COMM5000 Data Literacy for Business
The Complete Study & Assessment Guide · T3 2026COMM5000 Overview
- UNSW Sydney
- Postgraduate course
- Term 3, 2026
- Case study assessment
COMM5000 Data Literacy for Business develops statistical reasoning for business decisions through a connected case study based on a real business problem. The teaching sequence begins with data types, graphs, numerical summaries and bivariate relationships, then introduces probability, random variables and normal models.
- Assessment structure One connected case develops through staged analysis, feedback, proposal and reporting.
- Core method Move from descriptive evidence through inference and regression before making a recommendation.
- Claim boundary Separate association, prediction and causal evidence in every business conclusion.
- Final check Recompute every displayed number and audit privacy, fairness and review responsibilities.
What you get in the full COMM5000 Study & Assessment Guide
- 8 chapters, 36 pages — every assessed topic, week by week
- 18 original practice questions, each with a model answer
- 8 worked examples, solved the way COMM5000 marks
- 8 computed figures, redrawn from the course’s own numbers
- Before-you-submit checklist for the Case study business report (50%) — what markers reward, what costs marks
Free on this page: assessment table, chapter map, one worked example, key terms, FAQ. The rest is in the full study & assessment guide.
How COMM5000 is assessed
| Component | Weight | Format |
|---|---|---|
| Case study insight development (M1) | 20% | Individual case-study insight development; due Week 4, Friday at 5PM |
| Peer marking and feedback (M1) | 5% | Peer marking and feedback across Weeks 6 and 7 |
| Case study project proposal (M2) | 25% | Project proposal; due Week 7, Friday |
| Case study business report | 50% | Business report; due Week 10, Friday |
What other students are asking about COMM5000
- Where can I find lecture notes or study guides for COMM5000 at UNSW?
- The COMM5000 study guide in AskSia Library covers Data Literacy for Business in 8 chapters, with the assessment table, a chapter-by-chapter topic map, key terms and a worked example, free to read on this page.
- Is there a final exam in COMM5000?
- No final exam is listed in the offering this AskSia Library guide is built from; COMM5000 is assessed by Case study insight development (M1) 20%, Peer marking and feedback (M1) 5%, Case study project proposal (M2) 25%, Case study business report 50%.
- What does COMM5000 cover on probability and normal models?
- In AskSia Library's COMM5000 guide, the Probability and Normal Models chapter explains events, conditional probability, random variables, normal models and tail decisions.
- What does COMM5000 cover on sampling distributions and estimation?
- In AskSia Library's COMM5000 guide, the Sampling Distributions and Estimation chapter explains population parameters, sample statistics, standard errors and the central limit theorem.
- What does COMM5000 cover on confidence intervals and sample size?
- In AskSia Library's COMM5000 guide, the Confidence Intervals and Sample Size chapter explains interval interpretation, margins of error, precision trade-offs and survey planning.
What COMM5000 covers
The course moves from graphical and numerical summaries through probability, sampling, interval estimation and hypothesis testing, then applies simple and multiple regression before examining causality, privacy, governance and ethical data practice.
Describing Data for Business
Graphical and numerical summaries, covariance, correlation and defensible business interpretationLearn to classify variables, choose charts and numerical summaries, describe distribution shape, and interpret covariance and correlation without overstating what association can establish. Begin by naming the observational unit and target population.
- Observational units and variable types
- Graphical summaries
- Centre and spread
- Covariance and correlation
- Business interpretation
Probability and Normal Models
Events, conditional probability, random variables, normal models and tail decisionsTranslate business uncertainty into events, conditional probabilities, complements and normal-model tail areas while keeping denominators, assumptions and direction visible. Start by defining the experiment, sample space and event in ordinary language.
- Events and sample spaces
- Conditional probability
- Independence
- Random variables
- Normal standardisation
Sampling Distributions and Estimation
Population parameters, sample statistics, standard errors and the central limit theoremSeparate population parameters, sample statistics and sampling distributions, then use standard errors and the central limit theorem to describe estimator precision.
- Population and sample
- Point estimators
- Sampling distributions
- Standard errors
- Central limit theorem
Confidence Intervals and Sample Size
Interval interpretation, margins of error, precision trade-offs and survey planningConstruct and interpret confidence intervals, explain the long-run procedure, and plan sample size by balancing confidence, precision, variability and collection quality. An interval combines a point estimate with a margin of error.
- Point estimate
- Margin of error
- Confidence procedure
- Interval width
- Sample-size planning
Hypothesis Tests, Errors and Power
Null models, p-values, Type I and Type II errors, power and practical importanceFrame population claims as hypotheses, interpret p-values conditionally, distinguish testing errors, and connect power and practical importance to business decisions.
- Null and alternative hypotheses
- Test statistics
- P-values
- Type I and Type II errors
- Power and materiality
Simple Regression and Prediction
Slope interpretation, residuals, coefficient inference and prediction intervalsInterpret simple-regression coefficients with units, examine residuals, separate coefficient inference from prediction, and avoid extrapolation and unsupported causal language.
- Fitted line
- Slope and intercept
- Residual diagnostics
- Coefficient inference
- Mean and individual prediction
Multiple Regression and Robustness
Ceteris paribus effects, omitted variables, dummy variables and specification checksInterpret ceteris paribus coefficients, reason about omitted variables and indicators, and design robustness checks that challenge rather than decorate a model.
- Conditional coefficients
- Confounding and endogeneity
- Dummy variables
- Nonlinearities and multicollinearity
- Robustness analysis
Causality, Privacy and Data Ethics
Counterfactual reasoning, governance, fairness, responsible communication and reviewDistinguish association from intervention effects, evaluate identification assumptions, and connect privacy, governance, fairness and monitoring to responsible business action.
- Counterfactual questions
- Randomisation and confounding
- Privacy and data minimisation
- Fairness and accountability
- Monitoring and review
Sampling distributions provide the bridge from samples to populations, followed by point estimation, confidence intervals, sample-size planning and hypothesis tests. The later blocks use simple and multiple regression for explanation and prediction, examine omitted variables and robustness, then close with causal reasoning, privacy, governance and ethical statistical practice.
The current Term 3, 2026 Moodle syllabus publishes four graded rows that sum to 100%. Case study insight development is worth 20% and asks students to turn early descriptive and inferential work into a supported insight. Peer marking and feedback contributes 5%, so evaluating another analysis and responding to feedback are part of the assessed sequence.
The 25% project proposal requires a coherent question, method and evidence plan. The 50% business report is the major task and tests whether the whole chain from data and model to interpretation and recommendation remains defensible. The 2026 UNSW Handbook groups the first two Moodle rows into one Milestone 1 project, producing the same 25%, 25%, 50% overall allocation.
Across the course, a calculation is only one part of the answer. Define the population and variables, select a method that fits the question, show the arithmetic, interpret uncertainty and end with a recommendation whose strength matches the evidence. Descriptive patterns should not become causal claims by wording alone.
Predictions need a defined deployment population and range, while intervention claims need a credible comparison design. Use seminar questions to rehearse operations and solutions to diagnose the step that failed rather than memorising a final number. Keep one decision log for each analysis: question, population, data period, assumptions, estimate, uncertainty, practical meaning, limitation and next check.
That record supports consistent writing across insight development, peer feedback, the proposal and the final report. Before submission, reconcile spreadsheets, figures and prose so the same result has the same unit, sample size and interpretation everywhere. Review privacy, data minimisation, fairness, accountability and monitoring as parts of the analysis, not as an appendix added after the recommendation.
Worked example · free
A confidence interval for a customer proportion
- +1The four marks used here are not an official university marking scheme. Compute the sample proportion: 236/400 = 0.59.
- +1Compute the standard error: sqrt(0.59 x 0.41/400) = 0.02459.
- +1Use margin 1.96 x 0.02459 = 0.04820 and form 0.59 plus or minus 0.04820.
- +1Report about 54.2% to 63.8% for the active-customer population under the sampling assumptions.
Key terms
- Sampling distribution
- The probability distribution of a statistic over repeated samples drawn by the same method.
- Standard error
- The repeated-sample standard deviation of an estimator and a measure of its precision.
- Confidence interval
- A range produced by a calibrated procedure to estimate an unknown population parameter.
- P-value
- The probability, under the null model, of a result at least as extreme as the observation in the alternative's direction.
- Statistical power
- The probability that a test rejects the null for a specified alternative effect.
- Regression residual
- The vertical difference between an observed outcome and its value predicted by the fitted model.
- Omitted-variable bias
- Coefficient distortion that can arise when an excluded factor affects the outcome and is related to an included regressor.
- Causal effect
- A contrast between potential outcomes under different interventions for the target population.
- Contingency table
- A contingency table cross-classifies categorical variables and displays the cell counts used to obtain joint, marginal and conditional distributions.
- Central limit theorem
- The central limit theorem supports an approximately normal sampling distribution for a sample mean or proportion under suitable conditions as the sample size grows.
- Margin of error
- The margin of error is the critical value multiplied by the standard error and forms the distance from a point estimate to either confidence-interval bound.
- Type I error
- A Type I error occurs when a decision rule rejects a null hypothesis that is actually true, with probability set by the test's significance level.
- Type II error
- A Type II error occurs when a test fails to reject a false null hypothesis, so a real effect relevant to the business decision is missed.
- Ceteris paribus
- Ceteris paribus is the modelling condition that interprets a relationship while holding other relevant influences constant.
COMM5000 FAQ
How difficult is the course?
The difficulty comes from linking calculations to defensible business claims, not from arithmetic alone. The weekly sequence moves from descriptive summaries and probability into sampling distributions, confidence intervals, hypothesis tests and regression. Later work also asks you to separate association from causality and consider ethics.
Steady practice with units, assumptions, uncertainty and interpretation is therefore more useful than memorising isolated formulas.
How is the assessment organised?
The Term 3 Moodle syllabus separates one connected case into four graded tasks: 20% case study insight development, 5% peer marking and feedback, a 25% case study project proposal, and a 50% business report. Those published rows total 100%. The sequence means early analysis and feedback should feed the proposal and report, so keep a consistent record of data choices, calculations, assumptions and revisions.
Which statistics should I practise first?
Begin with variable types, graphical summaries, centre and spread, covariance and correlation, then practise conditional probability and normal models. Next, distinguish population parameters from sample statistics and learn how a sampling distribution creates a standard error.
Those foundations determine whether later confidence intervals, hypothesis tests and regression conclusions are set up correctly and whether the final business interpretation matches the quantity calculated.
What matters when interpreting regression?
Print the units for both the outcome and each regressor, then interpret a coefficient as the fitted change associated with a one-unit regressor change while other included variables are held constant. Distinguish inference about a coefficient from prediction of a mean or one future outcome.
Check residual behaviour, the observed range, omitted-variable risk and out-of-sample performance before turning the model into a recommendation.
How should causal claims be checked?
Ask which intervention and comparison define the claim, whether the groups are comparable, and which common causes, selection mechanisms or timing problems remain. The course materials warn that correlation is a relationship rather than automatic evidence of cause and effect. Regression adjustment can help with included confounders, but it does not by itself turn observational data into a credible intervention effect.
What should I check before submitting?
Recompute every displayed number and confirm that tables, figures and prose use the same sample, units and rounding. State the population, method, assumptions, uncertainty and practical limitation beside the recommendation.
Then confirm the current Moodle task instructions and audit the analysis for legitimate purpose, data minimisation, access, retention, fairness, transparency, monitoring and accountability before the report is finalised.
What does the 50% business report test?
The business report is the largest graded task in the connected case study. The course sequence indicates that it should bring together descriptive evidence, statistical inference, regression and an appropriately bounded recommendation for a real business problem.
A strong report keeps the question, population, variables, model, uncertainty and decision aligned, and it distinguishes patterns or predictions from causal effects that require stronger design evidence.
Do the diagnostic lessons, formative quizzes or GMetrix affect the grade?
The Moodle syllabus describes the diagnostic lessons, formative quizzes and GMetrix activities as non-graded learning aids. It says the diagnostic lessons check prerequisite mathematics, the quizzes provide formative self-assessment, and GMetrix supports Excel training.
The same section states that these activities do not contribute to the course grade and are not mandatory hurdle tasks, although students are strongly encouraged to use them.
Questions students ask in each COMM5000 chapter
- When should I use the median?
- Use the median when skew or extreme observations make the mean unrepresentative of the middle case. Report the unusual observations separately when they remain important to the business decision. Chapter 01 · Describing Data for Business
- How do I choose the denominator?
- For conditional probability, restrict the data to the group named after given and divide by that group total. For a joint probability, use the full relevant sample. Chapter 02 · Probability and Normal Models
- What does the central limit theorem say?
- Under suitable conditions, the sampling distribution of the mean becomes approximately normal as sample size grows. It does not claim that the individual observations become normal. Chapter 03 · Sampling Distributions and Estimation
- What does 95% confidence mean?
- Across repeated samples analysed with the same valid procedure, about 95% of the resulting intervals contain the fixed population parameter. It is not a probability statement about one completed interval. Chapter 04 · Confidence Intervals and Sample Size
- Does a large p-value prove no effect?
- No. It means the data do not provide sufficient evidence against the null under the stated test. The confidence interval shows which effect sizes remain compatible with the data. Chapter 05 · Hypothesis Tests, Errors and Power
- When is the intercept meaningful?
- The intercept is practically meaningful when zero is sensible and represented by the data. Otherwise it remains part of the fitted equation but should not receive an unsupported substantive story. Chapter 06 · Simple Regression and Prediction
- Should I add every available control?
- No. Controls should follow the question and causal logic. Mediators and colliders can distort the target comparison, while irrelevant variables can add noise or reduce interpretability. Chapter 07 · Multiple Regression and Robustness
- Can regression prove causality?
- Regression can adjust for included variables, but causal interpretation requires a credible design and assumptions that address confounding, selection, timing, measurement and the target intervention. Chapter 08 · Causality, Privacy and Data Ethics
How to prepare for the assessments
Work in the same sequence as the course and the connected assessment. In Weeks 1 and 2, build a reusable description sheet for variable types, graphs, centre, spread, covariance, correlation, conditional probability and normal tails. Recalculate seminar examples without looking at the solution, then explain what each result means for the case rather than stopping at the number.
In Weeks 3 and 4, separate population, sample and sampling distribution on paper before using a formula. Practise standard errors, confidence intervals and sample-size trade-offs, because the 20% insight-development task needs both a numerical result and a defensible interpretation.
When peer marking begins, use the 5% feedback stage to test whether another reader can reconstruct your denominator, units, assumptions and claim boundary. For the 25% proposal, write the business question first, then name the target population, variables, method, uncertainty and evidence needed for the recommendation.
In the hypothesis-testing and regression blocks, keep a decision table that distinguishes statistical significance, practical importance, association, prediction and causal evidence. Check residuals, observed ranges, omitted variables and robustness before treating a model as useful beyond its sample.
For the 50% report, trace every recommendation backward through the written claim, numerical result, model, variables and sampling process. Reconcile spreadsheet outputs, figures and prose, and finish with privacy, fairness, transparency, monitoring and accountability checks. Use the current Moodle task page as the authority for submission requirements and permitted assistance.
Your AI Statistics tutor for COMM5000
Stuck on a hard COMM5000 question? Sia is AskSia’s AI Statistics tutor — ask any COMM5000 Data Literacy for Business question and get a clear, step-by-step explanation grounded in how the course is actually taught and assessed. Read this whole study guide free, then take your hardest questions to Sia.
