Five problems on how samples drawn from 1,180 known accounts behave, each worked by rule and checked by simulation, form the Unit 3 probability problem set in GB701. Searches like "gb 701 unit 3 assignment example", "gb701 unit 3 sample" and "gb701 unit 3 example" land here.
What a finished GB701 Unit 3 probability problem set looks like
Five numbered problems over about six pages, each in the same frame: the question, the rule or distribution chosen, the working, a simulation check where one applies, and a sentence for the sampling plan. Problem 1 treats late payment, beyond 60 days, as a binomial event with p of 0.107 and finds that a sample of 25 accounts has a 5.9 percent chance of containing no late payer at all. Problem 2 builds a table of terms by lateness and shows that 57.9 percent of late accounts are on net-45 terms, which hold only 24.2 percent of the book. Problems 3 and 4 trace the sampling distributions of the mean and the median. Problem 5 asks how often a sample of 100 overstates the late share, and by how much.
How a GB701 Unit 3 example is structured
Each problem was chosen to answer a question the Unit 4 memo will face, and the set says so, which keeps probability from reading as a detached drill. The binomial problem warns that small samples can miss the late tail entirely. The conditional table shows why sampling must be stratified by terms. Problem 3 is the center of the set: with the population standard deviation known, the standard error of a mean from 10 accounts is 4.75 days and from 40 accounts 2.35, both with the finite population correction applied. Simulated sample means for n of 10 retain a skew of 0.34, which falls to 0.13 at n of 40, so the central limit theorem is shown converging rather than assumed. Problem 4 finds the median noisier than the mean, and Problem 5 catches the normal approximation understating a skewed tail.
Late payment as a binomial event
With 10.7 percent of accounts past 60 days, a sample of 25 is expected to hold 2.67 late payers. The chance of six or more is 4.4 percent and of none 5.9 percent. Both tails matter, since a sample with no late accounts would silently drop the costliest region of the variable.
Conditional probability from the ledger
A two-way table of 1,180 accounts shows 25.6 percent of net-45 accounts late against 5.9 percent of net-30. Reversed, 57.9 percent of late accounts sit on net-45 terms. The problem separates the two conditionals explicitly, because confusing them is the error such prompts commonly test.
Sampling distribution of the mean
Theory and simulation stand side by side for samples of 10 and 40. The chance that a sample mean exceeds 45 days is 16.8 percent by simulation for n of 10 against 17.0 percent from the normal curve, and 2.9 against 2.6 percent at n of 40.
A noisier median
Across 10,000 samples of 40, sample medians have a standard deviation of 2.81 days, compared with 2.35 for sample means. The robust statistic costs precision in this population, a trade-off the set records for Unit 5, where a bootstrap interval for the median is planned.
Where the normal approximation slips
For samples of 100, the normal curve puts the chance of a late share of 20 percent or more at 0.14 percent; simulation finds 0.40 percent. Small in absolute terms, the gap is nearly threefold, and the problem explains that skewed binomial tails are where approximation deserves least trust.
Where marks go in GB701 Unit 3
Probability sets at doctoral level in GB701 are graded less on reaching the right number than on choosing the right model and saying why. Applying the binomial without checking that draws are close to independent, or ignoring the finite population correction when the sample is a large share of the frame, draws a margin note. Confusing the probability that a late account is on net-45 terms with the probability that a net-45 account is late is a frequent conceptual slip. Results left as bare decimals, with no sentence on what they imply, earn partial credit at best. Instructors often reward simulation used as a check on theory, provided the seed and the number of replications are reported so the check can be repeated. Problems disconnected from any later decision read as exercises copied out of a textbook.
Get a GB701 Unit 3 example written to your instructions
Which problems were set, and what data do they draw on? Send both with the Unit 3 rubric and a custom set is prepared in 24-48h, each model named and justified, working shown and every result read in a plain sentence. First samples are free, and simulation code is included where your instructions allow.
GB701 Unit 3 questions, answered
Do GB701 problem sets require simulation?
Not always. Many sections accept analytic solutions alone, and the instructions decide. Simulation earns its place when an approximation is in doubt, as with skewed data or small samples, because it shows whether theory and reality agree. Where it is used, reporting the seed, the number of replications and the software makes the check reproducible, which doctoral readers expect.
Is Excel acceptable for GB701 probability work?
Generally yes. BINOM.DIST, NORM.DIST and a short resampling macro cover everything in a set like this one, and plenty of sections are built around a spreadsheet. Output from R or SPSS is accepted just as readily. What instructors check is whether the function matches the model named in the text and whether arguments such as cumulative or exact were set correctly.
Can the problems use a textbook dataset instead?
Usually, if the prompt permits it. The advantage of a workplace extract is continuity: the same variable can carry through estimation, testing and regression, so later units build on earlier findings. A textbook set works equally well for the mechanics, though the discussion of what each result means for a real decision tends to come out thinner.