The Data Cleaning & Codebook Kit

What it's for

Cleaning decides your results. This makes every decision deliberate, recorded and reproducible — so the dataset can be defended and re-created from raw.

What's inside

1. Set up for clean work — Preserve the raw file, decide scripted or manual, and name things so software can read them.
2. Audit before you change anything — The structural pass: dimensions, identifiers, duplicates and a missingness overview.
3. Types, dates and levels — Type conversions with a count of what silently became missing, plus dates and categorical levels.
4. Text and category harmonisation — The invisible problems, a mapping from raw values to standard ones, and field splits.
5. Recoding and derived variables — Every new variable with its rule, its reason and a verification against source cases.
6. Reverse coding and scale scores — The formula, the verification that catches a double reversal, and a rule for partial responses.
7. Missing data: diagnose — Codes that are not missing, the amount, the pattern and the evidence about the mechanism.
8. Missing data: handle and report — Strategy by situation, what each costs, and the five things every paper must report.
9. Outliers: triage, don't delete — Error or genuine, the treatment options, and a decision log with reasons.
10. Range, logic and consistency rules — A re-runnable rule set that finds impossible values automatically.
11. Build the codebook — Nine fields per variable, complete enough for someone who has never seen your data.
12. Freeze, reproduce, sign off — The ten-item gate before any test is run, plus the methods paragraph.

Swipe cards — the rules, formulas and phrasings worth keeping beside you, on one page to cut out.

The Log & Codebook (Excel / Google Sheets)

• Variable Inventory: types, permitted ranges, missing codes and a missingness flag
• Cleaning Log: one row per change, with cases affected and whether it was verified
• Missing Data: amount, mechanism, handling method and sensitivity, per variable
• Validation Rules: the re-runnable rule set, with violations found and resolved
• Codebook: one row per variable, with a completeness check
• Sign-Off: the readiness gate and the frozen file's details

Look inside

Low-resolution, watermarked previews of four of the 16 pages.

Who it's for

• Anyone about to analyse a dataset they collected, inherited or downloaded.
• Students whose supervisor has asked how a variable was derived — and who want the answer on file.

It is not software and it does not touch your data: it is the decision, documentation and sign-off layer around whatever you clean in.

From the author

Companion to The Data Cleaning & Dataset Preparation Workbook (Mastering Research Series), by Dr. Rafiq Muhammad, MD, PhD.

The kit is the worksheet layer; the workbook adds the method, the worked examples and the field notes behind each step.

Licence and refunds

One purchaser, personal academic use: print and annotate freely; please don't share, resell or upload the files to shared drives or course sites. Refund on request within 14 days if the files cannot be downloaded or opened. Full terms.

Lost your link?

Enter the email address you used at checkout and we'll email a fresh download link.

Other kits

• Sample-Size Justification — a computed sample size, a sourced effect size and the justification paragraph.
• Scoping Review Protocol — a search log, screening decisions, a charting form and the flow numbers.
• Abstract, Title & Keyword — a word budget, title scoring and a keyword map for the 250 words that matter.

See all kits