Showing papers on "Intraclass correlation published in 1986"

A Review of Inference Procedures for the Intraclass Correlation Coefficient in the One-Way Random Effects Model

Abstract: Summary Recent theory and methodology for inferences concerning the intraclass correlation coefficient are reviewed, under the assumption of an underlying random effects model. Topics discussed include point and interval estimation, significance-testing for nonzero values of the intraclass correlation, and inference procedures in multiple samples.

Percent Agreement, Pearson's Correlation, and Kappa as Measures of Inter-examiner Reliability

Abstract: Percent agreement and Pearson's correlation coefficient are frequently used to represent inter-examiner reliability, but these measures can be misleading. The use of percent agreement to measure inter-examiner agreement should be discouraged, because it does not take into account the agreement due solely to chance. Caution must be used in the interpretation of Pearson's correlation, because it is unaffected by the presence of any systematic biases. Analyses of data from a reliability study show that even though percent agreement and kappa were consistently high among three examiners, the reliability measured by Pearson's correlation was inconsistent. This study shows that correlation and kappa can be used together to uncover non-random examiner error.

Reliability of a Craniomandibular Index

Abstract: The Craniomandibular Index (CMI) was developed to provide a standardized measure of severity of problems in mandibular movement, TMJ noise, and muscle and joint tenderness for use in epidemiological and clinical outcome studies. The instrument was designed to have clearly defined objective criteria, simple clinical methods, and ease in scoring; it is divided into the Dysfunction Index and the Palpation Index. Inter-rater reliability (three raters) and intra-rater reliability (19 patients examined twice by one rater) were tested to determine whether the instrument has operational definitions sufficiently precise to allow for consistency in use between different raters and with one rater over time. Intraclass Correlation Coefficient for inter-rater reliability was 0.84 for the Dysfunction Index, 0.87 for the Palpation Index, and 0.95 for the CMI. Correlation for intra-rater reliability was 0.92 for the Dysfunction Index, 0.86 for the Palpation Index, and 0.96 for the CMI. These results support the reliability of the CMI for use in epidemiological and clinical studies. Users are cautioned about the subjectivity of numerous items within the CMI and the strict methodological guidelines that must be followed in order to assure accuracy and reproducibility of results.

A comparison of confidence interval methods for the intraclass correlation coefficient

Abstract: Different methods of obtaining confidence intervals for the intraclass correlation coefficient rho in the unbalanced one-way random-effects model are investigated, focusing on applications to family studies. Methods based on simple modifications of formulas for the case of equal group sizes are found to provide adequate coverage at small to moderate values of rho. A method based on the large-sample standard error of the sample intraclass correlation, as derived by Smith (1956, Annals of Human Genetics 21, 363-373), is shown to provide consistently good coverage at all values of rho. A method proposed by Thomas and Hultquist (1978, Annals of Statistics 6, 582-587) also provides consistently good coverage, but generates mean interval widths substantially greater than those generated by Smith's method at values of rho likely to arise in practice.

Reliability and validity of four instruments for measuring lumbar spine and pelvic positions.

Abstract: We studied the between-therapist reliability and the validity of four instruments in measuring lumbar spine curvature and pelvic tilt. The four instruments and their measurements were 1) a tape measure to measure the change in lumbar curvature during trunk flexion; 2) a gravity goniometer to measure pelvic angle and lumbar curvature during stance, trunk flexion, and trunk extension; 3) a parallelogram goniometer to measure lumbar curvature during stance, trunk flexion, and trunk extension; and 4) a standard goniometer to measure the angle between wooden pointers mounted perpendicularly to the spine to obtain pelvic angle and lumbar curvature during stance, trunk flexion, and trunk extension. We found no single instrument to be the most reliable or valid. Between-therapist reliability ranged from .64 to .93 (Pearson product-moment correlation) and from .60 to .92 (interclass correlation coefficient). The validities of the instruments compared with measurements from roentgenograms generally were low, ranging from -.13 to .76 (Pearson product-moment correlation) and -.73 to -.05 (interclass correlation coefficient).

Taste intensity perception in aging.

Abstract: Men (n = 91) and women (n = 79) between the ages of 23 and 88 years made cross-modal matches of distance to the perceived intensity of aqueous solutions representing the four basic taste qualities. Traditional analysis, which characterizes age-group performance in terms of the slope of a group's psychophysical function, yielded results consistent with previous findings in that some flattening of slope was observed in the oldest group. Analyses based on the slope of the functions generated by individual participants demonstrated significant regression on age for some qualities. None of these findings were observed in the analogous analysis of variance (ANOVA). Significant age-related changes in Intraclass Correlation Coefficient (ICC) were demonstrated by both ANOVA and regression analysis. The ICC reflects the consistency of an individual's discriminative response over repeated stimulus presentations. Its variation with age was dependent upon taste quality.

Regression analysis in the presence of heterogeneous intraclass correlations.

Abstract: In the statistical analysis of twinship and familial data, one often encounters the need for regression methods that control for the possibly different aggregation of the twins and the members of the families according to their type (e.g., monozygotic, dizygotic for twinship data). We present the maximum likelihood solution for the regression analysis of data in the presence of heterogeneous intraclass correlations. This work extends previous results for the case of a homogeneous intraclass correlation. An application of the methods for the analysis of twinship data is included.

A procedure for evaluating the reliability of a gingivitis index.

Abstract: A methodology is presented for assessing the reliability of an ordinarily-scaled index and is illustrated by using data from a clinical trial in which gingival inflammation was assessed with the PMGI index, independently, by 5 examiners. One of the examiners was an experienced examiner, the others newly trained. All subjects were evaluated by each examiner initially and at the end of the study period. The reliability of the average score per subject, maximum score per subject, and the % of affected sites per person are estimated by the intraclass correlation coefficient. Procedures are presented that utilize various forms of the weighted kappa statistic for dissecting patterns in examiner agreement for specific sites, types of site, all sites, and for the individual components and categories of the index. It is shown how these procedures can be useful for training and calibrating multiple examiners, who will be using such an index in a clinical study, so that adequate reliability levels can be realized.

Confidence intervals on the intraclass correlation in the unbalanced one-way classification

Abstract: Several methods are compared for constructing confidence intervals on the intraclass correlation coefficient in the unbalanced one-way classification. The results suggest that a conservative approximation of the exact procedure developed by Wald (1940) can be used for hand calculations, When the exact solution is desired, a solution procedure is recommended that is computationally convenient and allows the investigator to determine the precision of the estimate. In cases where a prior estimate of the correlation is available, researchers may select intervals based on either the analysis of variance or unweighted sums of squares estimator.

Effects of Intraclass Correlation on Weighted Averages

Abstract: Weighted averages of class means using different sets of weighting factors are compared in terms of sampling variances and of relative weights given to the class means. Details are given for the 1-way classification, and extensions to other models are indicated.

General Derivation of Intraclass Correlation Coefficients

Abstract: A general intraclass correlation coefficient is derived similarly to the general correlation coefficient given by KENDALL (1962). This coefficient II embraces as specific cases three intraclass correlations, related to Pearson's r, Kendall's tau and Spearman's rs, that were dealt with by FISHER (1921), WHITFIELD (1949), SHIRAHATA (1981) and SCHEMPER (1984). Formal relationships are presented and the qualification of the three coefficients for specific applications is discussed.

Intraclass correlations of periodontal measurements in beagle dogs.

Abstract: The intraclass correlations of Gingival Index, probing depth, and attachment level scores were estimated in a sample of beagle dogs. These estimates were made over a ten-week period at the time of ligature placement, eight weeks later at ligature removal, and two weeks postremoval. Estimates of the intraclass correlation were computed for the actual scores and for differences in scores between the first and second, and second and third observation points. All intraclass correlations were of sufficient magnitude to preclude statistical analyses which do not consider the subject as the sampling unit. For all conditions. intraclass correlations decreased upon experimental intervention and were generally larger than those reported in human studies. Intraclass correlations of Gingival Index scores were larger than those of probing depth and attachment level measurements. The results were discussed in terms of sources of variation within and between subjects.

A comparison of estimators of a common correlation coefficient

Abstract: It is often of interest to test the hypothesis that all off-diagonal elements of the correlation matrix of a multivariate normal distribution are equal. If the hypothesis of equal correlation can be accepted, it then may be of interest to estimate the common correlation coefficient. In this paper, four estimators of the common correlation are compared in terms of bias, variance, mean squared error, adequacy of the normal approximation, and ease of calculation. The average sample correlation is seen to be comparable to the other estimators and is recommended here since it is the easiest to calculate. The estimators are compared using simulation.