Kako odabrati pravi test za procjenu statističke značajnosti razlike između skupina?

Marusteri, Marius; Bacarea, Vladimir
April 2010
Biochemia Medica;2010, Vol. 20 Issue 1, p15
Academic Journal
Choosing the right statistical test may at times, be a very challenging task for a beginner in the field of biostatistics. This article will present a step by step guide about the test selection process used to compare two or more groups for statistical differences. We will need to know, for example, the type (nominal, ordinal, interval/ratio) of data we have, how the data are organized, how many sample/groups we have to deal with and if they are paired or unpaired. Also, we have to ask ourselves if the data are drawn from a Gaussian on non-Gaussian population. A key question is, if the proper conditions are met, should a one-tailed test or two-tailed test be used, the latter typically being the most powerful choice. The appropriate approach is presented in a Q/A (Question/Answer) manner to provide to the user an easier understanding of the basic concepts necessary to fulfill this task. Some of the necessary fundamental concepts are: statistical inference, statistical hypothesis tests, the steps required to apply a statistical test, parametric versus nonparametric tests, one tailed versus two tailed tests etc. In the final part of the article, a test selection algorithm will be proposed, based on a proper statistical decision-tree for the statistical comparison of one, two or more groups, for the purpose of demonstrating the practical application of the fundamental concepts. Some much disputed concepts will remain to be discussed in other future articles, such as outliers and their influence in statistical analysis, the impact of missing data and so on.


Related Articles

  • Significance Testing in Theory and Practice. Greco, Daniel // British Journal for the Philosophy of Science;Sep2011, Vol. 62 Issue 3, p607 

    Frequentism and Bayesianism represent very different approaches to hypothesis testing, and this presents a skeptical challenge for Bayesians. Given that the most empirical research uses frequentist methods, why (if at all) should we rely on it? While it is well known that there are conditions...

  • Estimation of Normal Variance Based on Balanced and Unbalanced Ranked Set Samples. Yu, Philip; Lam, Kin; Sinha, Bi // Environmental & Ecological Statistics;Apr1999, Vol. 6 Issue 2, p23 

    In this paper we address the problem of estimation of the variance of a normal population based on a balanced as well as an unbalanced ranked set sample (RSS), which is a modification of the original RSS of McIntyre (1952).We have proposed several methods of estimation of variance by combining...

  • A Test for Symmetry with Leptokurtic Financial Data. Premaratne, Gamini; Bera, Anil // Journal of Financial Econometrics;Spring2005, Vol. 3 Issue 2, p169 

    Most of the tests for symmetry are developed under the (implicit or explicit) null hypothesis of normal distribution. As is well known, many financial data exhibit fat tails, and therefore commonly used tests for symmetry (such as the standard √b1 test based on sample skewness) are not...

  • A STUDY OF THE HARTMAN-WATSON DISTRIBUTION MOTIVATED BY NUMERICAL PROBLEMS RELATED TO THE PRICING OF ASIAN OPTIONS. Barrieu, P.; Rouault, A.; Yor, M. // Journal of Applied Probability;Dec2004, Vol. 41 Issue 4, p1049 

    One approach to the computation of the price of an Asian option involves the Hartman- Watson distribution. However, numerical problems for its density occur for small values. This motivates the asymptotic study of its distribution function.

  • Statistical Significance Tests: Equivalence and Reverse Tests Should Reduce Misinterpretation. Parkhurst, David F. // BioScience;Dec2001, Vol. 51 Issue 12, p1051 

    Focuses on situations in which significance tests are used and a null hypothesis is not rejected, and the problem of interpreting the result as providing evidence for the null hypothesis without a logical basis. Description of a method for looking at data statistically but with different null...

  • Statistical tests for limit value transgression–how to deliver unambiguous results. Molt, Karl; Einax, Jürgen W.; Winterstein, Michael // Analytical & Bioanalytical Chemistry;Nov2009, Vol. 395 Issue 6, p1577 

    The article discusses the use of statistical tests for limit value transgression. Statistical tests can be applied to unobjectionably and reproducibly justify a decision on whether a limit value has been violated or seen, and on the other hand to calculate the analytical effort needed to get...

  • contingency table.  // Hutchinson Dictionary of Scientific Biography;2005, p1 

    Table listing information classified as variable according to two or more independent attributes. Such tables are used commonly in commerce, notably by insurance companies.

  • Model selection for Gaussian concentration graphs. Drton, Mathias; Perlman, Michael D. // Biometrika;Sep2004, Vol. 91 Issue 3, p591 

    A multivariate Gaussian graphical Markov model for an undirected graph G, also called a covariance selection model or concentration graph model, is defined in terms of the Markov properties, i.e. conditional independences associated with G, which in turn are equivalent to specified zeros among...

  • The robust binomial approach to chance-constrained optimization problems with application to stochastic partitioning of large process networks. Stan, Oana; Sirdey, Renaud; Carlier, Jacques; Nace, Dritan // Journal of Heuristics;Jun2014, Vol. 20 Issue 3, p261 

    In this paper, we study an interpretation of the sample-based approach to chance-constrained programming problems grounded in statistical testing theory. On top of being simple and pragmatic, this approach is theoretically well founded, non parametric and leads to a general method for leveraging...


Read the Article


Sorry, but this item is not currently available from your library.

Try another library?
Sign out of this library

Other Topics