The conditional distribution of Y given X is the probability distribution of Y when X has a specific value. In Intro to Probability, it comes from a joint distribution and shows how Y changes across different X values.
The conditional distribution of Y given X is the distribution you get when you fix X at a particular value and then look at how Y behaves. In Intro to Probability, this is one of the main ways to break a joint distribution into something easier to read.
If X and Y are discrete random variables, you find the conditional distribution with P(Y = y | X = x) = P(X = x, Y = y) / P(X = x), as long as P(X = x) is not zero. That formula says: take the joint probability for the pair (x, y), then rescale by the total probability of X = x. The result is a new probability distribution for Y that depends on the chosen x.
For example, if a table shows the probabilities for two variables, you can focus on one row or one column once X is fixed. The probabilities for all possible Y values should add up to 1 within that condition. That is what makes it a real distribution, not just a list of ratios.
This is different from the marginal distribution of Y, which ignores X entirely. The marginal tells you how Y looks overall, while the conditional distribution tells you how Y looks after you restrict attention to one slice of the data. If the conditional distributions change a lot as X changes, then X and Y are related in a meaningful way. If they stay the same, that often suggests independence.
A common mistake is to treat P(Y | X) like one single number. It is usually a whole distribution, meaning a full set of probabilities for every possible Y value given a fixed X value. In graph form, you might see separate bars, rows, or curves for different X values so you can compare the shape of Y under each condition.
This term is one of the cleanest ways to see dependence between random variables in Intro to Probability. Instead of asking only what Y looks like overall, you ask what Y looks like after you know X, which is the kind of question probability is built to answer.
It also connects directly to the rest of joint distributions. Once you can move from a joint table to a conditional distribution, you can read information more efficiently, compare groups, and check whether one variable changes the chances for another. That shows up in problem sets where you are given a table and asked to fill in missing probabilities or describe the pattern.
Conditional distributions also set up later ideas like independence and Bayes' theorem. If knowing X does not change the distribution of Y, that is a strong hint that the variables are independent. If it does change, then the conditional distribution gives you a precise way to describe how.
In applications, this is the tool you use when you already know part of the story and want the rest of the probability model to update with that information. That is a core move in probability, not just a table trick.
Keep studying Intro to Probability Unit 10
Visual cheatsheet
view galleryJoint Distribution
The joint distribution gives the starting point for conditional distributions because it lists probabilities for pairs of values, not just one variable at a time. To get P(Y | X = x), you pull the relevant joint probabilities from the table or formula and divide by the marginal probability of X = x. Without the joint distribution, you usually cannot build the conditional one.
Marginal Distribution
Marginal distributions show the overall behavior of one random variable, ignoring the other variable in the joint setup. Conditional distributions do the opposite: they freeze X at a chosen value and look only at Y. Comparing the two helps you see whether conditioning changes the shape, center, or spread of the probabilities.
Independence
Independence means the distribution of Y does not change when you know X. In conditional distribution language, that means P(Y | X = x) matches the marginal distribution of Y for every allowed value of x. If the conditional distributions are different from the marginal, then the variables are not independent.
Interpretation of Conditional Probabilities
Conditional distribution is the broader version of conditional probability. A single conditional probability gives one outcome, like P(Y = y | X = x), while the conditional distribution gives the full list of probabilities for all Y values given that same X. That difference matters when a question asks for a whole pattern instead of one event.
A quiz or homework problem will usually give you a joint probability table, a two-way frequency table, or a formula and ask you to find the distribution of Y when X takes a specific value. The move is to isolate the row, column, or relevant cases for that X, divide each joint probability by P(X = x), and check that the probabilities add to 1.
You may also be asked to compare conditional distributions for different X values and say whether X and Y look independent. A good answer does not just compute numbers, it explains what changes in the shape of Y after conditioning. If the topic appears in a short response or discussion, describe the pattern directly, such as whether larger values of X make larger values of Y more likely.
A marginal distribution describes one variable by itself, with the other variable ignored. A conditional distribution describes one variable after fixing the other, so it changes depending on the value of X. If you mix them up, you will read the table the wrong way and miss the effect of conditioning.
The conditional distribution of Y given X is the probability distribution of Y after X has been fixed at one value.
For discrete variables, you find it by dividing each joint probability P(X = x, Y = y) by the marginal probability P(X = x).
A conditional distribution is a full distribution, so the probabilities for all possible Y values must add up to 1 for that chosen X.
If the conditional distribution changes when X changes, that suggests a dependence between the variables.
If the conditional distribution of Y given X matches the marginal distribution of Y, that is a sign of independence.
It is the distribution of Y after you know that X has a particular value. In a joint probability table, you take the probabilities in the slice where X is fixed and divide by P(X = x). The result shows how likely each Y value is under that condition.
Use P(Y = y | X = x) = P(X = x, Y = y) / P(X = x) for each possible value of Y. If you are using a table, find the row or column for the chosen X value, then divide every joint probability in that slice by the total for X. The conditional probabilities should sum to 1.
A marginal distribution ignores the other variable and describes one random variable overall. A conditional distribution keeps one variable fixed and shows how the other variable behaves under that condition. So marginal is the whole picture for one variable, while conditional is one slice of the joint distribution.
No. Conditioning only says the distribution of Y is being viewed after X is known. That can reveal association or dependence, but it does not prove cause and effect. You need more information about the situation before making a causal claim.