Continuous Density Functions and the Normal Distribution
You request a ride and the app says the car will arrive "sometime in the next 12 minutes." What is the probability that it shows up at exactly 5 minutes, not a split second earlier or later? And what is the probability that it arrives between 3 and 8 minutes? Continuous probability gives surprising answers to both.
Density functions: area is probability
A continuous random variable can take any value in an interval, like a waiting time, a height, or a battery life. You can't list all the outcomes, so you describe the variable with a curve called a density function $f(x)$. The probability of landing in an interval is the area under the curve over that interval.
Here is why the rules make sense. The total area is $1$ because the variable has to take some value, so the probabilities add up to 100%. The curve can't go negative because there is no such thing as a negative probability.
The simplest density is a uniform distribution: a flat rectangle, because every value in the interval is equally likely.
The probability density function of a continuous random variable $X$ representing the waiting time (in minutes) for a match to start on a video game server is defined…
The probability density function of a continuous random variable $X$ representing the waiting time (in minutes) for a match to start on a video game server is defined as:
$f(x) = \begin{cases} k & \text{if } 0 \le x \le 5 \\ 0 & \text{otherwise} \end{cases}$
Determine the value of the constant $k$ for $f(x)$ to be a valid probability density function and graph the function.
The battery life of a new model of wireless headphones follows a uniform distribution between 4 and 10 hours. Its probability density function is constant in that…
The battery life of a new model of wireless headphones follows a uniform distribution between 4 and 10 hours. Its probability density function is constant in that interval: $f(x) = \frac{1}{6}$. Calculate the probability that the battery lasts between 5 and 8 hours, interpreting it as the area of a rectangle.
Finding k and computing areas
Many density functions are piecewise, made of lines. The graph is then a rectangle, a triangle, or a trapezoid, so you only need geometry. A missing constant $k$ is found from the one condition we know: the total area is $1$.
The plan is the same every time. First sketch the graph, then set the total area equal to $1$ and solve for $k$. Finally write each probability as an area, and use the complement ($1 -$ the other part) when that is shorter.
Some densities have a "tent" shape, with two pieces that meet at a peak.
For a trapezoid, split it into a rectangle plus a triangle, or use $\tfrac{1}{2}(b_1 + b_2)h$ on the part you need.
The reaction time, in seconds, of a group of teenagers to a visual stimulus in a video game is being studied. The probability density function of this continuous…
The reaction time, in seconds, of a group of teenagers to a visual stimulus in a video game is being studied. The probability density function of this continuous variable is given by the following expression:
$f(x) = \begin{cases} kx & \text{if } 0 \le x \le 2 \\ 0 & \text{otherwise} \end{cases}$
Determine:
- a) The value of the constant $k$ for $f(x)$ to be a valid probability density function.
- b) The probability that a randomly chosen player has a reaction time greater than $1,5$ seconds.
- c) The probability that the reaction time is exactly between $0,5$ and $1$ seconds.
The waiting time, in minutes, until an update is downloaded on a console follows a continuous distribution whose probability density function has the shape of a…
The waiting time, in minutes, until an update is downloaded on a console follows a continuous distribution whose probability density function has the shape of a trapezoid. The function is defined between $0$ and $10$ minutes. Between minutes $0$ and $4$, the function is constant with a value of $h$, and between minutes $4$ and $10$ it decreases linearly until reaching $0$.
- a) Calculate the value of $h$ for it to be a probability density function.
- b) Calculate the probability that the update takes less than $6$ minutes to download.
- c) If a user has been waiting for more than $4$ minutes, what is the probability that the download finishes before minute $8$?
Let $f(x)$ be a piecewise function defined as follows:…
Let $f(x)$ be a piecewise function defined as follows:
$f(x) = \begin{cases} k \cdot x & \text{if } 0 \le x < 2 \\ k \cdot (4 - x) & \text{if } 2 \le x \le 4 \\ 0 & \text{otherwise} \end{cases}$
Determine, justifying your answer, the value of the constant $k$ for $f(x)$ to be a probability density function of a continuous random variable $X$. Once $k$ is found, calculate the probability that the variable takes a value in the interval $[1, 3]$ and represent it graphically.
The cumulative function and the median
Calculating a new area for every question gets slow. The cumulative distribution function collects the area once.
With $F$, any probability is a subtraction:
$$P(a \le X \le b) = F(b) - F(a) \qquad P(X > a) = 1 - F(a)$$The area up to $b$ includes the area up to $a$, so you subtract the part you don't want.
The median $M$ splits the area into two equal halves, so it is the solution of $F(M) = 0.5$. For a conditional probability with nested events, use $P(A \mid B) = \dfrac{P(A \text{ and } B)}{P(B)}$. For example, $P(X > 4 \mid X > 3) = \dfrac{P(X > 4)}{P(X > 3)}$, because "$X > 4$" already implies "$X > 3$".
Given the probability density function $f(x) = \frac{1}{8}(x-1)$ defined for the interval $x \in [1, 5]$ and $f(x)=0$ elsewhere: a) Show that the total area under the…
Given the probability density function $f(x) = \frac{1}{8}(x-1)$ defined for the interval $x \in [1, 5]$ and $f(x)=0$ elsewhere:
- a) Show that the total area under the curve is indeed $1$.
- b) Find a value $M$ (median) such that $P(X \le M) = 0,5$.
- c) Calculate the conditional probability $P(X > 4 | X > 2)$.
The normal curve and the z-table
The most important density in statistics is the normal distribution, the bell curve. It describes heights, test scores, measurement errors, and many other things. It is written $N(\mu, \sigma)$, where $\mu$ is the mean and $\sigma$ is the standard deviation. The curve is symmetric about $\mu$, so the mean and median are both at the center, and $\sigma$ controls how spread out it is. The total area is $1$.
There is no easy area formula for the bell curve, so we use a table (or a calculator). To keep one table for all normal curves, we use the standard normal distribution $Z \sim N(0, 1)$, which has mean $0$ and standard deviation $1$. The z-table gives $P(Z \le z)$, the area to the left of $z$.
Selected values of the standard normal table, $P(Z \le z)$ (a full z-table or your calculator has the rest):
| $z$ | area | $z$ | area | $z$ | area |
|---|---|---|---|---|---|
| 0.00 | 0.5000 | 0.85 | 0.8023 | 1.50 | 0.9332 |
| 0.45 | 0.6736 | 1.00 | 0.8413 | 1.52 | 0.9357 |
| 0.48 | 0.6844 | 1.12 | 0.8686 | 1.62 | 0.9474 |
| 0.50 | 0.6915 | 1.15 | 0.8749 | 1.78 | 0.9625 |
| 0.62 | 0.7324 | 1.19 | 0.8830 | 1.96 | 0.9750 |
| 0.67 | 0.7486 | 1.20 | 0.8849 | 2.00 | 0.9772 |
| 0.75 | 0.7734 | 1.28 | 0.8997 | 2.05 | 0.9798 |
| 0.81 | 0.7910 | 1.33 | 0.9082 | 2.38 | 0.9913 |
| 0.84 | 0.7995 | 1.42 | 0.9222 | 3.00 | 0.9987 |
To read a value such as $z = 1.42$, find the row for $1.4$ and the column for $0.02$. Reading the table backward works too. If you are given the area $0.9332$, look for it in the body of the table, and the row and column give $z = 1.50$.
Given a standard normal distribution $Z \sim N(0, 1)$, use the probability table to find the following values directly: a) $P(Z \le 0,45)$ b) $P(Z \le 1,2)$ c)…
Given a standard normal distribution $Z \sim N(0, 1)$, use the probability table to find the following values directly:
-
a) $P(Z \le 0,45)$
-
b) $P(Z \le 1,2)$
-
c) $P(Z \le 2,38)$
Use the table of the $N(0, 1)$ distribution in reverse to find the value of the constant $k$ in the following cases: a) $P(Z \le k) = 0,8413$ b) $P(Z \le k) = 0,9750$ c)…
Use the table of the $N(0, 1)$ distribution in reverse to find the value of the constant $k$ in the following cases:
-
a) $P(Z \le k) = 0,8413$
-
b) $P(Z \le k) = 0,9750$
-
c) $P(Z \le k) = 0,5000$
Symmetry and areas between values
Since the table only shows areas to the left, you combine it with two facts: the total area is $1$, and the curve is symmetric about $0$.
For a variable $Z \sim N(0, 1)$, apply the properties of the normal curve (symmetry and total area) to calculate the following probabilities: a) $P(Z > 1,52)$ b)…
For a variable $Z \sim N(0, 1)$, apply the properties of the normal curve (symmetry and total area) to calculate the following probabilities:
-
a) $P(Z > 1,52)$
-
b) $P(Z < -0,84)$
-
c) $P(Z \ge -1,15)$
Calculate the probability that the standard normal variable $Z$ takes a value in the interval $(0,62 \le Z \le 1,96)$. Remember that to calculate the area of an interval…
Calculate the probability that the standard normal variable $Z$ takes a value in the interval $(0,62 \le Z \le 1,96)$. Remember that to calculate the area of an interval you must subtract the cumulative areas: $P(a \le Z \le b) = P(Z \le b) - P(Z \le a)$.
Given a random variable $Z$ that follows a standard normal distribution $N(0, 1)$, use the properties of symmetry and the total area to calculate the following…
Given a random variable $Z$ that follows a standard normal distribution $N(0, 1)$, use the properties of symmetry and the total area to calculate the following probabilities:
- a) $P(Z < 1,42)$
- b) $P(Z > -0,75)$
- c) $P(1,12 < Z < 2,05)$
- d) $P(-1,5 < Z < 1,5)$
- e) Find the value of $k$ such that $P(Z < k) = 0,8830$
Standardizing with z-scores
Real data rarely have mean $0$ and standard deviation $1$. To use the z-table for $X \sim N(\mu, \sigma)$, convert $x$ to a z-score: how many standard deviations $x$ is from the mean.
Subtracting $\mu$ slides the center to $0$, and dividing by $\sigma$ rescales the spread to $1$. The area under the curve doesn't change.
The scores of an English proficiency exam for high school seniors follow a normal distribution with mean $\mu = 75$ and standard deviation $\sigma = 8$. Standardize the…
The scores of an English proficiency exam for high school seniors follow a normal distribution with mean $\mu = 75$ and standard deviation $\sigma = 8$. Standardize the following values of the variable $X$ to find their corresponding value in $Z$:
-
a) A score of $83$.
-
b) A score of $67$.
-
c) A score of $75$.
The weight of the bags of potato chips produced by a machine follows a distribution of $N(150, 5)$ ounces. Calculate the probability that a randomly chosen bag weighs…
The weight of the bags of potato chips produced by a machine follows a distribution of $N(150, 5)$ ounces. Calculate the probability that a randomly chosen bag weighs less than $160$ ounces. To do this, first standardize the variable and look up the result in the table.
The battery life of a specific model of wireless headphones follows a normal distribution with a mean of $30$ hours and a standard deviation of $4$ hours. a) What…
The battery life of a specific model of wireless headphones follows a normal distribution with a mean of $30$ hours and a standard deviation of $4$ hours.
- a) What percentage of headphones have a battery that lasts between $26$ and $34$ hours?
- b) If the manufacturer guarantees that the headphones last at least $22$ hours, what is the probability of finding headphones that do not meet the guarantee?
- c) What is the probability that the battery lasts more than $35$ hours?
Percentages, counts, and cutoff values
Once you can standardize, three kinds of questions follow.
How many? Multiply the group size by the probability. In a group of $n$ people, you expect about $n \cdot P$ to land in the interval.
What interval? A symmetric interval around the mean that holds $95\%$ of the data is $\mu \pm 1.96\sigma$, because $P(-1.96 < Z < 1.96) = 2(0.9750) - 1 = 0.95$.
What cutoff? Work backward in two steps. First find $z$ from the area to the left, using the table in reverse. Then convert back with
$$x = \mu + z\sigma$$For the top $5\%$ you need an area to the left of $0.95$. That falls halfway between $z = 1.64$ and $z = 1.65$, so use $z \approx 1.645$. For the top $10\%$ (area $0.90$), use $z \approx 1.28$.
The scores obtained by the students of a school in the college entrance exam follow a normal distribution $N(6,8 ; 1,2)$. a) If accessing a highly demanded major…
The scores obtained by the students of a school in the college entrance exam follow a normal distribution $N(6,8 ; 1,2)$.
- a) If accessing a highly demanded major requires a score higher than $8,5$, what percentage of the school's students would have the option to enter?
- b) The school decides to award an honorable mention to the $5 \%$ of students with the best scores. From what score is this mention received?
- c) What is the probability that a student has a score lower than $5$?
In a study on the monthly leisure expenditure of young people in a city, it has been determined that it follows a normal distribution with a mean of $\$80…
In a study on the monthly leisure expenditure of young people in a city, it has been determined that it follows a normal distribution with a mean of $\$80$ and a standard deviation of $\$15$.
- a) Calculate the interval centered on the mean that comprises $95 \%$ of the young people.
- b) If we select a young person at random, what is the probability that their monthly expenditure does not deviate from the mean by more than $\$10$?
- c) If there are $5.000$ young people in the city, how many of them are estimated to spend more than $\$110$ a month?
A farmer classifies the oranges from his harvest by their diameter. The diameter follows a normal distribution $N(8 ; 0,6)$ in inches. a) Oranges with a diameter of less…
A farmer classifies the oranges from his harvest by their diameter. The diameter follows a normal distribution $N(8 ; 0,6)$ in inches.
- a) Oranges with a diameter of less than $7$ inches are destined for juice manufacturing. What probability does an orange have of ending up in the juice factory?
- b) "Extra" oranges are those whose diameter is between $8,5$ and $9,5$ inches. What percentage of the harvest is labeled as "Extra"?
- c) If the farmer wants to separate the largest $10 \%$ of the oranges for a luxury export, what must be the minimum diameter of these oranges?
When μ or σ is unknown
Sometimes the problem tells you a percentage and a value, and you must find $\sigma$ or $\mu$. The z-score formula still holds, but now $z$ comes from the table and the unknown is what you solve for: $z = \dfrac{x - \mu}{\sigma}$. Convert the percentage to a $z$ first, then solve the one equation.
The same idea finds a mean you can adjust. If a soda machine has $\sigma = 2$ oz and you want only $2.28\%$ of cups to have less than $12$ oz, then $z = -2$ and $12 = \mu - 2(2)$. So $\mu = 12 + 4 = 16$ oz. In general, $\mu = x - z\sigma$.
In a young musical talent contest, the scores follow a normal distribution with a mean of $72$ points. It is known that to obtain an "Honorable Mention," a score is…
In a young musical talent contest, the scores follow a normal distribution with a mean of $72$ points. It is known that to obtain an "Honorable Mention," a score is required such that only $10 \%$ of the participants achieve it. If the minimum score to receive this mention is $84,8$ points, what is the standard deviation of the scores? If $500$ participants enter, how many are expected to have a score lower than $60$ points?
It has been verified that the weight of the $500$-ounce coffee bags leaving a packaging machine follows a normal distribution $N(505, 10)$. Quality control rejects those…
It has been verified that the weight of the $500$-ounce coffee bags leaving a packaging machine follows a normal distribution $N(505, 10)$. Quality control rejects those bags that weigh less than $490$ ounces or more than $525$ ounces.
- a) What percentage of bags is rejected by the factory?
- b) If the manager decides to adjust the machine so that the mean weight is $\mu$, but keeping the standard deviation $\sigma = 10$, what value should $\mu$ have so that the risk of a bag weighing less than $500$ ounces is only $2,5 \%$?
On a streaming platform, the time users watch a show before abandoning it follows a normal distribution. It is known that the mean time is $45$ minutes. If $25 \%$ of…
On a streaming platform, the time users watch a show before abandoning it follows a normal distribution. It is known that the mean time is $45$ minutes. If $25 \%$ of users watch for more than $60$ minutes, calculate the standard deviation. Determine the interval symmetrical with respect to the mean $(45 - d, 45 + d)$ that contains $95 \%$ of the users' viewing times.
Finding both μ and σ
When both $\mu$ and $\sigma$ are unknown, you need two pieces of information. Each one gives an equation $x = \mu + z\sigma$, and you solve the pair of linear equations with elimination (or substitution).
A courier company has analyzed the weight of the parcels it sends daily. It has been verified that the weight follows a normal distribution. We know that $15,87 \%$ of…
A courier company has analyzed the weight of the parcels it sends daily. It has been verified that the weight follows a normal distribution. We know that $15,87 \%$ of the parcels weigh less than $2$ pounds and that $2,28 \%$ of the parcels weigh more than $8$ pounds.
- a) Determine the mean and standard deviation of the weight of the parcels.
- b) What is the probability that a randomly chosen parcel weighs between $3$ and $6$ pounds?
The daily usage time of a map application on the cell phones of delivery drivers in a city follows a normal distribution $N(\mu, \sigma)$. It is known that $15,87 \%$ of…
The daily usage time of a map application on the cell phones of delivery drivers in a city follows a normal distribution $N(\mu, \sigma)$. It is known that $15,87 \%$ of the delivery drivers use it for less than $40$ minutes a day, while $2,28 \%$ use it for more than $80$ minutes. Calculate the parameters $\mu$ and $\sigma$ of this distribution. What is the probability that a randomly chosen delivery driver uses the application between $50$ and $70$ minutes?
The battery life of a model of wireless headphones follows a normal distribution with a mean of $30$ hours. If it is known that the probability of a battery lasting more…
The battery life of a model of wireless headphones follows a normal distribution with a mean of $30$ hours. If it is known that the probability of a battery lasting more than $34$ hours is the same as the probability of it lasting less than $26$ hours, and both are $0,1587$, calculate the variance of the distribution. If a user needs the headphones to work during a $33$-hour trip without recharging, what is the probability that the battery runs out before reaching their destination?
Normal approximation to the binomial
A binomial variable counts successes in $n$ independent trials, each with success probability $p$. Its mean is $np$ and its standard deviation is $\sqrt{np(1-p)}$. For a large $n$, adding up binomial probabilities is painful, but the bar graph looks like a bell curve. So we can use $N\!\left(np,\ \sqrt{np(1-p)}\right)$ instead.
A binomial count is a whole number, but the normal curve is continuous. The bar for $X = 12$ really covers the stretch from $11.5$ to $12.5$. The continuity correction means you widen each endpoint by $0.5$ so that you include the whole bar.
Some problems call this Yates's correction. It is the same continuity correction.
An electronic component manufacturing company claims that $4 \%$ of its touch screens have some factory micro-defect. If a school acquires a batch of $250$ screens for…
An electronic component manufacturing company claims that $4 \%$ of its touch screens have some factory micro-defect. If a school acquires a batch of $250$ screens for its classrooms, calculate the probability that:
- a) There are exactly $12$ defective screens.
- b) There are between $8$ and $15$ defective screens (inclusive).
Note: Justify whether it is possible to approximate the binomial distribution to the normal distribution and apply Yates's continuity correction.
Back to the start: the chance that the car arrives at exactly 5 minutes is $0$, because a single point has no area. The chance it arrives between 3 and 8 minutes is the area of a rectangle, $\tfrac{5}{12} \approx 0.417$.