MO-index, a transformed score to present and compare results of Likert response format scales

Abstract

Evaluation of employees' opinions about work-related issues is commonly practised in organisations. Regardless of whether the surveys are conducted on a large-scale, e.g., organisation wide, or on a smallscale, e.g., work area, the mainstay is to develop initiatives to address challenges and build on achievements so as to benefit both the employees and the organisation. With resources invested in action plans to address problematic areas, organisations look for meaningful interpretation and maximum information from surveys, using follow-up surveys to facilitate better informed decision-making by management. The main focus of this paper was to address how best to accurately communicate results and assess change in scores. The Likert response format (e.g., 5-point ranging from strongly disagree to strongly agree) and summated scales are commonly used in surveys. The coding of Likert responses from the items with sequential integers and then summating them to produce a raw score percentage is widely practised as a means of presenting results and the comparison of scores from different survey cycles to employees and management. However, percentage scores are not interval measurement, which makes comparisons between scores (e.g., from two survey periods or two measures) inaccurate. It is proposed that there is a vast improvement in the interpretability of change scores when organisations adopt the MO-Index, a transformed score. It applies the log of odd ratios of the percentage of cumulative item endorsement to an inventory with a Likert response format to transform scores into an interval scale of logits. The interpretation of logits was made more meaningful and comprehensible to all stakeholders through the linear transformation of logits to a scale of Outcome Units that ranged from +100 to -100.