Chapter 1 — Research question and literature method

Learning outcomes

You will (1) frame a master's research question on data stewardship; (2) run a structured literature search; (3) record search decisions transparently.

1.1 The student and the setting

Tomás Ferreira starts a two-year master's programme at the fictional Kestrel Institute of Technology, supervised by Dr. Ingrid Solberg. His partner organisation is the fictional Arden County Council, where five departments keep their own copies of residents' addresses for council tax, waste collection, libraries, parking permits and social care. Addresses are keyed differently in each system, so the same household appears several times, letters go to old addresses and errors take weeks to fix. The council plans to appoint data stewards in two departments as a pilot.

1.2 Framing the question

Tomás drafts a question: does appointing data stewards reduce the time taken to correct address errors? Dr. Solberg asks him to make it answerable. He specifies the outcome as days from an error report to its correction in the system, the intervention as named stewards with defined responsibilities, the comparison as departments without stewards, and the period as six months before and after the pilot. DAMA-DMBOK2 describes stewardship as accountability for data on behalf of others, which gives him a definition to test rather than a slogan.

1.3 Searching the literature

Tomás searches two academic databases and the publications of DAMA and ISO using fixed terms: data stewardship, data governance, master data and data quality improvement. He records the date, database, terms and number of results for each search, and screens titles and abstracts against written criteria. Studies must report an outcome measured before and after a governance change. He finds many frameworks and case descriptions but few studies that measure outcomes, which matches the gap other reviewers report. Grey literature is logged separately.

1.4 Recording decisions

Every inclusion and exclusion is logged with a reason, such as no outcome measure or not about governance. Dr. Solberg checks a random tenth of his decisions and agrees with all but one, which they discuss and resolve. Tomás learns that a transparent search allows readers to judge whether evidence was missed and lets him update the review later without starting again. He keeps the log with his dissertation materials, so examiners can see how the evidence base was assembled rather than having to trust it.

Worked example

Tomás frames a stewardship question with outcome, intervention, comparison and period, searches with fixed terms, screens against written criteria and logs every decision with a supervisor check.

Synthetic case. Kestrel, Arden and all people are synthetic; DAMA-DMBOK2 is real.

Errors and limits

Search terms shape what is found; test alternative terms before finalising.

Exercises

• EX001 — Partner problem. What problem does Arden County Council face?
  Show marking criteria

  Rubric: 3 = duplicated addresses across five departments causing errors;; 2 = mentions duplicates;; 1 = vague;; 0 = none.
• EX002 — Outcome measure. What is Tomás's outcome measure?
  Show marking criteria

  Rubric: 3 = days from error report to correction;; 2 = partly correct;; 1 = vague;; 0 = none.
• EX003 — Comparison group. What is the comparison in his question?
  Show marking criteria

  Rubric: 3 = departments without stewards;; 2 = partly correct;; 1 = vague;; 0 = none.
• EX004 — Stewardship definition. How does DAMA-DMBOK2 describe stewardship?
  Show marking criteria

  Rubric: 3 = accountability for data on behalf of others;; 2 = mentions accountability;; 1 = vague;; 0 = none.
• EX005 — Inclusion criterion. What must included studies report?
  Show marking criteria

  Rubric: 3 = an outcome measured before and after a governance change;; 2 = partly correct;; 1 = vague;; 0 = none.
• EX006 — Supervisor check. What share of decisions did Dr. Solberg check?
  Show marking criteria

  Rubric: 3 = a random tenth;; 2 = partly correct;; 1 = vague;; 0 = none.
• EX007 — Literature gap. What did the search find lacking?
  Show marking criteria

  Rubric: 3 = studies that measure outcomes;; 2 = partly correct;; 1 = vague;; 0 = none.