Measures of Association for Cross Classifications. II: Further Discussion and References

I remember it like it was yesterday: Age-related differences in the subjective experience of remembering

It has been frequently described that older adults subjectively report the vividness of their memories as being as high, or even higher, than young adults, despite poorer objective memory performance. Here, we review studies that examined age-related differences in the subjective experience of memory vividness. By examining vividness calibration and resolution, studies using different types of approaches converge to suggest that older adults overestimate the intensity of their vividness ratings relative to young adults, and that they rely on retrieved memory details to a lesser extent to judge vividness. We discuss potential mechanisms underlying these observations. Inflation of memory vividness with regard to the richness of memory content may stem from age-differences in vividness criterion or scale interpretation and psycho-social factors. The reduced reliance on episodic memory details in older adults may stem from age-related differences in how they monitor these details to make their vividness ratings. Considered together, these findings emphasize the importance of examining age-differences in memory vividness using different analytical methods and they provide valuable evidence that the subjective experience of remembering is more than the reactivation of memory content. In this vein, we recommend that future studies explore the links between memory vividness and other subjective memory scales (e.g., ratings of details or memory confidence) in healthy aging and/or other populations, as it could be used as a window to better characterize the cognitive processes that underpin the subjective assessment of the quality of recollected events.

The modified Stulberg classification is a strong predictor of the radiological outcome 20 years after the diagnosis of Perthes' disease

AIMS

The aim of this study was to assess the prognostic value of the modified three-group Stulberg classification, which is based on the sphericity of the femoral head, in patients with Perthes' disease.

METHODS

A total of 88 patients were followed from the time of diagnosis until a mean follow-up of 21 years. Anteroposterior pelvic and frog-leg lateral radiographs were obtained at diagnosis and at follow-up of one, five, and 21 years. At the five- and 21-year follow-up, the femoral heads were classified using a modified three-group Stulberg classification (round, ovoid, or flat femoral head). Further radiological endpoints at long-term follow-up were osteoarthritis (OA) of the hip and the requirement for total hip arthroplasty (THA).

RESULTS

There were 71 males (81%) and 17 females. A total of 13 patients had bilateral Perthes' disease; thus 101 hips were analyzed. At five-year follow-up, 37 hips were round, 38 ovoid, and 26 flat. At that time, 66 hips (65%) were healed and 91 (90%) were skeletally immature. At long-term follow-up, when the mean age of the patients was 28 years (24 to 34), 20 hips had an unsatisfactory outcome (seven had OA and 13 had required THA). There was a strongly significant association between the modified Stulberg classification applied atfive-year follow-up and an unsatisfactory outcome at long-term follow-up (p < 0.001). Between the five- and 21-year follow-up, 67 hips (76%) stayed in their respective modified Stulberg group, indicating a strongly significant association between the Stulberg classifications at these follow-ups (p < 0.001).

CONCLUSION

The modified Stulberg classification is a strong predictor of long-term radiological outcome in patients with Perthes' disease. It can be applied at the healing stage, which is usually reached five years after the diagnosis is made and before skeletal maturity. Cite this article: Bone Joint J 2021;103-B(12):1815-1820.

Correlation and association analyses in microbiome study integrating multiomics in health and disease

Correlation and association analyses are one of the most widely used statistical methods in research fields, including microbiome and integrative multiomics studies. Correlation and association have two implications: dependence and co-occurrence. Microbiome data are structured as phylogenetic tree and have several unique characteristics, including high dimensionality, compositionality, sparsity with excess zeros, and heterogeneity. These unique characteristics cause several statistical issues when analyzing microbiome data and integrating multiomics data, such as large p and small n, dependency, overdispersion, and zero-inflation. In microbiome research, on the one hand, classic correlation and association methods are still applied in real studies and used for the development of new methods; on the other hand, new methods have been developed to target statistical issues arising from unique characteristics of microbiome data. Here, we first provide a comprehensive view of classic and newly developed univariate correlation and association-based methods. We discuss the appropriateness and limitations of using classic methods and demonstrate how the newly developed methods mitigate the issues of microbiome data. Second, we emphasize that concepts of correlation and association analyses have been shifted by introducing network analysis, microbe-metabolite interactions, functional analysis, etc. Third, we introduce multivariate correlation and association-based methods, which are organized by the categories of exploratory, interpretive, and discriminatory analyses and classification methods. Fourth, we focus on the hypothesis testing of univariate and multivariate regression-based association methods, including alpha and beta diversities-based, count-based, and relative abundance (or compositional)-based association analyses. We demonstrate the characteristics and limitations of each approaches. Fifth, we introduce two specific microbiome-based methods: phylogenetic tree-based association analysis and testing for survival outcomes. Sixth, we provide an overall view of longitudinal methods in analysis of microbiome and omics data, which cover standard, static, regression-based time series methods, principal trend analysis, and newly developed univariate overdispersed and zero-inflated as well as multivariate distance/kernel-based longitudinal models. Finally, we comment on current association analysis and future direction of association analysis in microbiome and multiomics studies.

A modified accident analysis and investigation model for the general aviation industry: Emphasizing on human and organizational factors

INTRODUCTION

Currently, there is a lack of specific analytical tools for general aviation accidents (GAAs). This has led to loopholes in the prevention of GAAs.

METHODS

A Swiss Cheese model for general aviation (SCM-GA) is proposed to identify the human and organizational factors involved in GAAs. In the proposed SCM-GA, 5 categories, 45 subcategories, a general aviation safety management system (GA-SMS) and safety culture were developed based on the classic accident causation models combined with the laws and regulations and safety management practices in the general aviation industry.

RESULTS

One GAA was analyzed using SCM-GA. The human and organizational causes revealed by SCM-GA were more complete than the causes revealed through the accident report. The identification results of the deficiencies in the subcategories of GA-SMS and the safety culture were more consistent with the requirements in the general aviation laws and regulations than the organizational factors in the accident report. Based on the subcategories of SCM-GA, 41 GAAs that occurred between 1996 and 2010 in China were statistically analyzed and χ² test analyses were performed to estimate the statistical strength of the association between two adjacent subcategories of SCM-GA. The results showed that two adjacent subcategories of SCM-GA were significantly associated. They helped to determine the hidden problems in the accident report based on the path of accident.

CONCLUSIONS

SCM-GA is an accident analysis tool that can comprehensively analyze the human and organizational deficiencies involved in GAAs. The accident causes revealed by SCM-GA were more consistent with the general aviation safety management practices.

PRACTICAL APPLICATIONS

General aviation companies should establish their own GA-SMS and safety culture based on the subcategories developed herein. Using SCM-GA for routine safety inspection and accident investigation will help the management and the staff make effective safety decisions to effectively prevent GAAs.

Radiographic classifications in Perthes disease

Background and purpose - Different radiographic classifications have been proposed for prediction of outcome in Perthes disease. We assessed whether the modified lateral pillar classification would provide more reliable interobserver agreement and prognostic value compared with the original lateral pillar classification and the Catterall classification. Patients and methods - 42 patients (38 boys) with Perthes disease were included in the interobserver study. Their mean age at diagnosis was 6.5 (3-11) years. 5 observers classified the radiographs in 2 separate sessions according to the Catterall classification, the original and the modified lateral pillar classifications. Interobserver agreement was analysed using weighted kappa statistics. We assessed the associations between the classifications and femoral head sphericity at 5-year follow-up in 37 non-operatively treated patients in a crosstable analysis (Gamma statistics for ordinal variables, γ). Results - The original lateral pillar and Catterall classifications showed moderate interobserver agreement (kappa 0.49 and 0.43, respectively) while the modified lateral pillar classification had fair agreement (kappa 0.40). The original lateral pillar classification was strongly associated with the 5-year radiographic outcome, with a mean γ correlation coefficient of 0.75 (95% CI: 0.61-0.95) among the 5 observers. The modified lateral pillar and Catterall classifications showed moderate associations (mean γ correlation coefficient 0.55 [95% CI: 0.38-0.66] and 0.64 [95% CI: 0.57-0.72], respectively). Interpretation - The Catterall classification and the original lateral pillar classification had sufficient interobserver agreement and association to late radiographic outcome to be suitable for clinical use. Adding the borderline B/C group did not increase the interobserver agreement or prognostic value of the original lateral pillar classification.

Nonparametric Statistics: Use and Nonuse

This paper clarifies the distinctions among parametric, non-parametric, and distribution free statistical tests. The principal circumstances prompting the use of nonparametric procedures, both rank order and qualitative data Techniques, are identified and illustrated. Historical factors and disciplinary orientations affecting the use of nonparametric procedures are discussed. Finally, some conditions which are conducive to the more frequent use of nonparametric statistics are identified.

An Index for a Domain of Completion or Short Answer Items

Characteristics of the Lazarsfeld-Kendall “turnover” model adapted for completion or short answer achievement items are examined, and an index describing a domain of items based on this model is proposed. Estimation is discussed, and data that indicate the model is useful for the analysis of addition, subtraction, multiplication, and division items are presented.

An Exact Test of Significance for Goodman and Kruslkal's Gamma

Goodman and Kruskal's gamma is a convenient measure of association between ordered polytomies. It is here proposed to perform an exact test of significance of a given gamma by: (a) setting forth all possible sets of all frequencies of the matrix in question, leaving the marginals fixed; (b) computing gamma for each such arrangement; and (c) computing and sum ming the exact probabilities of all such sets where gamma is at least as great as that observed, with the same sign. The method here proposed thus differs somewhat from that of Freeman and Halton (1951); one example was significant according to their test at the .0024 level, but according to the method here proposed at the .1572 level.

Preserving differential privacy for similarity measurement in smart environments

Advances in both sensor technologies and network infrastructures have encouraged the development of smart environments to enhance people's life and living styles. However, collecting and storing user's data in the smart environments pose severe privacy concerns because these data may contain sensitive information about the subject. Hence, privacy protection is now an emerging issue that we need to consider especially when data sharing is essential for analysis purpose. In this paper, we consider the case where two agents in the smart environment want to measure the similarity of their collected or stored data. We use similarity coefficient function (FSC) as the measurement metric for the comparison with differential privacy model. Unlike the existing solutions, our protocol can facilitate more than one request to compute FSC without modifying the protocol. Our solution ensures privacy protection for both the inputs and the computed FSC results.

The Effect of the Raters' Marginal Distributions on Their Matched Agreement: A Rescaling Framework for Interpreting Kappa

Cohen's κ measures the improvement in classification above chance level and it is the most popular measure of interjudge agreement. Yet, there is considerable confusion about its interpretation. Specifically, researchers often ignore the fact that the observed level of matched agreement is bounded from above and below and the bounds are a function of the particular marginal distributions of the table. We propose that these bounds should be used to rescale the components of κ (observed and expected agreement). Rescaling κ in this manner results in κ', a measure that was originally proposed by Cohen (1960) and was largely ignored in both research and practice. This measure provides a common scale for agreement measures of tables with different marginal distributions. It reaches the maximal value of 1 when the judges show the highest level of agreement possible, given their marginal disagreements. We conclude that κ' should be used to measure the level of matched agreement contingent on a particular set of marginal distributions. The article provides a framework and a set of guidelines that facilitate comparisons between various types of agreement tables. We illustrate our points with simulations and real data from two studies-one involving judges' ratings of baseball players and one involving ratings of essays in high-stakes tests.

Latent Class Procedures: Applications to Organizational Research

In this overview, the authors use the seven studies included in the feature topic as a platform to delineate three areas that latent class procedures are particularly useful for in advancing the field of organizational research. The first topic area focuses on dealing with the need to identify and understand unobserved subpopulations in organizational research. The second topic area focuses on recognizing the unobserved heterogeneity in measurement functioning. The third topic area focuses on addressing the challenges surrounding the existence of multiple longitudinal change (both quantitative and qualitative) patterns in organizational research. The authors conclude this overview by highlighting further thoughts on the ways that latent class procedures should be utilized to advance organizational research.

Seasonal variations of air pollutant concentrations within Krasnoyarsk City

This paper examines the significant differences in seasonal variations of criteria pollutant concentrations in various parts of a large urban area. These differences are caused by the microclimatic heterogeneity of the city and show the influence of breeze and orographic-type circulations on urban air pollution. The temperature heterogeneity of Krasnoyarsk territory during the winter leads to an increase of 150% in CO air pollution levels in the central part of city. During the summer the orographical heterogeneity of Krasnoyarsk City leads to increases of up to 400% in air pollution for different areas.

Correlation of multidetector CT findings with splenic arteriography and surgery: prospective study in 392 patients

BACKGROUND

To determine the accuracy of contrast-enhanced multidetector CT (MDCT) in demonstrating splenic vascular injury based on results of splenic angiography and operation.

STUDY DESIGN

This institutional review board-approved study included 392 hemodynamically stable blunt trauma patients whose admission MDCTs demonstrated splenic injury. Images were assessed for parenchymal injury grade, hemoperitoneum volume, and evidence of bleeding and nonbleeding splenic vascular injury. Splenic arteriography was performed for high splenic injury grade and splenic vascular injury. Medical records were reviewed to determine arteriographic interpretation, surgery indications and findings, outcomes, and demographics. Sensitivity, specificity, predictive values, and accuracy of MDCT in detecting vascular injury were calculated based on results of arteriography and operation.

RESULTS

Splenic vascular injury was seen in 22% of patients (86 of 392) on MDCT. Presence of a vascular injury correlated with the CT-based parenchymal splenic injury grade (p < 0.0001). Active splenic bleeding was associated with subsequent clinical deterioration (p < 0.0001). Overall, MDCT had a sensitivity of 76% (76 of 100); specificity of 90% (95 of 106); negative and positive predictive values of 80% (95 of 119) and 87% (76 of 87), respectively; and accuracy of 83% (171 of 206) in detecting vascular injury compared with reference standards. The success rate of nonoperative management was 96%.

CONCLUSIONS

MDCT provides valuable information to direct initial clinical management of patients with blunt splenic trauma by demonstrating both active bleeding and nonbleeding vascular injuries. Not all vascular injuries are detected on MDCT, and splenic angiography is still indicated for high-grade parenchymal injury.

An evaluation of olive-tree bark for the biological monitoring of airborne trace-elements at ground level

There is a need to start looking into the possibility of tree bark for biomonitoring. Bark from olive trees (Olea europaea Linn.) has been collected over an area in northwestern Portugal, featuring a blend of rural lands, industrial towns and coastal environments. Samples were analysed through INAA and PIXE for their elemental contents. Results from both techniques were reconciled and then compared with an extensive database on 46 trace elements in Parmelia spp. thalli from the same sites and mostly from the very bark substrates. Distribution-free, nonparametric statistics show that, despite signal magnitude, variation patterns of bark and lichen concentrations follow one another in a most significant way. Selected elements in bark also correlate to a superior extent. As far as this investigation goes, there is no reason whatsoever for discarding bark as an alternative to lower epiphytes.

Husbands’ versus wives’ fertility goals and use of contraception: The influence of gender context in five Asian countries

Using data from Pakistan, India, Malaysia, Thailand, and the Philippines, we explore how gender context influences (1) husband-wife concordance in the demand for children and (2) the impact of each spouse’s fertility preferences on contraceptive use. We also explore whether the husband’s pronatalism can explain the wife’s unmet need for contraception. The results suggest that gender context has little net effect on couples’ concordance, but influences the relative weight of husbands’ and wives’ preferences in determining contraceptive use. Analysis of women’s unmet need for contraception suggests that the husbands’ pronatalism contributes to wives’ unmet need, but only to a relatively small degree, especially in settings where unmet need is high. This is the case because the proportion of couples with differing fertility goals is small in most communities.

Sample size in clinical trials with dichotomous endpoints: use of covariables

In many clinical trials, the primary endpoint is dichotomous. In this article, we examine the possibility of reducing the required sample size by removing variation associated with baseline covariables. Three measures are used to study the size of the reduction. Simulation studies based on a database of head trauma and of stroke patients suggested that a substantial reduction in the sample size can be achieved when the correlation between the endpoint and covariables is strong. A simple ad hoc formula for approximating the required sample size is proposed.

Clinical implications of cyclins, cyclin-dependent kinases, RB and E2F1 in squamous-cell lung carcinoma

In the search for new risk factors at the molecular and cellular levels, clinical data [lymph-node involvement(LN)and stage] were used and 104 squamous-cell lung carcinomas were analyzed by immuno-histochemistry for expression of cyclin D1, cyclin A, cdk2, cdk4, RB, and E2F1. The results of the univariate analysis of all 8 factors showed that cyclin A and cdk2 gave the best prognostic information, while no prognostic value could be found associated with cyclin D1, cdk4, RB and E2F1. The subsequent multivariate analysis of all possible combinations of the important factors showed that the pairs LN/cyclin A, LN/cdk2 and cyclin A/cdk2, and the triplet LN/cyclin A/cdk2 yielded the best prognostic information. It was essentially better than the information given by a single factor.

Spectrum and outcome of congestive heart failure in a hospitalized population

There are very few contemporary studies on the frequency and cause of congestive heart failure (CHF) in a general population. In western Sweden, inhabited by 1.64 million people, a retrospective survey was performed. All hospital records of patients with CHF, ages 16 through 65 years, were examined in all hospitals in the region. During the study period 2711 patients fulfilled the criteria for CHF or cardiomyopathy. Patients were monitored for 37 +/- 28 months. The most common cause of heart failure was coronary artery disease (IHD) (40%). Other common causes were hypertension (17%), valvular disease (13%), alcohol (11%), diabetes mellitus (10%), and systemic diseases (10%). There were positive correlations between the male sex and IHD, alcohol, and dilated cardiomyopathy; the female sex was associated with systemic diseases, valvular heart disease, and diabetes. The incidence of CHF requiring hospitalization per 100,000 in the population was 1.2 to 263 men and 1.1 to 129 women, in the youngest (age 16 to 30 years) and oldest (61 to 65 years) age groups, respectively. The 5-year survival rate was 50%. Analysis of causes performed with Cox's proportional hazards model for survival showed that age, IHD, alcohol, and diabetes were independent and powerful predictors of mortality (p < 0.001). The mode of death was progressive heart failure in 54% and sudden death in 26%. We concluded that the prognosis in patients with CHF was still very poor, even among this young population. The most common cause of CHF was IHD, and the second was hypertension.

Measurement of handedness via self-report: the relationship between brief and extended inventories

Three self-report measures of handedness were explored. For simple dichotomous classification into right- versus left-handed groups, a brief (4-item) and an extended (12-item) inventory produced 98.8% classification concordance in a sample of 250 subjects. When handedness was treated as a continuous variable, indexes based upon the short and expanded scales correlated highly (r = 0.95). A single-item handedness measure was also created which produced handedness classifications which were better than 97% concordant with the brief and extended handedness scales for dichotomous classification. The methodological implications of these findings for handedness measurement are discussed.

Measurement of reliability for categorical data in medical research

The problem of measuring reliability of categorical measurements, particularly diagnostic categorizations, is addressed. The approach is based on classical measurement theory and requires interpretability of the reliability coefficients in terms of loss of precision in estimation or power in statistical tests. A general model is proposed, leading to definition of reliability indices. Design and estimation approaches are discussed. Issues and approaches found in the research literature that either lead to confusing or misleading results are presented. The signs and symptoms of unreliable diagnoses are identified, and strategies for improving the reliability of such diagnoses are discussed.

Studies of lectin binding to the human cervix uteri: II. Cervical intraepithelial neoplasia and invasive squamous carcinoma

The cell surface carbohydrate profile of formalin-fixed paraffin-embedded tissue sections of neoplastic cervical squamous epithelium was evaluated using lectins of Bauhinia purpurea (BPA), Canavalin ensiformis (Con A), Griffonia simplicifolia I (GS I), Griffonia simplicifolia II (GS II), Maclura pomifera (MPA), Archis hypogaea (PNA), Glycine max (SBA), Ulex europaeus I (UEA I) and Triticum vulgaris (WGA). Three lectins (BPA, Con A and PNA) showed a similar pattern of staining in both normal squamous epithelium and in cervical intraepithelial neoplasia (CIN). Variable alterations were seen in lectin-binding patterns in CIN with seven lectins (GS I, GS II, MPA, PNA, SBA, UEA I and WGA). A significant difference was seen between the intensity of staining of normal squamous epithelium and CIN with all lectins except WGA. The alteration in GS II-binding pattern and intensity was significantly related to grade of CIN. No correlation was found between lectin binding and the presence of koilocytes in squamous epithelium. Cases of invasive squamous carcinoma showed a heterogeneous lectin-binding pattern and no significant association was found between lectin binding and tumour differentiation of patient survival. These results indicate that neoplasia in cervical squamous epithelium is associated with alterations in terminal alpha-Man residues, alpha- and beta-GalNAc residues, alpha- and beta-GlcNAc residues, alpha- and beta-Gal residues, and alpha-Fuc-containing residues, present in the outer parts of both N-linked and O-linked glycoconjugates. The implications of these findings are discussed.

Goodman and Kruskal's lambda: a new look at an old measure of association

We examine Goodman and Kruskal's lambda using Efron's approach to regression and analysis of variance (ANOVA) for zero-one outcome data. For a binary response cross-classified by a single nominal predictor, we present a computationally simple ANOVA table in which lambda is analogous to Pearson's R-square. We characterize the relationship between lambda and the commonly used apparent error rate in logistic regression, and show that lambda is based implicitly on a prediction rule for a saturated model with classification level 0.5. This relationship suggests that we can correct the apparent error rate for chance by defining a natural generalization of lambda that we call PRE, the proportional reduction in error. We illustrate the use of lambda and PRE in an analysis of prognostic factors for one-year survival in children with the acquired immunodeficiency syndrome (AIDS).