LEADER 00000cam 2200505 a 4500 |
008 121002s20129999th 000 0 eng d |
020 ^a9781420099911 |
020 ^a1420099914 |
099 ^aEP9^bTu883S |
100 1 ^aTu, Yu-Kang |
245 10 ^aStatistical thinking in epidemiology /^cYu-Kang Tu, MarkS. Gilthorpe |
260 ^aBoca Raton, FL :^bCRC Press,^cc2012 |
300 ^axii, 219 p. :^bill. ;^c25 cm |
504 ^aIncludes bibliographical references and index |
505 0 ^aVector geometry of linear models for epidemiologists --Path diagrams and directed acyclic graphs -- Mathematicalcoupling and regression to the mean in the relationbetween change and initial value -- Analysis of change inpre-/post-test studies -- Collinearity andmulticollinearity -- Is reverse paradox a paradox? --Testing statistical interaction -- Finding growthtrajectories in lifecourse research -- Partial leastsquares regression for lifecourse research |
520 ^a^"While biomedical researchers may be able to followinstructions in the manuals accompanying the statisticalsoftware packages, they do not always have sufficientknowledge to choose the appropriate statistical methodsand correctly interpret their results. StatisticalThinking in Epidemiology examines common methodologicaland statistical problems in the use of correlation andregression in medical and epidemiological research:mathematical coupling, regression to the mean,collinearity, the reversal paradox, and statisticalinteraction. Statistical Thinking in Epidemiology is aboutthinking statistically when looking at problems inepidemiology. The authors focus on several methods andlook at them in detail: specific examples in epidemiologyillustrate how different model specifications can implydifferent causal relationships amongst variables, andmodel interpretation is undertaken with appropriateconsideration of the context of implicit or explicitcausal relationships. This book is intended for appliedstatisticians and epidemiologists, but can also be veryuseful for clinical and applied health researchers whowant to have a better understanding of statisticalthinking. Throughout the book, statistical softwarepackages R and Stata are used for general statisticalmodeling, and Amos and Mplus are used for structuralequation modeling^"--Provided by publisher |
530 ^aAlso available online |
650 0 ^aEpidemiology^xStatistical methods |
650 0 ^aEpidemiology^xMathematics |
650 12 ^aEpidemiologic Methods |
650 12 ^aStatistics as Topic |
650 22 ^aModels, Statistical |
700 1 ^aGilthorpe, Mark S |
999 ^aอรทัย โคตรธาดา |