Statistics › Section C · Probability › Topic C.5
Tree and Venn diagrams
Rules compute probabilities; pictures organise them. The Venn diagram lays two events over one sample space, the tree diagram unfolds a multi-stage experiment into branches - and a drawn problem is half solved.
Two pictures for chance
Rules compute probabilities; pictures organise them. The Venn diagram lays two events over one sample space so that overlap logic becomes visible. The tree diagram unfolds a multi-stage experiment into branches, each labelled with a conditional probability. Between them, almost every problem in this chapter can be drawn before it is computed - and once the picture is right, the arithmetic is usually the easy part.
Venn diagrams
Draw a rectangle for the sample space and a circle for each event. With two events, the picture has four regions: in \(A\) only, in \(B\) only, in both (the overlap), and in neither. Fill the regions with counts or probabilities and every earlier construction becomes visual: the complement is outside a circle, the intersection is the overlap, the union is both circles together. Venn diagrams suit questions about how two events overlap - they are logic pictures, not data pictures.
Reading union and complement from a Venn diagram
The addition rule becomes a picture. Shading both circles shows \(A\) or \(B\) - but a plain sum shades the overlap twice, which is exactly why the rule subtracts \(P(A \text{ and } B)\) back out. The complement shades everything outside a circle: the rectangle minus the circle, which is the complement rule made visible. When a problem's words tangle - "not \(A\), or \(A\) without \(B\)" - sketching the regions first usually untangles them faster than algebra does.
Worked example: the survey as a Venn diagram
Place the survey of 200 students in the picture: both French and band, 24 in the overlap; French only, 56; band only, 36; neither, 84. Read directly: \(P(\text{French and band}) = \frac{24}{200} = 0.12\). \(P(\text{French or band}) = \frac{56+36+24}{200} = \frac{116}{200} = 0.58\), counting the overlap exactly once. \(P(\text{neither}) = \frac{84}{200} = 0.42\), the complement of the union, and \(0.58 + 0.42 = 1\). Check: the four regions total \(56+36+24+84 = 200\), the whole sample space.
Tree diagrams
A tree diagram unfolds an experiment into stages: from each stage's outcome, branches fan out for the next stage, and every branch is labelled with its probability given what came before - a conditional. Each complete path from root to tip is one way the experiment can end, and its probability is the product of its branches. The tree's two laws: multiply along a path, add across paths. Because branches carry conditionals, dependence is built in - without-replacement counts update themselves branch by branch.
Worked example: two marbles from the jar, drawn as a tree
Draw two marbles from the jar (4 green, 8 other), without replacement. Stage one branches: green \(\frac{4}{12}\), other \(\frac{8}{12}\). Stage two branches conditionally: after green, the jar holds 3 green and 8 other among 11, so \(\frac{3}{11}\) and \(\frac{8}{11}\); after other, \(\frac{4}{11}\) and \(\frac{7}{11}\). The four path products: \(\frac{4}{12} \times \frac{3}{11} = \frac{12}{132}\) (both green), \(\frac{4}{12} \times \frac{8}{11} = \frac{32}{132}\), \(\frac{8}{12} \times \frac{4}{11} = \frac{32}{132}\), and \(\frac{8}{12} \times \frac{7}{11} = \frac{56}{132}\). Check: the paths total \(\frac{12+32+32+56}{132} = \frac{132}{132} = 1\) - every draw lands somewhere.
Worked example: one green and one other
What is the probability of drawing exactly one green and one other in the two draws? 1. Two paths qualify: green then other, and other then green. 2. Each path has probability \(\frac{32}{132}\). 3. Add across paths: \(\frac{32}{132} + \frac{32}{132} = \frac{64}{132} = \frac{16}{33} \approx 0.485\). 4. Check: the three outcomes - both green \(\frac{12}{132}\), one of each \(\frac{64}{132}\), both other \(\frac{56}{132}\) - total \(\frac{132}{132} = 1\), and every path is accounted for exactly once.
Worked example: a diner's dessert tree
At a restaurant every diner orders exactly one starter: 30% choose soup and 70% salad. If soup, 40% then order cake, and if salad, 60% do. What share of diners order cake? 1. Paths to cake: soup then cake, \(0.30 \times 0.40 = 0.12\); salad then cake, \(0.70 \times 0.60 = 0.42\). 2. Add across paths: \(P(\text{cake}) = 0.12 + 0.42 = 0.54\). 3. Check: the no-cake paths are \(0.30 \times 0.60 = 0.18\) and \(0.70 \times 0.40 = 0.28\), totalling 0.46, and \(0.54 + 0.46 = 1\). The tree turns a tangled two-stage question into four small multiplications.
Choosing the picture
Match the picture to the structure. Two events at one stage, questions of overlap and exclusion: the Venn diagram, or the contingency table when the data arrive cross-classified. Stages in sequence - draws, tests, decisions - where each step's chances depend on the last: the tree diagram. All three pictures enforce the same discipline: account for every outcome exactly once, and make the total come to 1. A picture whose regions or paths sum to something else is not simplified; it is wrong.
Common mistakes to avoid
First, adding branch probabilities that should multiply: a path is a chain, and chains multiply. Second, multiplying path probabilities that should add: two different routes to an outcome both count. Third, forgetting to update later branches after a without-replacement draw - every branch after a green must use 11 marbles, not 12. Fourth, double-counting the overlap when reading a union from a Venn diagram. Fifth, accepting a tree whose path probabilities total something other than 1. Sixth, using a Venn diagram for a sequential problem: circles have no order, and stages need branches.
Key takeaways
Venn diagrams lay events over the sample space: four regions for two events, with the overlap meaning and, the combined circles meaning or, and the outside meaning not. The addition rule's subtraction is the overlap, seen. Tree diagrams unfold stages into branches labelled with conditional probabilities: multiply along a path for one outcome chain, add across paths for an overall probability, and all path probabilities together must total 1. Use Venn diagrams for overlap logic, tables for cross-classified counts, and trees for sequences where each stage depends on the last.
Flashcards
Flashcard
Events as circles inside the sample-space rectangle
Test your understanding
Question 1 of 5
In a Venn diagram of two events, the overlap of the two circles represents:
Question 2 of 5
A Venn diagram of 100 people shows A only 30, B only 20, both 10 and neither 40. What is P(A or B)?
Question 3 of 5
To find the probability of one complete path through a tree diagram:
Question 4 of 5
Every shopper pays by card (40%) or cash (60%). Of card payers 25% buy a drink; of cash payers 50% do. What is P(drink)?
Question 5 of 5
A bag holds 3 red and 5 blue marbles. Two are drawn without replacement. What is the probability of exactly one of each colour?
