Student Response: $15$.
  • Instructor Clarification: With a base width of $15\,\text{units}$ and a uniform height of $\frac{1}{60}$, the area is calculated as $15 \times \frac{1}{60} = 0.25$, demonstrating a $25\%$ probability of arrival during that specific window.
• Question: Does calculating area beneath continuous distributions require students to perform calculus?
  • Student Concern: Apprehension regarding lack of prior coursework in calculus.
  • Instructor Clarification: Evaluating areas under non-rectangular continuous functions relies on calculus in theory. However, introductory statistics is structured for students without calculus backgrounds because software algorithms and pre-calculated normal probability tables perform the calculus computations automatically.
• Question: Where must the cutoff value of $2100\,\text{pounds}$ be located on the giraffe weight normal curve?
  • Student Response: To the left of the mean.
  • Instructor Clarification: Correct. On the horizontal axis, $2100$ is less than the mean of $2200$, placing it to the left of the center line, positioned between the lower inflection point at $2000$ and the mean at $2200$.