Academic journal article Research Quarterly for Exercise and Sport

Variability in Measurement of Swimming Forces: A Meta-Analysis of Passive and Active Drag

Academic journal article Research Quarterly for Exercise and Sport

Variability in Measurement of Swimming Forces: A Meta-Analysis of Passive and Active Drag

Article excerpt

An analysis was conducted to identify sources of true and error variance in measuring swimming drag force to draw valid conclusions about performance factor effects. Passive drag studies were grouped according to methodological differences: tow line in pool, tow line in flume, and carriage in tow tank. Active drag studies were grouped according to the theoretical basis: added and/or subtracted drag (AAS), added drag with equal power assumption (AAE), and no added drag (ANA). Data from 36 studies were examined using frequency distributions and meta-analytic procedures. It was concluded that two active methods (AAE and ANA) had sources of systematic error and that one active method (AAS) measured an effect that was different from that measured by passive methods. Consistency in drag coefficient ([C.sub.d]) values across all three passive methods made it possible to determine the effects of performance factors.

Key words: biomechanics, error, technique, variance


There is considerable variability in the force values reported for swimming positions and motions. These differences are found in studies on passive drag (the resistive force on a static body position) and active drag (resistance during actual swimming). For example, completely different results are reported for passive drag force on a body at the surface as opposed to below the surface. One research team (Clarys, Jiskoot, Rijken, & Brouwer, 1974; Jiskoot & Clarys, 1975) found a greater drag force below the surface, while another (Lyttle, Blanksby, Elliott, & Lloyd, 1998) found a greater drag at the surface.

Substantial differences were also found for the ratio of active to passive drag. Earlier studies reported values for active drag that were as much as double the values for passive drag (Clarys, 1978; diPrampero, Pendergast, Wilson, & Rennie, 1974). More recent studies (e.g. Kolmogorov & Duplishcheva, 1992; Shimonagata, Taguchi, Taba, & Aoyagi, 1999) reported roughly equivalent values for active and passive drag. Interpretation of the discrepancy in these findings ranged from "paradoxical" (Kolmogorov & Duplishcheva, 1992) to the claim that any difference between active and passive drag was "artificial" (Hollander et al., 1986). A disparity in results suggests that either different variables were actually measured or there was measurement error. In either case, measurement variability of a magnitude leading to completely different conclusions merits further study.

Sources of Variance

Measurement differences are due to true and error variance. An example of true variance (due to the expected factor effect) was shown as the difference between performance levels in the drag components of area and drag coefficient (Havriluk, 2005). Error variance can be random (i.e., there is a limitation in the precision of the measurement) or systematic (i.e., there is a problem with the measurement that always distorts the value in the same direction).

Random error occurs in every measurement, but systematic error may be controlled once it is identified. Systematic error is exemplified in two studies as: the difference in drag determined by two analysis methods (Berger, Hollander, & de Groot, 1996), and the difference in a drag component (area) determined by estimation and measurement (Cappaert, Gordon, & Frisbie, 1997).

There are many studies on swimming drag with similarities in experimental design and, therefore, similar potential sources of variance. A categorization of the different methodologies could account for true and error variance in the testing procedures and thereby accurately determine the effect of performance factors.

True Variance

Standardizing calculations of the component variables to produce the same resulting criterion can account for some of the variability in drag force measurement (Havriluk, 2005). Rearranging the drag equation shows the relationship of variables as: [C. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.