DATA4207 · Data Analysis in the Social Sciences
Conclusion and Delivering the Research Project
Week 13 pulls the unit's methods together into a delivery playbook for the individual Research Project — the 70% capstone (20% plan + 50% report) that is a 40%-combined hurdle to pass. It walks the full arc: choosing one of the four provided questions, sourcing extra data, running an appropriate analysis in R Markdown, and writing it up to a high standard with the Research Project Outline checklist. The point is to make the weighting and hurdle concrete so you spend effort where the marks are.
What this chapter covers
- 01Synthesis of the unit: theory (subject matter), uncertainty (bias, causality, confounders), communication
- 02Reproducible research revisited: R Markdown as transparent, verifiable, shareable analysis
- 03Choosing 1 of 4 provided research questions and sourcing an additional dataset
- 04Theory-driven variable selection — do not just maximise R² or rely on a selection algorithm
- 05Considering confounders and controls; distinguishing a substantive predictor from a mere predictor
- 06Justifying the model choice (linear / logistic / other — no correct default, but wrong answers exist)
- 07Report structure and requirements: PDF knitted from RMD, code hidden, high-quality ggplot and kable tables
- 08Assessment weighting: Research Plan 20% + Research Project 50%, a 40%-combined hurdle to pass
Assembling the individual Research Project
- +1Choose and scope: pick one of the four provided questions and its dataset, and source the required additional dataset. Frame a sharp, falsifiable hypothesis ('X causes Y because...') grounded in — but not a list of — the literature.
- +2Select variables by theory: choose predictors because a mechanism justifies them, not to maximise R² or on a selection algorithm alone; identify confounders to include as controls and distinguish substantive predictors from mere predictors.
- +2Analyse appropriately in R Markdown: match the model to the outcome (linear, logistic, or other) and justify it — there is no correct default, but there are wrong choices — and present findings with high-quality ggplot visuals and knitr::kable() tables, not raw R output.
- +1Write and submit: follow the Research Project Outline structure, knit to PDF with the R code hidden (appendix if shared), and remember the 20% plan and 50% report share a 40%-combined hurdle — so the report is where most of the mark is decided. Confirm the due date on Canvas.
Key terms
- Research Project (capstone)
- The individual R Markdown report (50%) answering the chosen research-plan question, submitted via Turnitin in the formal examination period. With the 20% plan it forms the unit's 70% capstone and a 40%-combined hurdle to pass.
- Hurdle assessment
- A requirement that the Research Plan (20%) and Research Project (50%) together reach at least 40% combined for you to pass the unit — a threshold on the pair, not on each separately.
- Theory-driven variable selection
- Choosing predictors because a substantive mechanism justifies them, rather than maximising R² or relying solely on a selection algorithm. The expected standard for the capstone.
- Substantive predictor
- A variable that operationalises the theory and helps explain the mechanism, as distinct from a mere predictor included only for fit. The report should distinguish the two and consider necessary controls.
- Reproducible research
- Analysis captured in an R Markdown file that anyone can re-run, giving transparency, verifiability and shareability — fundamental to the scientific method and to how the project is delivered.
- knitr::kable()
- The R function for rendering clean, formatted tables in a knitted report, used instead of pasting raw R console output — part of presenting results to a high standard.
Conclusion and Delivering the Research Project FAQ
How is the final Research Project weighted, and what's the hurdle?
The individual Research Plan is worth 20% and the individual Research Project report 50%, together the 70% capstone. They form a hurdle: you must score at least 40% combined across the two to pass the unit. Because the report alone is half your grade, it is where the marks are decided — plan your time accordingly and confirm the exact due date on Canvas.
What separates a high-scoring project from an average one?
Theory-driven choices and honest interpretation. Strong projects justify their variables by a mechanism rather than maximising R² or trusting a selection algorithm, control for confounders, match and justify the model to the outcome, and present polished ggplot visuals and kable tables. Weaker ones over-claim causation, dump raw R output, or throw in every variable. The report body carries no R code — it lives in the RMD or an appendix.
There's no exam — what do I hand in and when?
The capstone is a written R Markdown Research Project report, knitted to PDF with code hidden and submitted via Turnitin at the start of the formal examination period; there is no sat exam. The Research Plan is due earlier, in Week 5. Both dates and the outline are on Canvas — confirm them there rather than relying on any fixed calendar date.
Can AI help me deliver my Research Project in DATA4207?
Yes, as a study aid — for understanding and feedback, not for producing the work. Sia can explain how to scope a question, select variables by theory, justify a model, and structure the report, and give feedback on your own draft's logic. It does not write your graded assessment, and this unit does not permit generative AI on assessed work unless the coordinator explicitly allows it — always confirm on Canvas.
Assessment move
Reverse-engineer the week from the weighting. The 50% report decides your grade and shares a 40%-combined hurdle with the 20% plan, so allocate time there. Choose your question early, source the additional dataset, and write a falsifiable hypothesis grounded in the literature. Select variables by theory and list the confounders you will control — resist maximising R² or leaning on a selection algorithm. Match and justify your model to the outcome type, and present results with polished ggplot charts and knitr::kable() tables rather than raw output, knitting to a code-hidden PDF that follows the outline. Use the earlier writing chapters as your structure guide, keep the analysis reproducible, and confirm the outline, word counts and due dates on Canvas. Generative AI is not permitted on assessed work here.
Working through Conclusion and Delivering the Research Project in DATA4207? Sia is AskSia’s AI Statistics tutor — ask any DATA4207 Conclusion and Delivering the Research Project question and get a clear, step-by-step explanation grounded in how DATA4207 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
