Identifying superfluous survey items

Surveys provide critical insights into consumer satisfaction and experience. Excessive survey length, however, can reduce data quality. We propose using constrained principle components analysis to shorten the survey length in a data-driven way by identifying optimal items with maximum information. The method allows assessing item elimination potential, and explicitly identifies which items provide maximum information for a specified number of items. We use artificial data to explain the method, provide two illustrations with empirical survey data, and make code freely available in an online tool