Analysis of Reliability Coefficients In Electronic Survey Distribution Programs

Abstract

This study investigated the reliability coefficients of an electronic survey instrument, distributed via Qualtrics, designed to assess agriculturalists' perceptions of artificial intelligence in secondary agricultural education (SBAE). Building upon previous research that indicated negative correlations in a paper-based version of the instrument, this analysis aimed to determine if reliability would differ with electronic distribution and a new sample of SBAE teachers from Alabama and Georgia. Utilizing Rogers' diffusion of innovations theory as a framework, data were collected from six participants and analyzed using SPSS to assess the internal consistency (Cronbach's alpha) of eight constructs. The results revealed negative reliability coefficients for two constructs: perceived benefits and drawbacks, and support and guidance. Further examination identified specific items within these constructs, particularly those negatively phrased, as contributing to these negative values. The discussion highlights the importance of careful instrument design and the potential impact of negatively worded items on reliability analysis in electronic survey environments, emphasizing the need for reverse coding and thorough item analysis.