+44(0) 1494 435112 info@surveybe.com

Quality Data

The ability to consistently collect and deliver high quality data, is central to surveybe data collection (CAPI) software.  Ensuring high levels of data quality starts with being able to minimise data error and inconsistencies.  Surveybe facilitates this with easy to use functionality.  In essence:

 

  • surveybe removes mistakes that arise when re-keying data from paper into a computer, as data is entered directly by the interviewer
  • surveybe automates questionnaire skipping patterns that ensures only the right questions get asked
  • surveybe performs hundreds of consistency checks on the responses given during the interview and ensures that only clean responses get recorded
  • surveybe can automatically capture elaborate GIS data, which allows for more detailed, and ensures more accurate, GIS data
  • surveybe conducts computed answer types to provide instant feedback to enumerators, such as total calorific values or area and distance measurements.

validation Validating Responses

To be extra confident that the responses are valid, they can be checked in a number of ways. At the moment of entering a response, surveybe can either disallow an entry and show an error message, or simply show an error or a warning if the response does not meet the requirements set by the survey designer.

Read More

For example:

  • entering text in a field that requires a number would not be allowed and would trigger an error message to the interviewer
  • entering an age of 125 could trigger a warning message to the interviewer that this seems unlikely and they should confirm with a comment whether or not this is correct.
  • missing answering a question would trigger an error message that the question has not been answered
  • if the calculations of data entered during interview do not fit within pre-set parameters or e.g. do not total a set amount (e.g. 100%), then a warning will be given

android error message

In all cases Surveybe has a single-click navigation to the location of each error or warning so the interviewer can immediately see and deal with the issues.

At the end of the interview there is a validate-all option that will re-check the whole interview and identify any remaining errors or warnings. This function can be used by both the interviewer and anyone performing quality control functions to easily identify areas to address.

error Warning Messages

Users can ensure that they include customised and contextual error messages to enumerators.  These can be set in different languages and will be displayed during the live interview – even when working offline.  The ability to also distinguish between ‘hard’ errors and other warnings can help enumerators understand the nature of the issue.

Read More

Shown below

android warning message

In all cases Surveybe has a single-click navigation to the location of each error or warning so the interviewer can immediately see and deal with the issues.

Data Quality Monitoring

No matter how exacting the questionnaire design and enumerator training process may be, further quality controls can be applied during fieldwork to provide valuable insight into what is being recorded, by whom, when and how.

Surveybe enables researchers to understand if there are any systematic aspects of their data collection that may be impinging data quality.  Collecting paradata such as time stamps can allow assessment of the time taken to progress through interviews and comparisons between enumerators to check for outliers which can be dealt with to improve any under-performance. GPS readings allow you to be certain that your enumerators are where they are supposed to be.

Audio Audit recordings are particularly useful for ‘listening in’ on parts of the interview alongside the data set as received from the field.  The recordings can be defined to the frequency and duration that you desire for your project.

At the enumerator and supervisor levels we incorporate use of metadata headers to clearly display the file status of the interview file.  This helps ensure full validation and error detection has been conducted before leaving the enumeration area.

Our full suite of design features are tailored to facilitate the very highest data quality.  Our research team provide insight into using surveybe enhanced features via their various blog posts and conference papers.  See here for a recent paper entitled: USING CAPI TO IMPROVE AND EVALUATE THE QUALITY OF SOCIOECONOMIC SURVEYS