Establishing Some Measures of Absolute and Relative Reliability of a Motor Test / Određivanje nekih mjera apsolutne i relativne pouzdanosti motoričkih testova

Abstract

Abstract 

Relative and absolute reliability are discussed in this paper on the bases of some empirical motor data. Relative reliability was assessed via the calculation of intraclass correlation coefficients (ICC; Shrout & Fleiss, 1979; Nunnally & Berstein, 1994). Absolute reliability was assessed by calculating standard error of measurement (SEM), and additionally by calculating the smallest detectable change (SDC; de Veet et al., 2006a, b), a relatively new measure which originates from clinical disciplines but has an ever-growing use in other areas. Bland-Altman method (1986, 1995) for determining the limits of agreement and bias between two measurements, was also used. ICC coefficients were high with narrow limits of confidence but ICC masked some differences in trials revealed by SEM and Bland-Altman technique. As stated in Hopkins (2000), and Atkinson and Nevill (1998), more than one measure of reliability should be provided in reliability studies.

Key words: Bland-Altman limits of agreement; intraclass correlation; reliability of the measurement, smallest detectable change (SDC); standard error of the measurement (SEM). 

---

Sažetak

Rad u kojem se raspravlja o relativnoj i apsolutnoj pouzdanosti utemeljen je na empirijskim motoričkim podacima. Relativna pouzdanost određena je izračunavanjem koeficijenata intraklas korelacije (ICC; Shrout i Fleiss, 1979; Nunnally i Berstein, 1994). Određivanje apsolutne pouzdanosti izvedeno je izračunavanjem standardne pogreške mjerenja (SEM) i dodatno izračunavanjem najmanjih uočljivih promjena (smallest detectable change; SDC; de Veet i sur., 2006a,b), relativno nove mjere koja potječe iz kliničkih disciplina, a sve veću primjenu nalazi i u drugim područjima. U radu je upotrijebljen i Bland-Altmanova metoda (1986, 1995) određivanja razina slaganja (limits of agreement; LOA), odnosno biasa između dva mjerenja. Dobiveni su visoki ICC koeficijenti i suženi limiti pouzdanosti, no ICC koeficijenti su prikrili neke razlike u ponovljenim izvedbama testa koje su uočene primjenom SEM i Bland-Altman metode. Kao što su naveli Hopkins (2000), Atkinson i Nevill (1998) u istraživanjima pouzdanosti treba primijeniti više mjera. 

Ključne riječi: Bland-Altmanove razine slaganja (LOA); intraklas korelacija; najmanja uočljiva promjena (SDC); pouzdanost mjerenja, standardna pogreška mjerenja (SEM)