Continuous, Uniform, and Normal Probability Distributions

Foundations of Continuous Probability Distributions

• Definition of Continuous Random Variables:

  • A continuous random variable possesses an uncountable, infinite number of possible outcomes across a continuous scale or interval.

  • Continuous distributions contrast directly with discrete probability distributions, where random variables take on distinct, countable values represented graphically by discrete bars.

• Three Fundamental Characteristics of Continuous Probability Distributions:

  • Characteristic 1: The probability of an exact individual value is always zero:

  • Mathematically expressed as: $P(X = c) = 0$ for any exact value $c$.

  • Physical illustration via height measurement: If measuring the height of all adult men globally, the probability of finding an individual who is precisely $6\,\text{feet}$ tall ($P(X = 6) = 0$) is zero.

  • Mechanism of infinite precision: If an individual claims to be exactly $6\,\text{feet}$ tall, increasing measurement precision with micro-scale measuring devices reveals heights such as $5.99997\,\text{feet}$ or $6.00100\,\text{feet}$. Because continuous scales extend indefinitely, no individual measures exactly $6.00000\dots\,\text{feet}$.

  • Operational consequence: Continuous probability questions cannot evaluate exact single points; they must always be formulated using inequalities, such as $X < c$, $X > c$, $a < X < b$, or compound ranges.

  • Characteristic 2: Distributions are defined by a Probability Density Function ($\text{PDF}$):

  • The Probability Density Function, denoted as $\text{PDF}$ or $P(X)$, describes the geometric curve or line of the distribution across its domain on the horizontal $x$-axis.

  • Graph structure: The continuous random variable $X$ is plotted along the horizontal axis ($x$-axis), while the density function $P(X)$ is plotted along the vertical axis ($y$-axis).

  • Geometric profiles: The $\text{PDF}$ can take the form of a straight horizontal line, a bell curve, or any other continuous mathematical function.

  • Domain bounds: The random variable $X$ is defined between a minimum value $a$ and a maximum value $b$ (for instance, height bounded within an extreme theoretical range of $0\,\text{feet}$ to $10\,\text{feet}$).

  • Non-negativity condition: The probability density function cannot be negative; $P(X) \ge 0$ for all values of $X$ bounded by $a$ and $b$. The function can never extend below the horizontal $x$-axis.

  • Characteristic 3: The total area under the $\text{PDF}$ curve between $a$ and $b$ equals one:

  • Analogy to discrete distributions: In discrete probability distributions, the sum of all individual probabilities equals one ($\sum P(X) = 1$). In continuous probability distributions, the total area under the continuous curve from the minimum bound $a$ to the maximum bound $b$ equals $1$.

  • Central axiom: In all continuous probability distributions, area is equivalent to probability.

  • Calculation principle: To compute the probability of a continuous variable falling between two points (e.g., adult male height between $5.7\,\text{feet}$ and $6.2\,\text{feet}$), locate $5.7$ and $6.2$ on the $x$-axis, project vertical lines to the $\text{PDF}$ line/curve, and calculate the bounded area between those boundaries.

The Uniform Probability Distribution

• Definition and Overview:

  • A uniform distribution is a continuous probability distribution characterized by a constant probability density across its entire domain.

  • The $\text{PDF}$ line is completely flat and horizontal between its lower bound and upper bound.

• Three Defining Characteristics of a Uniform Distribution:

  • Characteristic 1: Rectangular geometric profile:

  • The distribution is defined over a closed domain from a minimum value $a$ to a maximum value $b$.

  • Dropping vertical lines from the constant $\text{PDF}$ height to the $x$-axis at points $a$ and $b$ forms a rectangle.

  • Characteristic 2: Perfect central symmetry:

  • The mean and the median of a uniform distribution are equal and positioned at the precise geometric midpoint of the distribution.

  • The formula for both the mean ($\mu$) and the median is:       $\mu = \text{median} = \frac{a + b}{2}$

  • Characteristic 3: Complete specification by minimum and maximum parameters:

  • The distribution is entirely determined by its minimum value $a$ and its maximum value $b$.

• Derivation of the Uniform Probability Density Function Height ($P(X)$):

  • The area of a rectangular box is given by the elementary geometry formula:     $\text{Area} = \text{base} \times \text{height}$

  • The horizontal base along the $x$-axis spans from $a$ to $b$, giving:     $\text{Base} = b - a$

  • By the third characteristic of continuous distributions, the total area under the curve must equal $1$:     $1 = (b - a) \times \text{height}$

  • Solving for height yields the probability density function formula:     $\text{Height} = P(X) = \frac{1}{b - a}$

  • Formal piece-wise definition of the Uniform $\text{PDF}$:     $P(X) = \frac{1}{b - a} \quad \text{for } a \le X \le b$     $P(X) = 0 \quad \text{elsewhere}$

• Formulas for the Uniform Distribution:

  • Minimum value parameter: $a$

  • Maximum value parameter: $b$

  • Probability Density Function: $P(X) = \frac{1}{b - a}$

  • Population Mean ($\mu$): $\mu = \frac{a + b}{2}$

  • Population Median: $\text{Median} = \frac{a + b}{2}$

  • Population Variance ($\sigma^2$): $\sigma^2 = \frac{(b - a)^2}{12}$

  • Population Standard Deviation ($\sigma$): $\sigma = \sqrt{\frac{(b - a)^2}{12}}$

  • Note on formula retention: The standard deviation formula $\sigma = \sqrt{\frac{(b - a)^2}{12}}$ is critical for recurring problem-solving in homework assignments, examples, and examinations.

Applications and Problem Solving for Uniform Distributions

• Bus Stop Waiting Time Problem:

  • Scenario: A municipal bus operates with a schedule arriving every $30\,\text{minutes}$ between 6:00 AM and 11:00 PM on weekdays. Students arrive at the bus stop at random times. The waiting time $X$ is a continuous random variable uniformly distributed from $0\,\text{minutes}$ to $30\,\text{minutes}$.

  • Parameter identification:

  • Minimum waiting time: $a = 0\,\text{minutes}$ (arriving right as the bus pulls up).

  • Maximum waiting time: $b = 30\,\text{minutes}$ (just missing the bus and waiting for the subsequent one).

  • Formulation of $\text{PDF}$ height:     $P(X) = \frac{1}{b - a} = \frac{1}{30 - 0} = \frac{1}{30}$

  • Graph representation:

  • A rectangular region spanning along the $x$-axis from $0$ to $30$, with a constant height of $\frac{1}{30}$.

  • Proof that the total probability over the domain equals $1$:

  • Area evaluation over interval $[0, 30]$:       $\text{Area} = \text{base} \times \text{height} = (30 - 0) \times \frac{1}{30} = 30 \times \frac{1}{30} = 1$

  • Therefore, $P(0 \le X \le 30) = 1$.

  • Computation of the distribution mean ($\mu$):     $\mu = \frac{a + b}{2} = \frac{0 + 30}{2} = 15\,\text{minutes}$

  • Computation of the distribution standard deviation ($\sigma$):     $\sigma = \sqrt{\frac{(b - a)^2}{12}} = \sqrt{\frac{(30 - 0)^2}{12}} = \sqrt{\frac{900}{12}} = \sqrt{75} \approx 8.66\,\text{minutes}$

  • Computation of probability of waiting more than $25\,\text{minutes}$ ($P(X > 25)$):

  • Lower limit of region: $25$

  • Upper limit of region: $30$

  • Base of shaded rectangle: $30 - 25 = 5$

  • Height of shaded rectangle: $\frac{1}{30}$

  • Area computation:       $P(X > 25) = 5 \times \frac{1}{30} = \frac{5}{30} = \frac{1}{6}$

  • Computation of probability of waiting between $10$ and $20\,\text{minutes}$ ($P(10 \le X \le 20)$):

  • Interval limits: From $10$ to $20$

  • Base of shaded rectangle: $20 - 10 = 10$

  • Height of shaded rectangle: $\frac{1}{30}$

  • Area computation:       $P(10 \le X \le 20) = 10 \times \frac{1}{30} = \frac{10}{30} = \frac{1}{3} \approx 0.3333$

  • Computation of probability of waiting exactly $15\,\text{minutes}$ ($P(X = 15)$):

  • The target is a single discrete point, representing a vertical line of zero width on the continuous graph.

  • Base width: $0$

  • Area computation:       $\text{Area} = 0 \times \frac{1}{30} = 0$

  • Therefore, $P(X = 15) = 0$.

• Microwave Oven Lifetime Problem:

  • Scenario: The lifetime of a microwave oven follows a uniform distribution bounded between $8\,\text{years}$ and $14\,\text{years}$.

  • Parameter identification:

  • Minimum lifetime: $a = 8\,\text{years}$

  • Maximum lifetime: $b = 14\,\text{years}$

  • Formulation of $\text{PDF}$ height:     $P(X) = \frac{1}{b - a} = \frac{1}{14 - 8} = \frac{1}{6}$

  • Verification of total area:

  • Base: $14 - 8 = 6$

  • Height: $\frac{1}{6}$

  • Area: $6 \times \frac{1}{6} = 1$

  • Mean calculation ($\mu$):     $\mu = \frac{a + b}{2} = \frac{8 + 14}{2} = \frac{22}{2} = 11\,\text{years}$

  • Standard deviation calculation ($\sigma$):     $\sigma = \sqrt{\frac{(b - a)^2}{12}} = \sqrt{\frac{(14 - 8)^2}{12}} = \sqrt{\frac{6^2}{12}} = \sqrt{\frac{36}{12}} = \sqrt{3} \approx 1.732\,\text{years}$

  • Computation of probability of lasting between $10$ and $14\,\text{years}$ ($P(10 \le X \le 14)$):

  • Base: $14 - 10 = 4$

  • Height: $\frac{1}{6}$

  • Area computation:       $P(10 \le X \le 14) = 4 \times \frac{1}{6} = \frac{4}{6} = \frac{2}{3} \approx 0.6667$

  • Computation of probability of lasting less than $9\,\text{years}$ ($P(X < 9)$):

  • Lower bound is the minimum value: $a = 8$

  • Base: $9 - 8 = 1$

  • Height: $\frac{1}{6}$

  • Area computation:       $P(X < 9) = 1 \times \frac{1}{6} = \frac{1}{6} \approx 0.1667$

  • Inclusion vs. Exclusion of Boundaries:

  • In continuous distributions, strict inequalities ($<$ or $>$) yield identical numerical probabilities to non-strict inequalities ($\le$ or $\ge$):       $P(X \le 10) = P(X < 10) + P(X = 10)$

  • Because $P(X = 10) = 0$, it follows that:       $P(X \le 10) = P(X < 10)$

  • Whether interval boundaries are included or excluded has no effect on the calculated probability.

• Family Insurance Spending Problem:

  • Scenario: Data from the Insurance Institute of America indicates that a family of four spends between $\$400$ and $\$3{,}800$ per year on all types of insurance, following a uniform distribution.

  • Parameter identification:

  • Minimum spending: $a = 400$

  • Maximum spending: $b = 3800$

  • Formulation of $\text{PDF}$ height:     $P(X) = \frac{1}{b - a} = \frac{1}{3800 - 400} = \frac{1}{3400}$

  • Mean calculation ($\mu$):     $\mu = \frac{a + b}{2} = \frac{400 + 3800}{2} = \frac{4200}{2} = 2100$     $\mu = \$2{,}100$

  • Standard deviation calculation ($\sigma$):     $\sigma = \sqrt{\frac{(b - a)^2}{12}} = \sqrt{\frac{(3800 - 400)^2}{12}} = \sqrt{\frac{3400^2}{12}} = \sqrt{\frac{11560000}{12}} \approx 981.5$     $\sigma \approx \$981.50$

  • Computation of probability of spending less than $\$2{,}000$ per year ($P(X < 2000)$) :

  • Base: $2000 - 400 = 1600$

  • Height: $\frac{1}{3400}$

  • Area computation:       $P(X < 2000) = 1600 \times \frac{1}{3400} = \frac{1600}{3400} = \frac{8}{17} \approx 0.4706$

  • Computation of probability of spending more than $\$3{,}000$ per year ($P(X > 3000)$) :

  • Upper limit is maximum value: $b = 3800$

  • Base: $3800 - 3000 = 800$

  • Height: $\frac{1}{3400}$

  • Area computation:       $P(X > 3000) = 800 \times \frac{1}{3400} = \frac{800}{3400} = \frac{4}{17} \approx 0.2353$

The Normal Probability Distribution

• Overview and Nomenclature:

  • The normal probability distribution is also widely termed the Gaussian distribution.

  • It serves as the foundational probability distribution for statistical inference and analysis.

• Four Fundamental Characteristics of the Normal Distribution:

  • Characteristic 1: Bell-shaped and Unimodal:

  • The curve exhibits a symmetric bell shape with a single peak located at the precise center of the distribution.

  • Characteristic 2: Perfect Central Symmetry:

  • The distribution is completely symmetrical about the mean ($\mu$).

  • The mean, median, and mode are identical values located at the exact center of the curve.

  • Characteristic 3: Unit Total Area:

  • Like all continuous distributions, the total area under the normal curve bounded between negative infinity and positive infinity is exactly equal to one:       $\text{Total Area} = 1$

  • Characteristic 4: Asymptotic Tails:

  • The curve is asymptotic with respect to the horizontal axis ($x$-axis).

  • As the tails extend indefinitely in both directions toward negative infinity ($-\infty$) and positive infinity ($+\infty$), they get closer and closer to the horizontal axis but never physically touch or intersect it.

• Probability Density Function Mathematical Formula:

  • The mathematical curve describing the normal probability density function is:     $P(X) = \frac{1}{\sigma \sqrt{2\pi}} e^{-\frac{(X - \mu)^2}{2\sigma^2}}$

  • Operational note: Manual integration or calculation using this formula is not required for standard coursework applications; table lookups and standardization transformations replace direct calculus integration.

• Prevalence in Nature:

  • The normal distribution appears frequently across empirical measurements in biological and physical systems:

  • Human adult heights (clustering heavily around population averages with symmetrical dispersion into high and low extremes).

  • Dimensions and growth sizes of trees.

  • Lifespan and age patterns of specific insect populations.

• Comparative Parameter Analysis ($\mu$ and $\sigma$):

  • Role of the Mean ($\mu$):

  • Determines the central location of the distribution along the horizontal axis.

  • Distributions sharing the same mean are centered at the exact same location.

  • Role of the Standard Deviation ($\sigma$):

  • Determines the spread, dispersion, and width of the distribution.

  • If distributions have the same mean but different standard deviations ($\sigma$), their peaks share the exact same horizontal coordinate, but their bell curves exhibit different widths and peak heights.

  • If distributions have different means but identical standard deviations ($\sigma$), their bell curves exhibit identical shapes and spreads but are translated to different positions along the horizontal axis.

• Calculus Integration vs. Standardization:

  • To calculate the probability of a normally distributed variable falling within an interval (such as male heights between $5.7\,\text{feet}$ and $6.2\,\text{feet}$), the theoretical mathematical procedure requires integration:     $P(5.7 \le X \le 6.2) = \int_{5.7}^{6.2} P(X)\,dX$

  • Because direct integration of the Gaussian function is complex, all general normal distributions are converted into the Standard Normal Distribution ($Z$-distribution) to determine probabilities.

Standard Normal Distribution and Z-Score Transformations

• Definition of the Standard Normal Distribution:

  • The standard normal distribution (designated as the $Z$-distribution) is a specialized case of the general normal distribution defined by two fixed parameters:

  • Population Mean: $\mu = 0$

  • Population Standard Deviation: $\sigma = 1$

  • Variable designation: The letter $Z$ is used specifically to prevent confusion with $s$, which represents sample standard deviation.

• The Universal Standardization Principle:

  • Any general normal distribution $X$ with arbitrary mean $\mu$ and arbitrary standard deviation $\sigma$ can be converted into the standard normal distribution $Z$.

  • Standardization geometrically translates the distribution center to $0$ and rescales its dispersion to a standard deviation of $1$.

  • Consequently, calculating the area under the curve of a transformed standard normal distribution yields the identical probability as finding the corresponding area under the original raw normal curve.

• The Z-Score Transformation Formula:

  • To convert any raw score $X$ into a standard score $Z$:     $Z = \frac{X - \mu}{\sigma}$

  • Component definitions:

  • $X$: The raw observation or value from the original distribution.

  • $\mu$: The mean of the original normal distribution.

  • $\sigma$: The standard deviation of the original normal distribution.

  • $Z$: The resulting standard score representing the signed distance of $X$ from $\mu$ in units of standard deviation.

  • Algebraic inversion of the formula:     $X - \mu = Z\sigma$     $X = \mu + Z\sigma$

  • Interpretation of the $Z$-score:

  • A positive $Z$-score (such as $Z = +1.5$) indicates that the raw observation $X$ is $1.5$ standard deviations greater than the population mean ($X = \mu + 1.5\sigma$).

  • A negative $Z$-score (such as $Z = -1.5$) indicates that the raw observation $X$ is $1.5$ standard deviations less than the population mean ($X = \mu - 1.5\sigma$).

• Structure of the Standard Normal Table ($Z$-Table):

  • Located in Appendix B.3 (page 530) of the course e-textbook; distributed as a physical handout during exams.

  • The table tabulates the cumulative area under the standard normal curve from the center ($Z = 0$) up to a specified positive $Z$-value.

  • Procedure for table lookup:

  • To find the area between $Z = 0$ and $Z = 1.96$:

    • Locate the row labeled 1.9 in the left-hand column.

    • Locate the column labeled .06 along the top row.

    • Find the intersection value of row 1.9 and column .06, which reads $0.4750$.

  • Interpretation: The area under the standard normal curve between the mean ($Z = 0$) and $Z = 1.96$ is equal to $0.4750$.

• Step-by-Step Z-Score Calculation Examples:

  • Uber Driver Weekly Earnings Example:

  • Scenario: Weekly income for Uber drivers is normally distributed with a mean of $\mu = \$1{,}000$ and a standard deviation of $\sigma = \$100$.

  • Case 1: Driver earning $\$1{,}100$ per week:       $Z = \frac{X - \mu}{\sigma} = \frac{1100 - 1000}{100} = \frac{100}{100} = 1.0$

    • Interpretation: Earnings of $\$1{,}100$ convert to $Z = +1.0$, exactly $1$ standard deviation above the mean.

  • Case 2: Driver earning $\$900$ per week:       $Z = \frac{X - \mu}{\sigma} = \frac{900 - 1000}{100} = \frac{-100}{100} = -1.0$

    • Interpretation: Earnings of $\$900$ convert to $Z = -1.0$, exactly $1$ standard deviation below the mean.

  • General Distribution Example ($\mu = 48$, $\sigma = 12.8$):

  • Case 1: Raw score $X = 64$:       $Z = \frac{64 - 48}{12.8} = \frac{16}{12.8} = 1.5$

    • Algebraic relationship: $64 = 48 + 1.5(12.8)$.

    • Interpretation: $64$ lies $1.5$ standard deviations above the mean.

  • Case 2: Raw score $X = 32$:       $Z = \frac{32 - 48}{12.8} = \frac{-16}{12.8} = -1.5$

    • Algebraic relationship: $32 = 48 - 1.5(12.8)$.

    • Interpretation: $32$ lies $1.5$ standard deviations below the mean.

The Empirical Rule and Standard Normal Table Analysis

• The Empirical Rule (68-95-99.7 Rule):

  • Applies specifically to symmetric, bell-shaped normal distributions.

  • Core intervals:

  • Approximately $68\%$ of observations lie within one standard deviation of the mean: $\mu \pm 1\sigma$.

  • Approximately $95\%$ of observations lie within two standard deviations of the mean: $\mu \pm 2\sigma$.

  • Approximately $99\%$ (or almost all) of observations lie within three standard deviations of the mean: $\mu \pm 3\sigma$.

• Mathematical Derivation of the $68\%$ Value via the $Z$-Table:

  • The interval $\mu \pm 1\sigma$ corresponds to standard scores of $Z = -1.0$ and $Z = +1.0$.

  • Locating $Z = 1.00$ in the Appendix B.3 table (row 1.0, column .00) yields an area of $0.3413$ between the center ($Z = 0$) and $Z = 1.00$.

  • By the property of bilateral symmetry, the area from $Z = -1.00$ to $Z = 0$ is also $0.3413$.

  • Adding both symmetric halves:     $\text{Total Area} = 0.3413 + 0.3413 = 0.6826$

  • Expressed as a percentage, $0.6826$ rounds to $68\%$, demonstrating the analytical origin of the Empirical Rule.

• Academic and Exam Context:

  • The Empirical Rule connects descriptive statistics (studied in Chapter 3) with continuous probability models (studied in Chapter 7).

  • Because it bridges both midterm curricula, it is a primary topic emphasized for the final examination.

• Orolife Battery Company Quality Assurance Application:

  • Scenario: Orolife battery company conducts quality control testing on D-cell battery life. Useful life is normally distributed with a sample mean of $\bar{X} = 19\,\text{hours}$ and a sample standard deviation of $s = 1.2\,\text{hours}$.

  • Calculation of Empirical Rule Intervals using sample statistics ($\bar{X}$ and $s$):

  • $68\%$ of battery lifespans lie within $\bar{X} \pm 1s$:       $\bar{X} \pm 1s = 19 \pm 1(1.2) = 19 \pm 1.2$       $\text{Lower Bound} = 19 - 1.2 = 17.8\,\text{hours}$       $\text{Upper Bound} = 19 + 1.2 = 20.2\,\text{hours}$

    • Interval: $[17.8\,\text{hours}, 20.2\,\text{hours}]$

  • $95\%$ of battery lifespans lie within $\bar{X} \pm 2s$:       $\bar{X} \pm 2s = 19 \pm 2(1.2) = 19 \pm 2.4$       $\text{Lower Bound} = 19 - 2.4 = 16.6\,\text{hours}$       $\text{Upper Bound} = 19 + 2.4 = 21.4\,\text{hours}$

    • Interval: $[16.6\,\text{hours}, 21.4\,\text{hours}]$

  • $99\%$ (almost all) of battery lifespans lie within $\bar{X} \pm 3s$:       $\bar{X} \pm 3s = 19 \pm 3(1.2) = 19 \pm 3.6$       $\text{Lower Bound} = 19 - 3.6 = 15.4\,\text{hours}$       $\text{Upper Bound} = 19 + 3.6 = 22.6\,\text{hours}$

    • Interval: $[15.4\,\text{hours}, 22.6\,\text{hours}]$