MKF2121 Chap.2 Conclusive and Descriptive Research Design
Conclusive and Descriptive Research Design
Conclusive research uses formal designs to answer specified questions or hypotheses. This chapter connects analytical models and variable roles to cross-sectional and longitudinal description, surveys, observation, measurement and pretesting. The key test is whether the promised conclusion is no stronger than the time structure, population route and comparison that produced it.
What this chapter covers
- 01
Exploratory, conclusive, descriptive and causal design
- 02
Verbal, graphical and mathematical models
- 03
Predictors, outcomes, controls, mediators and moderators
- 04
Cross-sectional and longitudinal evidence
- 05
Survey modes, coverage and mode effects
- 06
Structured and unstructured observation
- 07
Constructs, operational definitions and scale properties
- 08
Question comprehension, retrieval and response mapping
- 09
Cognitive pretesting, piloting and version control
Choose a mixed survey route
- 2Define the shared constructs and population before choosing modes.
- 2Use harmonised online and assisted routes while recording mode and accessibility.
- 2Pretest each path, monitor coverage and state the remaining response boundary.
Key terms
- Conclusive research
- A formal design used to evaluate specified questions or hypotheses.
- Descriptive research
- A design used to estimate characteristics, group differences and associations.
- Cross-sectional design
- A design collecting a bounded snapshot from one or more samples.
- Longitudinal design
- A design observing the same or comparable units repeatedly over time.
- Operational definition
- A rule stating how a construct is represented through observable indicators.
- Mode effect
- A response difference produced by collection conditions rather than the underlying construct.
Conclusive and Descriptive Research Design FAQ
How does a descriptive design differ from a causal design?
A descriptive design estimates characteristics, differences or associations as observed. A causal design manipulates a treatment and protects a counterfactual comparison so that an outcome difference can be attributed more credibly to the treatment.
Why build an analytical model before writing questions?
The model identifies proposed variables, directions, mechanisms and alternatives. It reveals which constructs need measurement and which comparisons answer the research problem, preventing the questionnaire from becoming an unplanned catalogue of interesting items.
What does questionnaire pretesting add?
Cognitive probing reveals how intended respondents interpret terms, retrieve experiences and choose options. Piloting tests routing, recruitment, devices, completion time and data capture so defects can be corrected before the main collection.
Assessment move
Draw the design taxonomy from memory and attach an inference verb to each branch. Practise turning a verbal mechanism into a graphical model, then name each variable role and a plausible alternative explanation. For survey and observation scenarios, identify coverage, measurement and ethical consequences instead of naming a mode alone.
Use a traceability matrix to link every research question to population, construct, indicator, collection route and planned analysis. End by writing the strongest conclusion the proposed design can honestly support. Revise analytical approaches in three passes. First describe the proposed relationships verbally, including why each link should exist. Next draw the graph and read every arrow as a directional sentence.
Finally write a simple symbolic form that identifies the outcome and predictors without pretending the equation proves causality. Add one rival model and note which additional variable or comparison would help distinguish it. For descriptive designs, take one question and write both a cross-sectional and longitudinal route. Compare the estimand, time claim, attrition risk and collection cost.
Practise separating within-unit change from a shift in sample composition. For survey modes, make a reach table covering who can participate, the answering conditions, likely mode effects, privacy and accessibility. For observation, define a coding rule that another observer could apply, then list motives the behaviour does not reveal.
Build questionnaire drills by repairing leading, double-barrelled, ambiguous and unrealistic-recall items. Specify the reference episode, period, balanced response options and not-applicable route. Ask how a respondent would comprehend, retrieve, judge and map the answer. Then create a routing diagram and test every eligible and ineligible path.
Finish with a pretest script that probes interpretation rather than asking whether an item was clear. Record a pilot version, completion time by path, break-off points, missingness and device problems. Close the exercise by writing the exact claim the revised instrument can support and one stronger claim it still cannot establish. Add a final claim audit after the blueprint is complete.
Circle every verb that implies cause, change or generalisation, then trace it back to assignment, time structure and population coverage. Rewrite any verb the design cannot carry. Check that each survey route measures the same construct and that observation conclusions describe behaviour rather than private motive. This review converts a plausible methods section into a defensible inference plan.
Working through Conclusive and Descriptive Research Design in MKF2121? Sia is AskSia’s AI Marketing tutor — ask any MKF2121 Conclusive and Descriptive Research Design question and get a clear, step-by-step explanation grounded in how MKF2121 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
