Chapter 37 — Dissertation I: proposal

Lesson map

Lesson 37 published. Map, example, EX109.

flowchart TD
  N0["Learning outcomes"]
  N1["37.1 The proposal"]
  N2["37.2 Literature"]
  N3["37.3 Analysis plan"]
  N4["37.4 Timeline"]
  N0 --> N1 --> N2 --> N3 --> N4
  N4 --> EX["Worked example"]
  EX --> Q["Exercise EX109"]

Logical drawing

flowchart LR
  N0["Learning outcomes"]
  N1["37.1 The proposal"]
  N2["37.2 Literature"]
  N3["37.3 Analysis plan"]
  N4["37.4 Timeline"]
  N0 --> N1 --> N2 --> N3 --> N4 --> Q["EX109"]

[Learning outcomes]
    |
    v
[37.1 The proposal]
    |
    v
[37.2 Literature]
    |
    v
[37.3 Analysis plan]
    |
    v
[37.4 Timeline]
    |
    v
[Worked example] --> [Exercise EX109]

Learning outcomes

You will (1) write a dissertation proposal; (2) review literature; (3) plan analysis and timeline.

37.1 The proposal

Amara writes a proposal with background, research question, literature, method, ethics, timeline and expected contribution. The background explains how invalid entries harm Harbourline's stock and reports. The question asks whether data champions reduce invalid entry rates. The contribution is evidence on a low-cost governance practice that small organisations could adopt. She keeps the proposal to five pages, following the course guidance, and submits it to Dr. Lindqvist for comments before the deadline. Her tutor returns comments within a week, and she revises the draft.

37.2 Literature

Amara reviews literature on data quality, governance roles and organisational change. She draws on Wang and Strong for quality dimensions, Khatri and Brown for decision domains and DAMA-DMBOK2 for stewardship. She finds little research on champion networks specifically for data, though some on similar roles in other fields. She notes that this gap supports her contribution. She checks each source she cites by reading it, rather than relying on summaries written by others. She records each source in a reference manager with notes on its relevance.

37.3 Analysis plan

Amara plans her analysis: generate synthetic data for twenty stores over twelve months, with champions introduced in ten stores at month six; compute invalid entry rates; estimate the difference-in-differences; and compute a bootstrap interval using Python's random module. She will also check pre-introduction trends. She writes the plan before generating data, so she cannot adjust the analysis to produce a preferred result, a practice known as pre-registration. She sends the plan to her tutor, who keeps a dated copy, so that any later change to the analysis must be explained in the dissertation.

37.4 Timeline

She sets a timeline: literature complete by month two, data and analysis by month four, drafting by month six and revision by month seven. Each milestone has a check-in with her tutor. Dr. Lindqvist approves the proposal, asking her to explain how synthetic data patterns were chosen, since readers will want to know how realistic they are. She adds a section describing the placement observations that informed the patterns. She also builds two weeks of slack into the timeline, which proves useful when an illness delays her drafting.

Worked example

Amara writes a five-page proposal, reviews literature and finds a gap on data champions, pre-registers a difference-in-differences analysis with bootstrap intervals and sets a seven-month timeline.

Synthetic case. The proposal is synthetic; cited works are real.

Errors and limits

Planning analysis after seeing data invites biased choices.

Exercises

• EX109 — Literature gap. What gap did Amara find?
  Show marking criteria

  Rubric: 3 = little research on data champion networks;; 2 = partly correct;; 1 = vague;; 0 = none.
• EX110 — Store design. How many stores receive champions?
  Show marking criteria

  Rubric: 3 = ten of twenty;; 2 = partly correct;; 1 = wrong;; 0 = none.
• EX111 — Pre-registration. Why write the plan before generating data?
  Show marking criteria

  Rubric: 3 = to avoid adjusting analysis toward a preferred result;; 2 = partly correct;; 1 = vague;; 0 = none.