SCIENCE POINT OF VIEW

Changes to M Score in PI2

Overview

M score refers to the number of adjectives selected on the Behavioral Assessment. M is measured because it serves an important purpose in scoring the behavioral assessment. Specifically, respondents' behavioral responses are scored based on norm tables where the expression or score of each factor is relative to the total number of adjectives selected. For example, someone who selects 10 words measuring Factor A (Dominance) and 30 words total will have a greater expression of Dominance than someone who selects 10 words measuring A, but 60 words total.

For PI2, we revisited the way that our software represents and communicates M. M is still and will continue to be used as a scoring mechanism. However, while M will still be used in scoring on the backend, it will no longer be reported in PI software or reports. This decision is rooted in the history of M and internal research findings – specifically, this decision was made because there is currently no evidence to support use of M in workforce decision-making. Internal research has not provided a strong indication that M provides any meaning or value beyond its scoring mechanism.

History of M

The PI Behavioral Assessment was created in 1955. There is no mention of M in the original BA technical documentation. Instead, the first mentions of M are found in Arnold Daniels’ notebooks from an unknown date and from another PI researcher, Mattern in 1966. In Daniels’ notebooks, several different claims are made regarding M, including that M scores can be used to identify if a person’s behavior is dominated by their self or self-concept (based on which M is higher); Daniels also refers to M as energy or aliveness.

Less than a decade after these initial definitions, the interpretation of M had been expanded despite lack of scientific research. In workshop materials from 1974, M is referred to as norm level, energy level, and the resultant factor. Daniels defines it as a level of responsiveness to environmental stimulus, and goes beyond that to claim people with higher Factor M scores have more intelligence, awareness, capacity for activity, and potential for achievement. These materials may have been supported by theory or anecdotal evidence, but did not contain empirical evidence substantiating these claims.

The lack of scientific research on the interpretation and validity of M precluded its use in the candidate selection process after 1978 when the Uniform Guidelines on Employee Selection Procedures were released by the US Departments of Justice and Labor and The Equal Employment Opportunity Commission. These guidelines were designed to help employers comply with federal laws that prohibit employment discrimination. The guidelines serve as legal precedent and require that employers utilize selection procedures that are consistent with legal standards and validation standards accepted by psychologists.

The release of the uniform guidelines necessitate construct and criterion validity evidence on any practice utilized for employee selection. While it may have been permissible to support usage of M with theory or anecdotal evidence as it was in the 1974 workshop materials, the uniform guidelines now require robust empirical evidence behind selection techniques.

Validation Research on M

Without available reference or robust empirical research, interpretations of M shifted over time and unsubstantiated usage of M proliferated. Although The Predictive Index never sanctioned use of M for hiring in the modern workplace, the PI Science team attempted to validate usage outside of M’s scoring mechanism in 2013. This research was conducted with a group of partners aimed at identifying whether a standard definition of M existed and whether it was utilized consistently across various clients. This research discovered that M was most commonly labeled as “Response Level” but was also defined as energy, activity level, stamina, capacity to sustain activity, capacity to tolerate stress, and ability to adapt or modify behavior to changing responsibilities. The conclusion of this research was that there was no consistent, singular definition of M.

In 2016, the PI Science Team conducted a large-scale study to clarify M’s definition and any practical purposes it served. The first step of this project was to collect the whole domain of how M was referred to. The team identified approximately 550 items to measure various interpretations of M and built a survey containing 140 questions to test those interpretations and to determine how many constructs were being measured with that collection of items. They gathered 395 respondents and determined that the collection of interpretations could be categorized into four constructs: engagement, connectivity, emotional stability, and adaptability.

The results of the research found that Self M appeared to loosely relate to two of these four interpretations: connectivity and adaptability. Connectivity refers to one’s comfort in handling a variety of situations or expectations, specifically in a work environment. Adaptability refers to one’s ability to multitask. The relationships between M and connectivity and adaptability were small and loadings for items on both constructs were low, meaning that M is only a coarse measurement of these constructs. The Science team concluded that there was no definition for M that was clear or valuable enough to warrant providing recommendations around using M as anything other than a scoring parameter.

Removing M in PI2

To summarize, the decision was made to remove M from the behavioral assessment reports in PI2 because there is currently no evidence to support use of M in workforce decision-making. Further, internal research has not provided a strong indication that M provides any meaning or value beyond its scoring mechanism.