Home About us Contact
|
Home About us Contact | ||
|
|
||
Ordinary Least Squares Regression (ordinary + least_square_regression)
Selected AbstractsGamma regression improves Haseman-Elston and variance components linkage analysis for sib-pairsGENETIC EPIDEMIOLOGY, Issue 2 2004Mathew J. Barber Abstract Existing standard methods of linkage analysis for quantitative phenotypes rest on the assumptions of either ordinary least squares (Haseman and Elston [1972] Behav. Genet. 2:3,19; Sham and Purcell [2001] Am. J. Hum. Genet. 68:1527,1532) or phenotypic normality (Almasy and Blangero [1998] Am. J. Hum. Genet. 68:1198,1199; Kruglyak and Lander [1995] Am. J. Hum. Genet. 57:439,454). The limitations of both these methods lie in the specification of the error distribution in the respective regression analyses. In ordinary least squares regression, the residual distribution is misspecified as being independent of the mean level. Using variance components and assuming phenotypic normality, the dependency on the mean level is correctly specified, but the remaining residual coefficient of variation is constrained a priori. Here it is shown that these limitations can be addressed (for a sample of unselected sib-pairs) using a generalized linear model based on the gamma distribution, which can be readily implemented in any standard statistical software package. The generalized linear model approach can emulate variance components when phenotypic multivariate normality is assumed (Almasy and Blangero [1998] Am. J. Hum Genet. 68: 1198,1211) and is therefore more powerful than ordinary least squares, but has the added advantage of being robust to deviations from multivariate normality and provides (often overlooked) model-fit diagnostics for linkage analysis. Genet Epidemiol 26:97,107, 2004. © 2004 Wiley-Liss, Inc. [source] Can a publicly funded home care system successfully allocate service based on perceived need rather than socioeconomic status?HEALTH & SOCIAL CARE IN THE COMMUNITY, Issue 2 2007A Canadian experience Abstract The present quantitative study evaluates the degree to which socioeconomic status (SES), as opposed to perceived need, determines utilisation of publicly funded home care in Ontario, Canada. The Registered Persons Data Base of the Ontario Health Insurance Plan was used to identify the age, sex and place of residence for all Ontarians who had coverage for the complete calendar year 1998. Utilisation was characterised in two dimensions: (1) propensity , the probability that an individual received service, which was estimated using a multinomial logit equation; and (2) intensity , the amount of service received, conditional on receipt. Short- and long-term service intensity were modelled separately using ordinary least squares regression. Age, sex and co-morbidity were the best predictors (P < 0.0001) of whether or not an individual received publicly funded home care as well as how much care was received, with sicker individuals having increased utilisation. The propensity and intensity of service receipt increased with lower SES (P < 0.0001), and decreased with the proportion of recent immigrants in the region (P < 0.0001), after controlling for age, sex and co-morbidity. Although the allocation of publicly funded home care service was primarily based on perceived need rather than ability to pay, barriers to utilisation for those from areas with a high proportion of recent immigrants were identified. Future research is needed to assess whether the current mix and level of publicly funded resources are indeed sufficient to offset the added costs associated with the provision of high-quality home care. [source] Rejecting the mean: Estimating the response of fen plant species to environmental factors by non-linear quantile regressionJOURNAL OF VEGETATION SCIENCE, Issue 4 2005Henning K. Schröder Abstract Question: Is quantile regression an appropriate statistical approach to estimate the response of fen species to single environmental factors? Background: Data sets in vegetation field studies are often characterized by a large number of zeros and they are generally incomplete in respect to the factors which possibly influence plant species distribution. Thus, it is problematic to relate plant species abundance to single environmental factors by the ordinary least squares regression technique of the conditional mean. Location: Riparian herbaceous fen in central Jutland (Denmark). Methods: Semi-parametric quantile regression was used to estimate the response of 18 plant species to six environmental factors, 95% regression quantiles were chosen to reduce the impact of multiple unmeasured factors on the regression analyses. Results of 95% quantile regression and ordinary least squares regression were compared. Results: The standard regression of the conditional mean underestimated the rates of change of species cover due to the selected factor in comparison to 95% regression quantiles. The fitted response curves indicated a general broad tolerance of the studied fen species to different flooding durations but a narrower range concerning groundwater amplitude. The cover of all species was related to soil exchangeable phosphate and base-richness. A relationship between soil exchangeable potassium and species cover was only found for 11 species. Conclusion: Considering the characteristics of data sets in vegetation science, non-linear quantile regression is a useful method for gradient analyses. [source] Diversity in fertility patterns in GuatemalaPOPULATION, SPACE AND PLACE (PREVIOUSLY:-INT JOURNAL OF POPULATION GEOGRAPHY), Issue 6 2006Sofie De Broe Abstract This study investigates urban and rural fertility trends in Guatemala up to 2002. It also aims to establish, using the theory of diffusion as its theoretical framework, the extent to which ethnicity and ethnic diversity are associated with geographical patterns in local-level fertility after controlling for socio-economic indicators. Data from the Demographic and Health Surveys of 1987, 1995,96 and 1998,99, the National Maternal and Child Health Survey of 2002 and the Census of 2002 were used. P/F ratios were calculated and used as an analytical tool and quality control measure of the data in order to establish the timing of changes in fertility patterns as measured by age-specific fertility rates (ASFRs) based on exact exposure in four-year periods from 1972 to 2002. Finally, the 2002 census data were used to analyse and model fertility at the municipio- level using ordinary least squares regression. The results suggest a steady but very slow decline in fertility from 1972 until the mid-1990s. Both the P/F ratios and ASFRs calculated using the Maternal and Child Health Survey and Census of 2002 show a sharp decline in fertility since 1998. The regression results for the census data suggest an independent and significant effect of ,proportion of indigenous people' and an almost significant effect of ethnic diversity on fertility at the municipio -level. The very slow decline in fertility in Guatemala until fairly recently can be attributed to the fact that Guatemala has been lagging behind in terms of socio-economic development and the additional challenge of having a culturally very diverse and segregated population, preventing the spread of modern reproductive ideas and behaviour. The accelerated fertility decline since the end of the 1990s seems likely to be associated with the widespread availability and increased uptake of family planning following declining fertility desires among its indigenous population. Copyright © 2006 John Wiley & Sons, Ltd. [source] Inflation of Type I error rate in multiple regression when independent variables are measured with errorTHE CANADIAN JOURNAL OF STATISTICS, Issue 1 2009Jerry Brunner MSC 2000: Primary 62J99; secondary 62H15 Abstract When independent variables are measured with error, ordinary least squares regression can yield parameter estimates that are biased and inconsistent. This article documents an inflation of Type I error rate that can also occur. In addition to analytic results, a large-scale Monte Carlo study shows unacceptably high Type I error rates under circumstances that could easily be encountered in practice. A set of smaller-scale simulations indicate that the problem applies to various types of regression and various types of measurement error. The Canadian Journal of Statistics 37: 33-46; 2009 © 2009 Statistical Society of Canada Lorsque les variables indépendantes sont mesurées avec erreur, la régression des moindres carrés ordinaires peut conduire à une estimation biaisée et incohérente des paramètres. Cet article montre qu'un accroissement de l'erreur de type I peut aussi se produire. En plus de résultats analytiques, une étude par simulations Monte-Carlo de grande envergure montre que, dans certaines conditions que nous pouvons rencontrer facilement en pratique, l'erreur de type I peut être trop élevée. Une autre étude de Monte-Carlo de moindre envergure suggère que ce problème se rencontre aussi dans plusieurs types de régression et différents types d'erreur de mesure. La revue canadienne de statistique 37: 33-46; 2009 © 2009 Société statistique du Canada [source] Coefficient shifts in geographical ecology: an empirical evaluation of spatial and non-spatial regressionECOGRAPHY, Issue 2 2009L. Mauricio Bini A major focus of geographical ecology and macroecology is to understand the causes of spatially structured ecological patterns. However, achieving this understanding can be complicated when using multiple regression, because the relative importance of explanatory variables, as measured by regression coefficients, can shift depending on whether spatially explicit or non-spatial modeling is used. However, the extent to which coefficients may shift and why shifts occur are unclear. Here, we analyze the relationship between environmental predictors and the geographical distribution of species richness, body size, range size and abundance in 97 multi-factorial data sets. Our goal was to compare standardized partial regression coefficients of non-spatial ordinary least squares regressions (i.e. models fitted using ordinary least squares without taking autocorrelation into account; "OLS models" hereafter) and eight spatial methods to evaluate the frequency of coefficient shifts and identify characteristics of data that might predict when shifts are likely. We generated three metrics of coefficient shifts and eight characteristics of the data sets as predictors of shifts. Typical of ecological data, spatial autocorrelation in the residuals of OLS models was found in most data sets. The spatial models varied in the extent to which they minimized residual spatial autocorrelation. Patterns of coefficient shifts also varied among methods and datasets, although the magnitudes of shifts tended to be small in all cases. We were unable to identify strong predictors of shifts, including the levels of autocorrelation in either explanatory variables or model residuals. Thus, changes in coefficients between spatial and non-spatial methods depend on the method used and are largely idiosyncratic, making it difficult to predict when or why shifts occur. We conclude that the ecological importance of regression coefficients cannot be evaluated with confidence irrespective of whether spatially explicit modelling is used or not. Researchers may have little choice but to be more explicit about the uncertainty of models and more cautious in their interpretation. [source] Nurse and resident satisfaction in magnet long-term care organizations: do high involvement approaches matter?JOURNAL OF NURSING MANAGEMENT, Issue 3 2006KENT V. RONDEAU PhD Aim, This study examines the association of high involvement nursing work practices with employer-of-choice (magnet) status in a sample of Canadian nursing homes. Background, In response to a severe shortage of registered nursing personnel, it is imperative for health care organizations to more effectively recruit and retain nursing personnel. Some long-term care organizations are developing employee-centred cultures that allow them to effectively enhance nurse and resident satisfaction. At the same time, many nursing homes have adopted progressive nursing workplace practices (high involvement work practices) that emphasize greater employee empowerment, participation and commitment. Method, A mail survey was sent to the director of nursing in 300 nursing homes in western Canada. In total, 125 useable questionnaires were returned and constituted the data set for this study. Separate ordinary least squares regressions are performed with magnet strength, nurse satisfaction and resident satisfaction used as dependent variables. Results, Nursing homes that demonstrate strong magnet (employer-of-choice) characteristics are more likely to have higher levels of nurse and patient satisfaction, even after controlling for a number of significant factors at the establishment level. Magnet nursing homes are more likely to have progressive participatory decision-making cultures and much more likely to spend considerable resources on job-related training for their nursing staff. The presence of high involvement work practices is not found to be a significant predictor in magnet strength, nurse or resident satisfaction. Conclusion, Merely adopting more high involvement nursing work practices may be insufficient for nursing homes, which desire to become ,employers-of-choice' in their marketplaces, especially if these practices are adopted without a concomitant investment in nurse training or an enhanced commitment to establishing a more democratic and participatory decision-making style involving all nursing staff. [source] | |||||||||||||