THE STYLO INDICATOR: A REVIEW OF PSYCHOMETRIC QUALITIES

S.M.de Ciantis Ph.D. Consulting Psychologist.

Purpose of the Indicator

The STYLO Indicator has been designed to provide a superior tool for stimulating a person's better understanding of his or her preferences for process and outcomes in tasks and relationships and of how others see him or her, and why they do so. The improved understanding will encourage wider toleration and utilization of differences between people, leading to better collaboration and management of resources.

The Indicator is to enable people, at all levels in the organization, to manage their behavior through learning from the interpretation of the data. Its purpose is therefore primarily developmental, although it could be useful in the toolkit of the assessor and selector as a means of evaluating compatibility with co-workers and the organizational culture, as well as 'suitability', rather than 'capability", for the function or role.

The Model Behind the Indicator

The behavioral model underpinning the STYLO instrument is well-grounded in classic behavioral research spanning more than 50 years, representing motivational drives that manifest themselves in an individual's overt behavior. As such, the theoretical structure is well established in terms of the defined domains and their linkages with the behaviors, the order and intensity of the preferences for which the instrument aims to measure.

The Integrity of the Indicator

The intention is to do the measurement in a psychometrically sound and reliable way so that it should be possible to have confidence that the scores, or profiles, derived from the instrument are valid indicators of the four behavioral Approaches in the basic model and allow valid inferences and assumptions to be made about an individual's preferences or lack of preferences for each of the Approaches. There are many behavioral measuring tools available, some well established, which rely totally on "Face Validity", their reliability and other psychometric data either unproved or presented in ambiguous ways. It is all too easy to be misled by apparent "Face Validity", where all the key words and phrases seem to ring true. Without the substance of reliable measurement, practical use is simply at best anecdotal and at worst specious. It is certainly unwise and possibly unethical to encourage individuals to act upon insights from such unscientific methods.

The significance of STYLO from a consulting psychologist's point of view is that it has been developed to maintain the highest standards of psychometric excellence. STYLO has been developed through a series of stages, whereby item analysis has raised all its scales to exceed the standard recommended levels. Consequently practitioners will have no anxieties about using data generated by the instrument.

Research on the Indicator

Work to provide direction for the development of the instrument has concentrated first of all on the internal reliability of the eight individual scales (the four different Approaches in both the Encouraging and Threatening Environments). In other words, do the statement endings accurately correlate with the constructs of the model which the STYLO instrument and its application seek to explore? This priority reflects one of the fundamental concerns of the psychometrician - that while the reliability of the instrument as a means of measuring the dimensions of the model does not guarantee validity (the model itself may be "wrong", based on false premises), validity is not possible without an inherently reliable instrument (the model itself may be "right", but the instrument is giving a false, or hit-and-miss, picture of people's relationship to and with it).
Each stage of the creation of the STYLO instrument has been researched to establish:

Item Distribution
The respondents agree or disagree with an item in a relatively balanced way. The problem here is often the well-known one of 'socially desirable behavior" to which almost everyone would gravitate, and the opposite, to which almost no one wants to admit.

Item-total Correlation
How closely the items within a scale relate to the scale as a whole. Too high a correlation and it appears that literally the same item is being offered over and again; too low or negative, and it appears that respondents do not see this item as associated with the construct, or that they see it associated with another one.

Scale Internal Reliability
The overall extent to which a group of items "hangs together' as a homogeneous set. A low level suggests that the items have little or no internal relationship to one another so there is little justification for aggregating them into a score indicative of a construct.

Item Distribution

In creating a sound version of the STYLO instrument, we found initially that there were problems with item distribution until we eliminated items which suggested negative or undesirable behaviors. The technique of 'paired comparisons' is also used to force the respondent into considering every item three times. The target is to ensure that not more than 80% or fewer than 20% of responses fall at the extreme ends of the possible range of scores (Maxwell, 1971). Research will continue to monitor the ongoing achievement of the target.

Item-total Correlation

With 180 items, we have tried to ensure that each item correlates to the total scale of which it is meant to be a representative, within the .2 to .6 range. (Cronbach, 1970) Our initial research showed where items needed to be worded more carefully and freed from associations which contaminated the response. Items in the published version of the Indicator have correlations which fall within these criteria.

Scale Internal Reliability

The target here is an alpha coefficient of, or higher than, 0.7 for each of the 8 scales (Nunally, 1978). Latest analysis indicates (N=97):

Important Update from the Developers of STYLO on the Psychometric Properties of the STYLO Instrument

April 1997

The original development of the STYLO Indicator in 1993 - 1994 was accompanied by rigorous reliability testing leading to a report on the psychometric properties of the Instrument, by a consultant psychologist, in October 1994, concurrently with publishing the Indicator for general use. Since then the authors have continued to work diligently to ensure that the Indicator achieves a high level of quality and sophistication and produces accurate, meaningful data about individuals' preferences for process at work.

A significant contribution to this work has been to use the data generated in the original research to improve the homogeneity and reliability of the elements of each scale. Using the model of those elements which performed at, or above, the desired level on item distribution, item total correlation and scale internal reliability, the authors have modified statements and endings so they contribute to a better overall result. These modifications appear to maintain the accuracy of the original version while yielding a more finely-tuned picture of behavioral preferences. The reactions of colleagues who have been part of this ongoing development effort have been uniformly positive, and the authors have incorporated the improvements into the STYLO Indicator from the start of 1997.

At the same time the authors have used valuable feedback from users of the Indicator in Europe and North America to simplify some of the original language and to incorporate changes which reflect specific differences between British and American usage and spelling.

We continue to look for, and value, the feedback from the growing number of users of the STYLO Indicator. Further research into the psychometric quality of the Indicator is to be undertaken. The results of this research will, of course, be circulated amongst all our users.

Alan Hamilton Msc., MA.

Ron Jackson BA., MA., MEd.

Format Systems Ltd.

The Vineries,

Northleach, Cheltenham, Glos. GL54 3EU

UK

Phone/Fax 01451 861211