Human Reliability as a Source of Error in Research

Abstract

Human error is embedded in every human endeavour. Given that research is conducted by humans, it is therefore prone to human error. This paper focuses on the errors that derive from transcribing data across formats. These 'inputs' errors arise largely from the monotony of the data entry process and may mean that an otherwise thoroughly designed research can potentially produce misleading conclusions. The paper reports the results of a quality checking process developed to monitor the transcription of data from paper-based questionnaires, collected as part of current PhD research, into the computer. Following the same entry method, the data from all questionnaires received were input twice by the PhD Candidate, then twice again by another participant. The 28,140 entries were matched and any differences analysed in order to quantify the occurrence of input errors were committed and identify the nature of these errors. The results suggest that where the input errors were committed had more impact on the findings revealed from each question than the total number of input errors committed.