Survey Data Analysis

Clean, documented survey data, weights you can defend, honest margins of error, coded open text and the routines that turn each wave into findings people act on.

Free · Mac (Apple silicon) and Windows · v1.0.5 · Other platforms

Then open the Marketplace in the app, search for Survey Data Analysis, and click Install. It goes into your own files, and works offline from then on.

What it installs.

1

Pillar

1

Area

50

Projects

206

Starter tasks

📊 Data & Analytics

Survey Data Analysis

Cleaning, weighting and analysing survey responses, including open text, to produce trustworthy findings and avoid common biases.

  1. Written analysis plan before fieldwork closes
  2. Survey codebook and data dictionary
  3. Read-only raw export and tidy analysis file
  4. Response rate and fieldwork disposition record
  5. Response quality screening rules
  6. Skip logic and missing data audit
  7. Derived variables and recode log
  8. Respondent confidentiality and small-cell rules
  9. Choosing a tool for weighted survey analysis
  10. Live fieldwork monitoring against quotas
  11. Weighting targets and benchmark upkeep
  12. Tracker survey wave routine
  13. Survey data archive and version register
  14. Second-analyst check before results go out
  15. Survey fatigue and contact frequency rules
  16. Tested question bank for future surveys
  17. Findings to actions follow-up log
  18. Annual survey programme review
  19. Margins of error, design effects and effective sample size
  20. Raking weights to population margins
  21. Crosstab significance testing without fishing
  22. Analysing Likert and rating scale questions
  23. Scale reliability and factor structure checks
  24. Coding open-ended survey responses
  25. Key driver analysis for satisfaction surveys
  26. Spotting questionnaire bias before analysing
  27. Nonresponse bias assessment
  28. Deciding how to report Net Promoter and single scores
  29. Survey platform comparison for clean exports
  30. Scripted cleaning and tabulation pipeline
  31. Crosstab banner and table specification
  32. Shortening the questionnaire with item analysis
  33. Briefing a research agency on data delivery
  34. Annual staff survey results readout
  35. Topline results within two weeks of close
  36. Board or trustee presentation of survey findings
  37. Public release with a methods statement
  38. Soft launch data check before full fieldwork
  39. Beneficiary feedback survey on a small budget
  40. Before and after programme survey for a funder
  41. Dissertation or thesis survey analysis chapter
  42. Comparing results across translated questionnaires
  43. Taking over an inherited customer survey
  44. Community needs assessment survey
  45. Item nonresponse and multiple imputation
  46. Assessing a change of survey mode
  47. MaxDiff and conjoint preference studies
  48. Variance estimation for stratified and clustered samples
  49. Panel attrition and longitudinal weights
  50. Multilevel regression and poststratification

Where it starts.

Each project arrives with its purpose, milestones and first tasks written out. Here is how the first three read.

Written analysis plan before fieldwork closes

Analysis decided after the data arrives tends to chase whatever looks interesting, and a survey with no stated questions produces fifty tables and no answer. Writing down the three to five decisions the survey must inform, the key measures, the subgroups to compare and the tests to run, before the survey closes, keeps the findings honest and the report short.

First tasks

  • Write the three to five decisions the survey results must inform
  • Map each research question to the question numbers that answer it
  • List the subgroups to compare and estimate the base size of each
  • Mark the primary measures and send the plan to the survey sponsor

Survey codebook and data dictionary

Exports from survey platforms arrive with cryptic column names like Q7_3, numeric codes with no labels and multi-select questions split across a dozen columns. A codebook giving every variable its question wording, value labels, routing condition and missing codes means anyone, including you in six months, can read the file without the questionnaire open beside it.

First tasks

  • Export the questionnaire and the data map from the survey platform
  • Rename cryptic columns to short readable names and record the mapping
  • Write value labels for every coded answer, including don't know
  • Note the routing condition beside each filtered question
  • Save the codebook beside the data with today's version date

Read-only raw export and tidy analysis file

Editing the raw download by hand is how responses quietly vanish and nobody can say what changed. Keeping the original export untouched in its own folder, and building a separate analysis file with one row per respondent and every change recorded, means any figure in the report can be traced back to what respondents actually submitted.

First tasks

  • Download the full raw export and note the date and number of records
  • Set the raw file to read-only in its own folder
  • Create the analysis file with one row per respondent and a unique ID
  • Start a change log with columns for date, change, records affected and reason
  • Reconcile the record count against the platform's own completes figure

Make it yours.

Install Survey Data Analysis, then tell Aurum the date you're aiming for. It checks whether the plan fits the weeks you actually have.

Free · Mac (Apple silicon) and Windows · v1.0.5 · Other platforms

Blueprints are general information to help you organise your own work. They are not medical, financial, legal or other professional advice. See the Terms.