BUSINFO 705 Chap.2 Probability Evidence and Distribution Summaries
Probability Evidence and Distribution Summaries
This chapter builds probability evidence from experiments, outcomes and observed frequencies. It distinguishes discrete counts from continuous measurements and checks that every discrete probability is nonnegative and that total mass equals one. A fresh urgent-call distribution is then used to compute expected value, variance and standard deviation, with attention to units and interpretation.
The closing drills separate exact probabilities, tail events and unsupported claims about future processes. The denominator, support and observation period remain visible because those details determine what an empirical probability can represent. Summary measures are used as decision evidence, not substitutes for the full distribution.
What this chapter covers
- 01
Experiments, outcomes and sample spaces
- 02
Classical, empirical and subjective probability
- 03
Discrete and continuous random variables
- 04
Valid probability mass functions
- 05
Expected value as a weighted mean
- 06
Variance and standard deviation
Worked example · free
Summarise an urgent-call distribution
- 1Check that the four probabilities total 1.00.
- 2Compute E(X)=1(0.10)+2(0.25)+3(0.40)+4(0.25)=2.80 calls.
- 2Compute Σ(x-2.80)²P(x)=0.86 calls squared.
- 2Take the square root: standard deviation is about 0.93 calls.
Key terms
- Sample space
- The full set of outcomes considered possible for an experiment.
- Random variable
- A numerical value assigned to the outcome of an uncertain process.
- Empirical distribution
- A probability distribution built from observed relative frequencies.
- Expected value
- The probability-weighted long-run centre of a random variable.
- Variance
- The probability-weighted average squared deviation from the mean.
- Standard deviation
- The square root of variance, expressed in the variable's original unit.
Probability Evidence and Distribution Summaries FAQ
What makes a discrete distribution valid?
Every assigned probability must be at least zero, no probability may exceed one, and all probabilities over the complete support must sum to one. Also check that the listed support contains every value treated as possible by the model.
What does expected value mean?
Expected value is the probability-weighted centre that would emerge across many repetitions under a stable process. It need not be an attainable single outcome. For a count distribution, a non-integer expected value is still meaningful as a long-run average.
Why report standard deviation with the mean?
The mean gives location while standard deviation gives the typical scale of movement around it. Two alternatives can have the same expected result but very different variability, which may change a risk-sensitive business decision.
How are exact and upper-tail events different?
An exact event selects one value from a discrete support. An upper-tail event includes every supported value at or above its boundary, depending on the wording. List or mark the included outcomes before adding probabilities so a correct distribution is not used to answer the wrong event.
When can observed frequencies represent future probabilities?
The interpretation is strongest when the population, measurement rule, observation period and operating process remain comparable. A policy change, seasonal shift or different customer mix can weaken that link. Record the scope of the data and update the empirical distribution when the process changes materially.
Assessment move
Start every probability question by writing the experiment, random variable and support. For empirical distributions, convert each frequency using the same denominator and confirm the probabilities total one. Memorise the three summary relations, but calculate through a table with columns for x, P(x), xP(x) and the squared-deviation contribution. Keep extra precision until the final square root.
When comparing alternatives, state what the mean and standard deviation each add to the decision and what shape information remains hidden. Practise one exact event, one cumulative event and one upper-tail event so the event notation becomes automatic. Build one deliberately invalid probability table and diagnose whether it has a negative entry, incomplete support or a total other than one.
Repair it before calculating a summary. For each valid table, identify the row that contributes most to the mean and the row that contributes most to variance; these need not be the same. Rewrite “at least,” “more than,” “no more than” and “exactly” as included support values before adding probability mass. Change one rare outcome to a more severe value and observe how expected value and variance respond.
Compare two distributions with equal means but different spreads, then explain what additional tail probability a manager would need. Keep the sample denominator, branch, season and observation period beside any empirical estimate. Write a short challenge to the stability assumption by naming a policy or population change that would make those frequencies less relevant.
Finish every numerical result with the original unit and horizon, and explicitly state that a long-run centre does not guarantee the next observed value.
Working through Probability Evidence and Distribution Summaries in BUSINFO 705? Sia is AskSia’s AI Management tutor — ask any BUSINFO 705 Probability Evidence and Distribution Summaries question and get a clear, step-by-step explanation grounded in how BUSINFO 705 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.