By Francesco Bartolucci
Statistical research of Questionnaires: A Unified strategy in accordance with R and Stata offers certain statistical tools for examining facts gathered via questionnaires. The e-book takes an utilized method of checking out and dimension projects, mirroring the starting to be use of statistical equipment and software program in schooling, psychology, sociology, and different fields. it truly is appropriate for graduate scholars in utilized data and psychometrics and practitioners in schooling, overall healthiness, and marketing.
The booklet covers the rules of classical try out idea (CTT), attempt reliability, validity, and scaling in addition to merchandise reaction concept (IRT) basics and IRT for dichotomous and polytomous goods. The authors discover the most recent IRT extensions, corresponding to IRT versions with covariates, multidimensional IRT types, IRT versions for hierarchical and longitudinal info, and latent type IRT types. in addition they describe estimation equipment and diagnostics, together with graphical diagnostic instruments, parametric and nonparametric checks, and differential merchandise functioning.
Stata and R software program codes are incorporated for every strategy. to augment comprehension, the booklet employs actual datasets within the examples and illustrates the software program outputs intimately. The datasets can be found at the authors’ net page.
Read or Download Statistical analysis of questionnaires : A unified approach based on R and Stata PDF
Similar probability & statistics books
Here's a sensible and mathematically rigorous advent to the sphere of asymptotic facts. as well as many of the ordinary issues of an asymptotics course--likelihood inference, M-estimation, the idea of asymptotic potency, U-statistics, and rank procedures--the ebook additionally provides contemporary study subject matters akin to semiparametric types, the bootstrap, and empirical methods and their functions.
The e-book offers commonly with 3 difficulties regarding Gaussian desk bound methods. the 1st challenge contains clarifying the stipulations for mutual absolute continuity (equivalence) of chance distributions of a "random procedure section" and of discovering powerful formulation for densities of the equiva lent distributions.
The publication goals to provide quite a lot of the most recent effects on multivariate statistical versions, distribution concept and functions of multivariate statistical tools. A paper on Pearson-Kotz-Dirichlet distributions via Professor N Balakrishnan comprises major result of the Samuel Kotz Memorial Lecture.
- Handbook of Nonresponse in Household Surveys
- Multivariate Statistics: A practical approach
- Mathematical Theory of Entropy
- Introductory Statistics for the Behavioral Sciences. Workbook
Additional info for Statistical analysis of questionnaires : A unified approach based on R and Stata
70. , redundant items). 90 should be regarded with suspicion. 60) indicate that items are very different from each other or are ambiguously defined. © 2016 by Taylor & Francis Group, LLC 35 Classical Test Theory Practically, the generalized Spearman–Brown coefficient is estimated as αˆ st = J cor , 1 + (J − 1)cor where cor = 1 J(J − 1) J J corj1 j2 j1 =1 j2 =1 j2 =j1 and corj1 j2 means the estimated correlation between the score on item j1 and the score on item j2 , that is, covj j corj1 j2 = √ 1 2 .
We also consider the (t) mean individual score for the overall test, indicated by y¯ ·· . As will be clear in the following, it may also be important to consider the correlation index between two test scores (or two halves of the same test or two scores on the same test repeated twice). This is defined as follows: cor(t1 t2 ) = (t1 ) n i=1 (yi· (t1 ) n i=1 (yi· (t ) (t ) (t ) − y¯ ·· 1 )(yi· 2 − y¯ ·· 2 ) (t ) − y¯ ·· 1 )2 (t2 ) n i=1 (yi· (t ) − y¯ ·· 2 )2 , t 1 = t2 . 4) Finally, given one or more random variables, we use the following notation throughout the book: E(·) for the expected value, V(·) for the variance, Cov(·) for the covariance, and Cor(·) for the correlation coefficient.
The assumption of equality of measurement units is relaxed within the model of congeneric tests, which, different from the previous models, assumes that the true scores at different tests are linearly related to one another so that any test true score can be expressed as a linear combination of any other test true score. To summarize, all three of the aforementioned CTT models assume a single underlying construct or latent dimension that is assessed by the set of T tests. However, these models are nested, with the model of parallel tests as the most restrictive and the model of congeneric test as the least restrictive one.
Statistical analysis of questionnaires : A unified approach based on R and Stata by Francesco Bartolucci