University of Sydney · FACULTY OF STATISTICS

BUSS1020 Chap.1 Data, Variables and Sampling

- one subject, every graph, every model, every mark
6 Chapters3-page Bible
Our own words - no uploaded lecturer files
Updated for this semester
Chapter 1 of 10 · BUSS1020

Data, Variables and Sampling

Statistics starts with a decision, a target population and a defensible way to observe it. This chapter distinguishes descriptive from inferential statistics, classifies qualitative and quantitative variables by measurement scale, and separates parameters from statistics.

It then compares probability sampling designs with convenience and judgement methods, showing why a large sample cannot repair systematic exclusion or non-response. The worked example designs a regional service survey and limits the claims made from a convenience sample.

A second theme is the observational unit: a transaction, customer, store and region support different sample sizes and conclusions even when their values sit in one spreadsheet. Coverage, non-response, response and sampling error are separated because each has a different repair. The final drills ask you to turn a vague bias warning into a named mechanism, a likely direction and a concrete redesign.

That reasoning prepares the ground for every later interval and test, whose formula quantifies sampling uncertainty but cannot certify a flawed frame. A short classification exercise also links variable scale to a suitable summary, ensuring the analysis does not treat ordered categories as if their gaps were measured exactly.

The closing claim audit makes the evidence boundary explicit before later formulas add apparent precision. Rehearse that boundary until every conclusion identifies who was observed, who was not and which selection rule supports the stated scope.

In this chapter

What this chapter covers

  • 01

    Descriptive and inferential statistics

  • 02

    Populations, samples, parameters and statistics

  • 03

    Qualitative and quantitative variables

  • 04

    Nominal, ordinal, interval and ratio scales

  • 05

    Simple random, systematic, stratified and cluster samples

  • 06

    Coverage, non-response, response and sampling error

Worked example · free

Design a regional service survey

Q [5 marks]. A retailer operates stores in three regions but has feedback only from customers at one busy city store. Propose a sampling design that supports a regional service estimate and limit the claim from the existing feedback. This is an AskSia-authored drill; the mark allocation shown here is illustrative and not an official university assessment scheme.
The arrows show the inferential chain: a defensible sampling design links a population to a sample, a statistic and finally a business decision.The arrows show the inferential chain: a defensible sampling design links a population to a sample, a statistic and finally a business decision.PopulationSampleStatisticDecisiondesigncalculateinterpret
The arrows show the inferential chain: a defensible sampling design links a population to a sample, a statistic and finally a business decision.
  • 1Define the target population as service experiences across every store and the stated time period.
  • 1Use the three regions as strata and select stores, time slots and customers with known random rules.
  • 1Define the service-time variable consistently so every observer uses the same start and end points.
  • 1Treat the old feedback as descriptive of its respondents because inclusion probabilities for all customers are unknown.
  • 1Report the new sample estimate with its unit and a later uncertainty measure that respects the design.
A stratified probability sample supports representation from every region. The existing convenience responses describe their respondents but do not justify a probability margin of error for all customers.
Sia tip — Write the target population, sampling frame and selection rule before claiming that a sample is representative.
Glossary

Key terms

Population
The complete set of units or outcomes named by the business question.
Sampling frame
The operational list or mechanism from which units can actually be selected.
Parameter
A numerical feature of a population that sample evidence seeks to estimate.
Statistic
A numerical summary computed from an observed sample.
Coverage error
Systematic exclusion of part of the target population from the sampling process.
FAQ

Data, Variables and Sampling FAQ

Why does measurement scale matter?

The scale determines which comparisons and summaries preserve meaning. An ordinal code has an order but not guaranteed equal gaps, while a ratio measurement has equal units and a meaningful zero. Treating every code as ratio data can manufacture precision.

Does a large sample remove bias?

A large sample reduces random sampling variation under a suitable design, but it does not repair a frame that excludes part of the population or responses that systematically differ from non-responses. Design quality and sample size solve different problems.

When is stratified sampling useful?

Stratification is useful when named subgroups must appear in the sample or when they are internally similar but differ from one another. Sampling within each stratum can improve representation and precision, provided the final analysis respects the allocation.

Study strategy

Exam move

Practise each new dataset with a four-line preface: observational unit, variable, measurement scale and target population. Then identify the sampling frame and selection rule. For every claim, mark whether it is descriptive of observed cases or inferential about a wider population. Build a table of sampling designs with one business use and one failure mode for each.

When a question mentions bias, name the mechanism rather than writing only that the sample is unrepresentative. Finish by stating how the design could be changed and what wider claim the repair would support. Create your own examples of nominal, ordinal, interval and ratio variables from one business, then explain what arithmetic becomes meaningful at each scale.

Recode the same customer feedback as categories and as a numerical score to see how easy it is to create false precision. For sampling methods, sketch the population as a box and mark exactly how units enter a simple random, systematic, stratified or cluster sample. Add a sentence about cost and a sentence about inferential strength.

Rehearse the distinction between a population parameter and sample statistic aloud until the wording remains stable when symbols change. When reviewing a survey claim, inspect coverage before sample size, because thousands of responses from the wrong frame do not answer the target question.

Use workshop solutions to check the order of reasoning, but rebuild the scenario with different units so you are learning the method rather than a stem. End a revision session by rewriting one overconfident claim into a scoped conclusion that names respondents, excluded groups and the additional design evidence needed for generalisation.

Working through Data, Variables and Sampling in BUSS1020? Sia is AskSia’s AI Statistics tutor — ask any BUSS1020 Data, Variables and Sampling question and get a clear, step-by-step explanation grounded in how BUSS1020 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.

A+Everything unlocked
Unlocks this Bible + all 202 of your subjects at University of Sydney, plus 1,000+ more Bibles across the AskSia library.
Sia - your BUSS1020 tutor, unlimited, worked the way the exam marks it
The full 3-page Bible + practice bank with worked solutions
Chrome extension - sync your LMS so Sia knows your deadlines
Bilingual EN / Chinese on every Bible and every Sia answer
$0.99 Trial
30-day money-back · cancel in one tap · how it works