Data management Key Issues in Data Entry and

Data management Key Issues in Data Entry and Management Cleaning Data, Who Should do What, When? Juan Muñoz

Levels of quality control • Range checking • Simple consistency checks • Inter-record checks

Range checking • Age should be a number less than 100 • Gender should be coded either “ 1” for male or “ 2” for female • No numbers in the name field • etc

Simple consistency checks • Age and birth date should be consistent with the date of the interview • Head of the household should be 18 years or older • A doctor should have completed university studies • etc.

Inter-record checks • Sub-totals • Food consumption • Checks with reference tables (anthropometrics) • Cash balance • Item-specific unit prices • etc.

Levels of quality control • Ranges Easier to conceive and program • Simple consistency • Inter-record Harder to conceive and program

Levels of quality control • Ranges Error likely to be due to miscoding or data entry • Simple consistency • Inter-record Errors likely to be due to interviewing

Why concurrent data entry? • Quality control • Turnaround time • General improvement of field procedures • Eliminates encoding as a separate task

Data entry is an integral part of the LSMS, not an afterthought

What is needed to develop an LSMS data entry program? • Integrated skills / tasks • Development time • Data entry software

Integrated skills • Questionnaire and data entry program evolve synergistically (example: subtotals) • Data manager needs to be a part of the core team from the beginning • Integration will become even more important in future surveys

Tasks • Data entry program development - screen design - range checks - consistency checks - reference data • • Data entry manual Training Supervision during field work Compilation and documentation

Development time • First version to be completed before the field test • Working version needs to be ready for training • Debugging needed during first weeks of data collection

Data entry software • In-house package • Six packages evaluated in 1994, IMPS and ISSA found adequate

Data entry software • In-house package • Has evolved synergistically with acquired experience and available technology • Six packages evaluated • IMPS and ISSA now in 1994, superseded by CSPro IMPS and ISSA found adequate