How to Enter Paper Questionnaire Data Accurately: Double Data Entry Explained
Collected your data on paper? Here is how to enter it accurately, from building an entry template to double data entry, so typing errors do not end up in your results.
Paper questionnaires are still everywhere: in fieldwork with no signal, with respondents who prefer paper, and in projects that started before anyone thought about tablets. But the data only becomes useful once it's typed into a computer, and that's where errors creep in. A misplaced decimal, a skipped row or a 7 typed as a 1 can quietly distort your results.
Here's how researchers and data teams enter paper data accurately.
1. Build the entry template before anyone types
Never just open a blank spreadsheet and start typing. First create a template that mirrors the questionnaire:
- One row per questionnaire, one column per question (or per option, for "tick all that apply" questions).
- A unique ID for every form, written on the paper copy and entered in the first column. It's how you trace any value back to its source.
- Short variable names such as
q1_ageandq7_income, in the same order as the paper form so typists don't jump around. - Numeric codes for answer options (1 = Yes, 2 = No) with consistent codes for missing values, for example 98 = Don't know, 99 = Refused, and blank only for questions that were legitimately skipped.
Write these decisions down in a codebook. It will also save you hours at the analysis stage.
2. Add validation so errors are caught as they're typed
Entry tools can reject impossible values on the spot:
- Range checks: age between 15 and 100, household size between 1 and 30.
- Allowed codes: a yes/no question accepts only 1, 2, 98 or 99.
- Consistency checks: number of children can't exceed household size.
- Skip patterns: questions that should be blank, given an earlier answer, stay blank.
In Excel, use Data → Data Validation with drop-down lists and number limits, and lock the header rows. For larger projects, a data entry screen in KoboToolbox, ODK, CSPro or EpiData enforces the rules more strictly, and the form can mirror the paper layout.
3. Decide between single and double data entry
Single entry means one person types each form once, perhaps with spot checks on a sample. It's faster and cheaper, and it can be acceptable for short forms or lower-stakes data.
Double data entry means two people independently type every form into identical templates. The two datasets are compared value by value, and every mismatch is checked against the original paper and corrected. Two people rarely make the same typing error on the same value, so most keying errors surface. That's why double entry is a standard expectation in clinical trials and many academic studies.
| Single entry | Double entry | |
|---|---|---|
| Effort | 1× | About 2× plus reconciliation |
| Catches keying errors | Only through spot checks | Most of them, systematically |
| Best for | Short forms, internal records | Research data, theses, M&E, clinical data |
4. Compare and reconcile
To compare the two entries, sort both datasets by form ID and compare cell by cell. In Excel, a formula like =IF(Sheet1!B2=Sheet2!B2,"","CHECK") across the grid flags differences. In Stata, cf _all using entry2.dta, verbose lists every mismatch. For each flag, pull the paper form, read the correct value, and fix the master dataset. Keep a log of what was changed.
5. Flag, don't guess
Some handwriting is genuinely unreadable, and some answers are missing or contradictory. Typists should never guess. Record these in a query log (form ID, question, issue) so the researcher can decide, ideally by checking with the field team. Guessed values are invisible errors.
6. Final checks before analysis
- Count records against the number of paper forms. Are any missing or duplicated?
- Run frequencies on every variable to spot stray codes.
- Check minimum and maximum values for numeric variables.
- Cross-tabulate related questions to catch inconsistencies.
- Store the paper forms safely until the study is complete, since you may need them again.
Hundreds of questionnaires waiting to be typed? We enter paper questionnaires into Excel, SPSS, Stata or CSV with a custom template, validation checks and optional double data entry, and deliver a codebook and query log with your dataset.
Got a pile of paper to digitise?
Accurate, double-entry verified data entry of questionnaires, registers and PDFs into Excel, SPSS, Stata or CSV.
Get a free quote →