INST0098 Chap.6 Data Types and Survey Design
Data Types and Survey Design
Data Types and Survey Design explains how constructs become variables, question items, responses and claims. Data characteristics constrain the analyses that remain meaningful, while sampling and nonresponse constrain inference.
The chapter covers measurement levels, balanced response options, piloting, missing values and the difference between a numeric code and a genuinely quantitative scale.
Start with an intended conclusion and work backward. Name the target population, the construct and the observable indicator. Then ask whether one survey response can genuinely represent that indicator.
Trust, access and engagement are broad constructs that may need several dimensions or a task-specific definition. Write rival operationalisations and explain what each includes or leaves out before settling on an item.
Review the data type independently of its storage format. A number may label a category, preserve an order or represent equal intervals.
The permitted comparison follows the measurement meaning, not the spreadsheet column type. Create a data dictionary before collection, defining variables, allowable values, missing codes, multiple responses and any derived measures. This prevents silent reinterpretation once attractive patterns appear.
Question wording should minimise avoidable ambiguity.
Use one concept per item, define the relevant time period and avoid assumptions about experience. Balance response options and include a legitimate non-applicable route where needed. Check order effects by asking whether an earlier item supplies a frame for a later answer.
On small screens, inspect whether labels remain visible and whether long grids create careless straight-line responses.
Create a response-option audit using imagined participants at the edge of the sampling frame. Test whether each person can answer accurately without accepting a false premise. Examine whether neutral, unsure and not-applicable positions are conceptually different.
Check that labels remain symmetrical and understandable without relying on numbers alone. When an option is added, reconsider the analysis because a measurement improvement can require a new coding decision.
Pilot for interpretation, not merely technical success. Ask pilot participants to explain what they believed each item meant and how they chose an answer.
Record missing options, unclear terms and points where they stopped. Revise the instrument and keep a version trail. If a change alters the construct, do not combine responses across versions as though measurement remained constant.
Sampling analysis continues after launch. Draw the path from target population to frame, invitation, start and completion.
Compare available characteristics at each stage and discuss who may be absent. A large convenience sample can remain systematically narrow. In reporting, name the reachable population, recruitment route and response pattern. Treat nonresponse as a possible source of bias linked to the topic, not simply a lower count.
Archive the questionnaire, recruitment wording and data dictionary together.
A reader should connect every reported variable to the exact item and response route that produced it.
What this chapter covers
- 01
Data characteristics
- 02
Construct measurement
- 03
Question wording
- 04
Response scales
- 05
Sampling frames
- 06
Nonresponse analysis
Worked example · free
Operationalise archive confidence
- 2Define confidence for a specified archive task.
- 2Use a balanced item and an optional explanation after the task.
- 2Report recruitment and limit inference to reachable respondents.
Key terms
- Construct
- An abstract concept that a research design translates into observable indicators.
- Sampling frame
- The accessible set from which survey participants can be recruited or selected.
- Nonresponse bias
- Potential distortion when participation relates to the subject being measured.
Data Types and Survey Design FAQ
Does coding an answer as a number make it quantitative?
No. Numbers can label categories or order responses without establishing equal intervals or meaningful ratios. The interpretation depends on the measurement design and its assumptions about comparison.
What should a survey pilot test?
A pilot should examine how respondents interpret items, whether options fit, how navigation works and where participants stop. Successful form submission alone is not enough.
Why report recruitment details?
Recruitment explains who had a chance to respond and which groups may be absent. It helps readers judge whether the final claim reaches beyond the observed sample or the accessible sampling frame.
Assessment move
Take one broad construct and write three competing operational definitions. For each definition, propose an observable indicator and a question item. Label the response level and list the analyses it permits. Read every item from the position of a respondent whose experience differs from your own, noting forced assumptions and missing options.
Draw the target population, sampling frame, invitations, starts and completed responses as separate boxes. Practise writing a conclusion that names the reachable population rather than an imagined universal audience. End by building a data dictionary with variable names, values, missing codes and derivation rules so analysis decisions are visible before collection.
Working through Data Types and Survey Design in INST0098? Sia is AskSia’s AI Arts and Humanities tutor — ask any INST0098 Data Types and Survey Design question and get a clear, step-by-step explanation grounded in how INST0098 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.