Rosella       Machine Intelligence & Data Mining

Interviewer Falsification and Detection

Survey is very important part of scientific and marketing research. When survey responses are collected with falsified information, the consequence is very damaging. Survey falsification can occur if interviewers do not collect information as described by survey designers. Types of falsification include;

  • When survey is paid, respondents may create a number of bogus accounts. Responses are likely to be inconsistent for the same questions asked for different surveys.
  • Survey interviewers may have incentives to improve productivity or increate the number of interviews artificially.

Ways to detect falsified interviews include;

  • Categorical questions: Deviation of distribution of categories from the overall distribution.
  • Yes/No questions: Deviation of average responses over the overall averages.
  • Numerical questions: Deviation of averages over the overall averages.
  • Logical tests: The same questions asked differently and placed at different places. Different survey responses may be compared for consistency.

Falsification Detection Expert System

Detecting falsified interviews can be difficult without the help of sophisticated s/w systems. The following shows an example report generated with expert system rule engine and powered by Rosella DBMS.

interviewer falsification detection analysis.


If you are interested in s/w solutions, please contact us.