Abstract
Background: Originally developed for educational use, the Leadership Practice Inventory (LPI) is used to measure leadership practices in nursing research. There is limited reporting of LPI psychometric properties when used to measure leadership practices of nurses.
Objective: This study aimed to investigate psychometric properties of the LPI when used to measure the leadership practices of nurses.
Method: Data from 67 LPI-self and 347 LPI-observer respondents were used to establish LPI psychometric properties. Dimensionality of the LPI was investigated using exploratory principal components analysis, and LPI construct validity was established by exploring correlations with theoretically related concepts and a known-groups approach. The predictive validity of the LPI was investigated using regression analysis to determine whether observer-reported leadership practices of established and aspiring nurse leaders predict observer ratings of the effectiveness of the organization environment. Reliabilities of the new factor solution were explored.
Results: Factor analysis found that the identified three-factor solution has psychometric properties at least as strong as those found with the original five-factor LPI solution.
Discussion: The three-factor solution is advocated for use in nursing research because of the strong psychometric properties, lighter respondent burden, and decrease in research costs, as compared with the traditional five-factor solution. When used as an educational tool, the five-factor LPI may be preferred because it may be more useful for examining a greater number of leadership behaviors.