Mental imagery, or the capacity to represent in the mind experiences of things that are not physically present (Matlin, 1989), is attracting increased attention from sports psychologists, coaches and athletes. For example, Murphy, Jowdy & Durtschi (1989) discovered that 90% of athletes, 94% of coaches and 100% of sports psychologists in the United States report using imagery techniques regularly in their training program, a finding later corroborated by Hall, Rodgers and Barr (1990). This interest in imagery is attributable mainly to the discovery that, under appropriate circumstances (see Murphy, 1990; Smith, 1987), the imaginary rehearsal by an athlete of a motor skill can lead to improvements in its subsequent performance (i.e. the "mental practice" effect).
The typical research paradigm used in the field of mental practice involves a "before-after" treatment comparison between people who have been exposed to imagery training for a given task/sport and those in various control conditions (e.g. physical practice only). Unfortunately, little attention has been devoted to a rather obvious flaw in this strategy: What happens if the treatment effects are confounded by individual differences in imagery ability? As Hall (1985) points out, "if the subjects in an experimental condition are asked to use an imagery strategy and these subjects are all low imagers, it is likely no effect or only a small effect for the condition will be shown". Clearly, therefore, the use of imagery tests can enhance the accuracy of mental practice research by ensuring that subjects are matched for visualization abilities before experimental treatments are administered. But which imagery tests are most suitable for this purpose?
As Table 1 shows, a variety of psychological tests have been developed to measure individual differences in imagery ability. But how valid and reliable are these tests? Unless they satisfy conventional psychometric criteria, they will have only limited value to coaches and researchers who wish to design or evaluate imagery training programs for athletes. Therefore, the main purpose of this paper is to evaluate the psychometric adequacy of the most popular tests available for the assessment of the mental imagery skills of athletes. A secondary objective will be to consider the principal conceptual and methodological issues encountered in imagery research.
The impetus for this paper comes from two sources. First, although there have been several reviews of imagery tests (see Anderson, 1981; Ernest, 1977; White, Sheehan & Ashton, 1977), the most recent one (by Sheehan, Ashton & White, 1983) was completed almost a decade ago. Since then, additional tests have been published which are potentially useful to athletes because of their focus on kinesthetic imagery. These measures include the Movement Imagery Questionnaire (MIQ) (Hall & Pongrac, 1983) and the Vividness of Movement Imagery Questionnaire (VMIQ) (Isaac, Marks & Russell, 1986). Therefore, an up-to-date evaluation of imagery measures is required. The second reason for exploring this field is that research on individual differences in imagery abilities is plagued by a variety of conceptual and methodological problems. For example, Hiscock (1978) concluded that "it is not clear what imagery questionnaires really measure or what criteria are appropriate for validating them". The implications of these issues for the assessment of the visualization skills of athletes must be examined.
This paper includes a brief outline of the nature and characteristics of mental imagery, examination of the psychometric properties of the most commonly used imagery tests and analysis of the main conceptual and methodological problems afflicting imagery research in sport psychology. Finally, recommendations will be provided to enable future researchers to address these difficulties.
Mental Imagery: Nature and Characteristics
According to Solso (1991), mental imagery refers to "a mental representation of a nonpresent object or event". …