For the research on test and measurement, the evaluation of measurement error is inevitable. The basic data for evaluation are obtained by the process of measurement, and also, for better evaluation, it is very important to evaluate the measurement error and causes of error. At the evaluating processes, such error must be considered reasonably. Therefore, the error which is included in flexibility measurement is discussed in this paper by the way of analysis of variance. At the same time, by the way of analysis of variance, the objectivity of flexibility measurement which is done by angular-method is discussed, too. In other words, through the computation of measurement error which is caused by repetition of measurement carried by the same testee and also through the computation of measurement error which is caused by repetition of measurement carried by the different testers, the measurement error is investigated.
1) The measurement error of trunk flexion is 1.7 degree for angular-method and 2.0cm for distance-method at the five repetitions, and also 1.5 degree and 1.7cm at the three repetitions, respectively.
2) For the trunk flexion measurement, such measurement values as dropping into the interval
x±1.7 degree is thought as the good representative value of measurement. This interval is called as a sort of critical region.
3) Through the above mentioned investigation, the measurement error affects the first digit of degree unit and also first digit of cm unit, respectively. Therefore, it is good enough only to read the degree unit and cm unit at angular-method and distance-method, respectively.
4) At the comparison of measurement error between trunk flexion and extension, it is the least for trunk flexion and the most for trunk extension, and medium for trunk sideward bend. Furthermore, the extent to which the measurement error is included in the mean value of each measurement item tends to show the same thing.
5) At the comparison of measurement error which is caused at three-times-repeated and fivetimes-repeated measurement of angular-method the significant difference of measurement error could not be found. However, in general, the measurement error ranging from minimum, 1.5 degree to maximum, 3.5 degree is found. Minimum error, 1.5 degree was found at trunk flexion, and maximum one, 3.5 degree was found at trunk extension.
6) Analysis of variance was applied to the trunk flexion, extension and sideward bend data, the differences of measurement error due to different testees are not significant statistically, at significance level 0.01 and 0.05. In other words, any difference between testees do not influence to measurment error variation.
7) And also, the differences of measurement error due to different testers are not significant statistically at significance level, 0.01 and 0.05. In other words, the extent of measurement error variation does not influence to the measures significantly. Therefore, this means that this test is highly objective.
抄録全体を表示