scispace - formally typeset
Search or ask a question

Showing papers on "Intraclass correlation published in 1990"


Journal ArticleDOI
TL;DR: It is shown that neither technique is appropriate for assessing the interchangeability of measurement methods, and an alternative approach based on estimation of the mean and standard deviation of differences between measurements by the two methods is described.

675 citations


Journal ArticleDOI
Susan Chinn1
TL;DR: Criteria are given for the choice of scale prior to estimation of repeatability, and recommendations of Bland and Altman are used for expressing repeatability and agreement of methods of measurement on the same scale.
Abstract: Criteria are given for the choice of scale prior to estimation of repeatability. Recommendations of Bland and Altman should then be used for expressing repeatability and agreement of methods of measurement on the same scale. Repeatability of measurements on different scales should be compared using the appropriate ratio of variances, or intraclass correlation coefficient. A reference range for diagnosis requires a high ratio of between-subject variation to total variation. The index of separation between diseased and healthy subjects should be used whenever possible. Changes within patients should be compared with reference change ranges, and not against the diagnostic range.

162 citations


Journal ArticleDOI
TL;DR: The use of the echo-Doppler technique mainly for the determination of rapid and large changes in portal hemodynamics within a short period of time is supported.

160 citations


Journal ArticleDOI
TL;DR: In this article, the authors compared responses to interview questions among 492 individuals interviewed more than once in a hospital-based case-control surveillance system in the United States, Canada, and Israel between 1976 and 1982.
Abstract: Responses to interview questions were compared for concordance among 492 individuals interviewed more than once in a hospital-based case-control surveillance system in the United States, Canada, and Israel between 1976 and 1982. Reliability of the data was determined using the Kappa statistic and the intraclass correlation coefficient. Reliability was good to excellent for demographic factors, such as birthplace, and for medical conditions/procedures that require hospitalization or continuing medical care, such as hysterectomy. Reliability was fair to good for less serious or less well-defined medical conditions/procedures, such as cystic breast disease, and for current habits, such as daily coffee consumption. Regarding medication use, reliability was poor to fair for drugs taken intermittently, such as aspirin and penicillin, and good to excellent for drugs taken on a regular basis, such as oral contraceptives. As expected, medications were reported more consistently when duration of use was prolonged. The data were also analyzed according to two intervals between interviews (less than 1 year and greater than or equal to 1 year). For most factors, reliability was not materially affected by interval. Where differences were observed, reliability tended to be better when the second interview followed the first by less than 1 year. These results suggest that structured interviews administered to hospital patients by trained personnel can elicit reliable data on demographic and medical history factors.

121 citations


Journal ArticleDOI
TL;DR: In this article, a generalized model for estimating intraclass correlation in the analysis of familial data is proposed. But the model is not suitable for the analysis with multiple classes of families, and the maximum likelihood estimates of the parameters of the model are not known.
Abstract: SUMMARY We introduce a generalized model for estimating intraclass correlation in the analysis of familial data. This model incorporates the models of Donner & Koval (1980), Donner & Bull (1983), Rosner (1984) and Miunoz, Rosner & Carey (1986) as special cases. We then obtain the maximum likelihood estimates of the parameters of the model. In particular we obtain a single estimating equation, to be solved iteratively, for the estimation of the intraclass correlation p and discuss all the special cases. Asymptotic variance for the estimate of the intraclass correlation is obtained for all the models and some examples are given.

19 citations


Journal ArticleDOI
TL;DR: In this article, two methods of estimating the intraclass correlation coefficient (p) for the one-way random effects model were compared in several simulation experiments using balanced and unbalanced designs.
Abstract: Two methods of estimating the intraclass correlation coefficient (p) for the one-way random effects model were compared in several simulation experiments using balanced and unbalanced designs. Estimates based on a Bayes approach and a maximum likelihood approach were compared on the basis of their biases (differences between estimates and true values of p) and mean square errors (mean square errors of estimates of p) in each of the simulation experiments. The Bayes approach used the median of a conditional posterior density as its estimator.

17 citations



Journal ArticleDOI
TL;DR: In this article, the authors point out that these two indices reflect two different aspects of interobserver reliability and that both aspects should be reported on the same data and suggest that Cohen's kappa and the intraclass correlation be used in their place.
Abstract: Research in infant cognition has relied heavily on designs in which the dependent variable is visual fixation and the measuring instrument is a human observer. The reliability of these data typically is expressed either as the average percent agreement or the Pearson correlation between two independent observers. In this commentary, we point out that these two indices reflect two different aspects of interobserver reliability and that both aspects should be reported on the same data. Moreover, we illustrate how average percent agreement and the Pearson correlation can produce misleading, usually spuriously high, estimates of interobserver reliability and suggest that Cohen's kappa and the intraclass correlation be used in their place.

12 citations


Journal ArticleDOI
TL;DR: Calculation of intraclass correlation coefficients for such data can indicate the degree of variability present in multiple samples of crevicular fluid collected from individual patients, provide information about the source of host mediators in the fluid, and help identify appropriate sampling strategies for the fluid.

12 citations


Journal ArticleDOI
TL;DR: In this article, a computer program that uses the intraclass correlation to estimate inter-rater reliability is presented. But the program is not suitable for the use of more than two judges.
Abstract: INTRACLS is a computer program that uses the intraclass correlation to estimate inter-rater reliability. The utility of the intraclass correlation lies in its capability of estimating the reliability of ratings using more than two judges. INTRACLS computes reliability using analysis of variance (Winer, 1971, p. 288), supplemented with the formulas for intraclass correlation found in Armstrong (1981). The results may be interpreted within the framework of generalizability theory (Cronbach, Gleser, Nanda, & Rajaratnam, 1972). The program computes the F ratio, which permits testing for significance (see discussion on assumptions in Tinsley & Weiss, 1975).

3 citations


Journal ArticleDOI
TL;DR: A rating scale, the Psychological Skills Inventory, was completed by a sample of parents of Head Start children, as well as by research assistants who observed these children in classrooms as mentioned in this paper.
Abstract: A rating scale, the Psychological Skills Inventory, was completed by a sample of parents of Head Start children, as well as by research assistants who observed these children in classrooms. The scale yields a total score, plus a subscale score for each of 22 psychological skills. The scale demonstrated high internal consistency (coefficient alpha = .97) and good interrater reliability among classroom observers (intraclass correlation = .76). The pattern of correlations with other measures of parent and child behavior and mental health suggests that the total score measures overall adjustment or psychological health of the child. Individual skills as measured by the inventory were fairly reliable (average intraclass correlation = .58). The reliability of the subscales in measuring individual psychological skills provides evidence that the scale can be useful in constructing individual treatment plans for children.

Journal ArticleDOI
TL;DR: The argument against the use of the ICC to evaluate the clinical significance of any error in an SMBG system is that two SMBG measurements that deviate by the same amount from their true values may have very different clinical meanings depending on the true values.
Abstract: where BMS is the between-blood sample mean square, EMS is the residual mean square, MMS is the betweenmethod mean square, k is the number of methods used to measure blood glucose (2 in this case), and n is the number of blood samples examined. The resulting ICC for the data in Table 2A is 0.97, which indicates a high level of agreement between the SMBG and laboratory measurements. One of our reservations with the ICC is that the size of the coefficient is influenced greatly by the amount of variability in the blood glucose measurements. The data in Table 2, B and C, illustrate how a reduction in the range of blood glucose values leads to a reduction in the ICC. (Note that these are extreme examples used only to illustrate what happens when a restricted range of blood values is examined; studies examining the performance of SMBG devices should include a representative range of blood glucose values.) The ICCs for the data in Table 2, B and C, are 0.83 and 0.01, respectively, despite the fact that the laboratory measurements are 20 U higher than the SMBG measurements in each example. The second problem with the ICC is that it provides little information about the clinical significance of the errors associated with SMBG. The issues of statistical and clinical significance have been discussed by various investigators in reference to the use of the Pearson r and regression analyses and apply equally to the use of the ICC (4-7). The crux of the argument against the use of the ICC to evaluate the clinical significance of any error in an SMBG system is that two SMBG measurements that deviate by the same amount from their true values may have very different clinical meanings depending on the true values. For example, a 50% underestimate of the true value of 2.8 mM blood glucose will not make any difference in terms of the treatment actions taken by a patient. Under these circumstances, the patient will take corrective action to raise blood glucose. However, a 50% underestimate when the true value is 16.8 mM will result in the patient not taking action to lower a clinically elevated blood glucose level. In view of these comments, we do not recommend the ICC as the primary method for evaluating the quality of the measurements from self-monitoring systems. Currently, the most useful method for evaluating the data from self-monitoring systems is the error-grid analysis graphic display technique developed by Cox et al. (4-6).

Journal ArticleDOI
01 Oct 1990
TL;DR: Using two job analysis data sets, it is found that ICCs cannot be interpreted unambiguously as indices of interrater agreement because they are substantially affected by the level of between-group variance.
Abstract: Intraclass correlation coefficients (ICCs) often are used to index the level of interrater agreement for job analysis ratings. Typically, ICCs are calculated from a one-way ANOVA design for each of several rated job characteristics in which multiple job analysts rate multiple jobs. ICCs are highest when between mean square (BMS) is large relative to within mean square (WMS), but they can be low either when (a) WMS is large, in which case ICCs appropriately indicate low interrater agreement, or (b) BMS is small, in which case ICCs reflect the attenuating influence of range restriction on the grouping factor (i.e., jobs). Using two job analysis data sets, we found that (a) ICCs covaried significantly and negatively with WMS components and (b) ICCs covaried significantly and positively with BMS. These restults indicate that ICCs cannot be interpreted unambiguously as indices of interrater agreement because they are substantially affected by the level of between-group variance. Interpretational confounding fo...