Likert Versus Cronbach's Psychometric Thresholds: Reducing Error and Maximizing Agricultural Education's Scholarship Impacts
Institutional Repository Document
Overview
Research
Identity
View All
Overview
abstract
Instrumentation is a critical function in measuring social and behavioral science impacts on stakeholders, teachers, and change agents. Internal validity and reliability have long been considered social sciences quality gatekeepers. A systematic review uses a comprehensive search based on explicit protocols to review existing literature with a synthesis of data focusing on key questions. Systematic reviews are five steps; identify the critical question, formulate search parameters, systematically search databases, analyze data, and data summary interpretation (Lee et al., 2021). Using the five steps, authors systematically reviewed all articles from Advancements in Agricultural Development (AAD), Journal of Agricultural Education (JAE), Journal of Extension (JOE), and The Journal of Agricultural Education and Extension (TJAEE) from 2018 to 2022. The authors reviewed eight hundred ninety-six (N = 896) articles from the four publications. Fewer items produced lower construct reliability coefficients and thus, produced higher levels of error. Much of agricultural educations, broadly defined, published scholarship has not utilized instruments to collect data over the last five years; when they have, smaller numbers of items measured constructs. Likerts convention in his quintessential work on measuring social variables suggested that for measurements to be reliable an alpha of .9 should be achieved. Researchers should include a maximum number of statements and questions and eliminate those that do not contribute to reliability and add additional questions when acceptable levels of reliability are not achieved.
