COMPARATIVE METHODOLOGICAL CRITIQUE

Comparative Methodological Critique

Comparative Methodological Critique

Introduction

In a companion paper (Ekkekakis & Petruzzello, 2000), we examined a number of important issues surrounding the measurement of affect in the context of exercise and identified several reasons for concern. Because measurement is the foundation of any research endeavor, the close inspection and resolution of these problems is necessary before substantial progress can be made in the study of the exercise-affect relationship.

The present paper extends our analysis by focusing on a specific measure, namely the Exercise-induced Feeling Inventory (EFI; Gauvin & Rejeski, 1993). The EFI was the product of escalating dissatisfaction with traditionally used measures of affective variables in the context of exercise, such as the Profile of Mood States (POMS; McNair, Lorr & Droppleman, 1971) and the State-Trait Anxiety Inventory (STAI; Spielberger, Gorsuch & Lushene, 1970). The EFI is a 12-item measure of “exercise-induced feeling.” It includes four subscales, namely Positive Engagement, Revitalization, Tranquility, and Physical Exhaustion. The popularity of the scale is growing as evidenced by its adoption by independent researchers (e.g., Annesi; Szabo; Szabo and Treasure). In the first review to focus on the measurement of affect in the context of exercise, Gauvin and Spence (1998) noted that, although a number of questions regarding the EFI remain to be investigated, the preliminary indications regarding the validity and reliability of the scale are promising. An independent examination of the factor structure of the scale also led to generally positive conclusions ( Vlachopoulos, Biddle & Fox, 1996).

The preliminary evaluations of the EFI, however, have focused exclusively on the computation of psychometric indices and have failed to provide a critical assessment of the conceptual underpinnings of the scale and the methods that were employed in its development. This is an important omission, especially given the fact that the EFI introduces a novel construct with unique structure. Evaluating whether the structure of a data set is consistent with a proposed model without considering whether the model was appropriately conceptualized in the first place seems problematic. Statistical fit has no bearing on whether the model itself is theoretically meaningful. The theoretical merit must be evaluated independently and prior to any psychometric assessments. It is also noteworthy that the researchers who have used the EFI in applied studies have not cited any conceptual reasons as a basis for selecting this particular instrument as opposed to some other relevant measure.

In summary, the EFI has yet to undergo a critical and thorough evaluation. Given its growing popularity, it is essential that such an evaluation be performed. To this effect, the present analysis concentrates on the conceptual foundations of the EFI and the methodological steps that were followed in its development. In both areas, the theoretical and methodological literatures on affective phenomena have served as guides. Thus, the discussion of the conceptual issues is embedded within the broader framework of affect theorizing and the methodological steps that were followed in the development of the EFI are contrasted with established guidelines.

1. Conceptual foundation

1.1. Delineation of content domain

As its name declares, the ...