FREE ELECTRONIC LIBRARY - Theses, dissertations, documentation

Pages:   || 2 |

«Marmolejo-Ramos, Fernando Current exploratory and confirmatory issues in data analysis in Psychology. International Journal of Psychological ...»

-- [ Page 1 ] --

International Journal of Psychological


ISSN: 2011-2084


Universidad de San Buenaventura


Marmolejo-Ramos, Fernando

Current exploratory and confirmatory issues in data analysis in Psychology.

International Journal of Psychological Research, vol. 3, núm. 1, 2010, pp. 4-8

Universidad de San Buenaventura

Medellín, Colombia

Available in: http://www.redalyc.org/articulo.oa?id=299023509001

How to cite

Complete issue

Scientific Information System More information about this article Network of Scientific Journals from Latin America, the Caribbean, Spain and Portugal Journal's homepage in redalyc.org Non-profit academic project, developed under the open access initiative International Journal of Psychological Research, 2009. Vol. 3 No. 1. Marmolejo-Ramos, F. (2010). Current exploratory and confirmatory issues in ISSN impresa (printed) 2011-2084 data analysis in Psychology. International Journal of Psychological Research, 3(1), 4-8.

ISSN electrónica (online) 2011-2079


Current exploratory and confirmatory issues in data analysis in Psychology.

Fernando Marmolejo-Ramos University of Adelaide - Universidad del Valle In 1968 John Tukey gave a speech at the American Psychological Association in San Francisco about the relevance of proper data analysis in Psychology (Tukey, 1969). His closing message was that “data analysis needs to be both exploratory and confirmatory” (p. 90). Exploratory data analysis (or EDA) is an approach to analysing data in order to formulate sound hypotheses, whereas confirmatory data analysis (CDA) is a method to test those hypotheses (a.k.a., statistical hypothesis testing). As Tukey announced in his speech, these two analytical tools have been, and are somewhat still, at odds. This special issue presents sixteen papers that cover relevant topics in EDA and CDA with the purpose of bringing together seemingly disparate issues.

The first section comprises eight papers that present a fresh look into basic and essential topics in descriptive statistics, data visualization, and data exploration. Thus, in the first paper, Yu discusses the role of EDA when performing data mining and resampling. His paper offers an updated view of what EDA stands for nowadays and how it can be useful in statistical modelling.

In a second paper, Teknomo and Estuar stretch the analytical power of EDA by using it as a tool to finding patterns in data. The authors show how some clustering methods can be used as EDA in order to identify patterns in pedestrian behaviour. Of particular interest is that the authors consider the role of a novel method called Affinity Propagation in relation to EDA.

EDA is characterised by relying on graphical techniques to explore and report data, e.g., via barcharts, scatterplots, or boxplots. In a third paper, Marmolejo-Ramos and Tian present a variation of the boxplot, a classic graphical technique.

Traditionally, the boxplot represents summary statistics around the median; nevertheless most statistical tests use the mean.

Thus, this paper proposes a variation of the boxplot which displays information around the mean.

An advantage of the boxplot is that it not only displays the mean value obtained but also the reliability of such an estimate, i.e., its confidence interval (CI). In a fourth paper, Beaulieu-Prévost discusses what CIs mean, how they are computed, and how they relate to hypothesis testing. Several authors have different views on these questions and this paper aims to reconcile them.

Parametric statistical tests assume that data are distributed following a bell-shaped curve. However, in actual research it is rare to find such well-behaved data. Typically, data distribution is affected because of some observations that sit far away from the mean, i.e., they are outliers. In a fifth paper, Cousineau and Chartier consider the impact of outliers in data distribution and propose various ways to deal with them. This paper considers the influence of outliers not only in univariate data sets but also in multivariate ones.

One of the approaches to dealing with outliers is by re-expressing the data so that they are fit for further analysis, i.e., data are transformed. There are various types of transformations, for instance, inverse, logarithmic, and square-root. In the sixth paper, Olivier and orberg focus on a particular type of transformation called Box-Cox. The authors argue that this type of transformation is suitable when dealing with positively skewed data, like those of reaction times.

A rather recent alternative approach to classical statistical methods investigates methods and techniques to find better estimators of central tendency. It is well known that outliers affect the distribution of data and therefore bias their estimators of location and spread, i.e., the mean and the standard deviation respectively. Therefore, the use of outlier elimination is commonly employed. In a seventh paper Courvoisier and Renaud argue that outlier elimination can have some shortcomings too. The authors emphasize that not all outliers are detected and therefore results of classical statistical tests can be biased. The paper presents a robust alternative to the Analysis of Variance.

The first section of this special issue concludes with a paper by Fischer and Milfont in which the use of the term standardization is clarified. The authors of this eighth paper particularly focus on the standardization of scores and what tests are available to analyse standardized data. The paper contains applications of standardization procedures in personality, organizational, and cross-cultural psychology.

The second section of this special issue presents various methods to test hypotheses. Factor Analysis (FA) is a method to describe variability among a set of variables in relation to some unobserved variables (or factors). Thus, in the ninth paper Matsunaga presents a primer on a special form of FA called Confirmatory Factor Analysis (CFA). The paper

–  –  –

differentiates between what Exploratory Factor Analysis (EFA) and CFA are. The paper provides worked-out examples as to how to perform both EFA and CFA in SPSS and LISREL.

Because in FA several variables are being compared to each other, it is important that the instrument used measures the same psychological construct across variables. In a tenth paper Milfont and Fischer review the relevance of equivalence in Psychological research. The authors also consider theoretical and methodological issues regarding measurement invariance in CFA. Examples are provided using LISREL.

A common technique in data analysis is called clustering or cluster analysis. In this technique sets of observations are placed in groups (clusters) based on some similarities. In an eleventh paper, Dutta Roy shows the importance of a type of clustering called hierarchical clustering. The author uses this method in determining test-retest reliability of questionnaires. The paper presents data from a reading motivation study to illustrate the power of cluster analysis over classical paired t-tests.

A popular statistical method to fit statistical models to data and provide estimates for the model is known as Maximum Likelihood Estimation (MLE). In the twelfth paper, Dunn provides an introductory tutorial on how to fit models of recognition memory data using MLE. In his paper, data from a recognition memory study are used to illustrate how different models can be fit to the data and graphically represented using Receiver Operating Characteristic curves (ROC curves). The author offers a step-by-step example using EXCEL.

Traditional literature reviews are characterised by presenting in a narrative way results from past studies. Another approach called meta-analysis (MA) forces a more careful literature review in which the results from several empirical studies are combined in order to produce an overall effect size of the construct of interest. In the thirteenth paper, SánchezMeca and Marín-Martínez present the importance of implementing meta-analysis in Psychological research. Their paper presents a tutorial on how to carry out MA and report results using real data.

Classical statistics, a.k.a. frequentist, has been criticised for offering only dichotomous answers when hypotheses are tested: they are rejected or not rejected. Another type of statistics, called Bayesian, argues that it is possible to assign a probability to any hypothesis being tested. In the fourteenth paper, González presents what Bayesian statistics is and how it can be used in psychometrics. The author presents how Bayesian models can be used to analyse data from questionnaires, tests, or similar instruments that measure different variables.

Finally, the fifteenth contribution is a full-length paper on other applications of Bayesian statistics to Psychological research. In this paper Hsieh and von Eye propose a model in which a measurement model and a structural model are combined. The authors use data taken from British Social Attitudes Panel Survey 1983-1986. Their results throw light on how to resolve some difficult modelling issues commonly encountered in developmental research.

All the papers in this special issue are characterised by centring on the application of techniques rather than on their mathematical technicalities. For that reason, not only Psychologists but social researchers in general can benefit from the topics addressed here. Various important topics, e.g., homogeneity of variance and normality, visualization of statistical ideas and models, relative distribution methods, bootstrap and jackknife, and logistic regression models just to mention a few, were left out given space constraints. Nevertheless, we believe this special issue offers a series of papers that can broaden researchers’ knowledge in statistics and motivate further research.


Tukey, J. W. (1969). Analysing data: Sanctification or detective work? American Psychologist, 24 (2), 83-91.

Dirección correspondencia/Mail Address:

Fernando Marmolejo-Ramos, Fernando Marmolejo-Ramos, School of Psychology, Faculty of Health Sciences, University of Adelaide, Australia and Institute of Psychology, Universidad del Valle, Colombia. Email: fernando.marmolejoramos@adelaide.edu.au, web page: www.firthunands.com


OPEN JOURNAL SYSTEM, BIBLIOTECA VIRTUAL DE PSICOLOGIA (ULAPSY-BIREME), DIALNET y GOOGLE SCHOLARS. Algunos de sus articulos aparecen en SOCIAL SCIENCE RESEARCH NETWORK y está en proceso de inclusion en diversas fuentes y bases de datos internacionales.


JOURNAL SYSTEM, BIBLIOTECA VIRTUAL DE PSICOLOGIA (ULAPSY-BIREME ), DIALNET and GOOGLE SCHOLARS. Some of its articles are in SOCIAL SCIENCE RESEARCH NETWORK, and it is in the process of inclusion in a variety of sources and international databases.

–  –  –

En 1968 John Tukey pronunció un discurso en la Asociación Americana de Psicología, en San Francisco, acerca de la importancia de llevar a cabo un adecuado análisis de datos en psicología (Tukey, 1969). El mensaje final de su discurso era que “el análisis de datos debe ser, tanto exploratorio como confirmatorio” (p. 90). El análisis exploratorio de datos (EDA, por sus siglas en ingles) es un método para analizar datos cuyo fin es la formulación de hipótesis sólidas, mientras que el análisis confirmatorio de datos (CDA, por sus siglas en inglés) es un método para poner a prueba dichas hipótesis (también conocido como prueba de hipótesis). Como Tukey lo anunció en su discurso, el uso de estas dos herramientas analíticas ha sido, y de algún modo, sigue siendo dispar. En esta edición especial presentamos dieciséis artículos que abarcan temas relevantes concernientes al EDA y al CDA, con el propósito de reunir asuntos supuestamente disímiles.

La primera sección está compuesta por ocho artículos que presentan una mirada fresca sobre los temas básicos y esenciales concernientes a la estadística descriptiva, la visualización de datos y la exploración de datos. Así, por ejemplo, en el primer trabajo, Yu discute el papel del EDA en la minería de datos y el método de remuestreo. Su trabajo ofrece una visión actualizada de lo que significa el EDA hoy en día y de cómo puede ser de utilidad en el modelamiento estadístico.

En un segundo trabajo, Teknomo y Estuar extienden el poder analítico del EDA al emplearlo como una herramienta para hallar patrones en los datos. Los autores muestran cómo algunos métodos de conglomerados (clustering) se pueden emplear como EDA, a fin de identificar patrones en el comportamiento de peatones. Es de particular interés el hecho de que los autores abordan el papel de un nuevo método denominado propagación de afinidad con relación al EDA.

El EDA se caracteriza por el empleo de técnicas gráficas para explorar y reportar los datos, por ejemplo, gráficas de barras, diagramas de puntos o diagramas de cajas. En un tercer artículo, Marmolejo-Ramos y Tian presentan una variación de una técnica gráfica clásica, el diagrama de cajas. Tradicionalmente, el diagrama de cajas representa el resumen estadístico según la mediana; sin embargo, la mayoría de pruebas estadísticas usan la media. Así, este artículo propone una variación del diagrama de cajas que muestra la información según la media.

Una ventaja del diagrama de cajas es que no solamente muestra el valor medio obtenido, sino también la fiabilidad de dicho cálculo, es decir, su intervalo de confianza (CI, por sus siglas en inglés). En un cuarto artículo, Beaulieu-Prévost discuten el significado de los intervalos de confianza, la forma de computarlos y la manera en que se relacionan con las pruebas de hipótesis. Varios autores tienen distintas opiniones sobre estos asuntos y este artículo tiene como objetivo reconciliarlas.

Pages:   || 2 |

Similar works:

«Essential TRUTHS OF THE CHRISTIAN FAITH R.C. SPROUL TYNDALE HOUSE PUBLISHERS, INC. WHEATON ILLINOIS Copyright © 1992 by R. C. Sproul PREFACE Orlando July, 1992 Every Christian is a theologian. We are always engaged in the activity of learning about the things of God. We are not all theologians in the professional or academic sense, but theologians we are, for better or for worse. The “for worse” is no small matter. Second Peter warns that heresies are destructive to the people of God and...»

«La Lucha Continues Mujerista Theology Ada María Isasi-Díaz Founded in 1970, Orbis Books endeavors to publish works that enlighten the mind, nourish the spirit, and challenge the conscience. The publishing arm of the Maryknoll Fathers and Brothers, Orbis seeks to explore the global dimensions of the Christian faith and mission, to invite dialogue with diverse cultures and religious traditions, and to serve the cause of reconciliation and peace. The books published reflect the views of their...»

«Oracle Utilities Customer Care and Billing Release 2.4.0 Utility Reference Model Select Customers and Manage Programs December 2015 Oracle Utilities Customer Care and Billing Utility Reference Model, Release 2.4.0 Copyright © 2015, Oracle and/or its affiliates. All rights reserved. This software and related documentation are provided under a license agreement containing restrictions on use and disclosure and are protected by intellectual property laws. Except as expressly...»

«―TO THE STARS THROUGH DIFFICULTIES‖: THE LEGAL CONSTRUCTION OF PRIVATE SPACE AND THE WIZARD OF OZ DONNA E. YOUNG* I. INTRODUCTION Kansas. Thirty-fourth State of the Union.1 Located in the center of the forty-eight contiguous states.2 Home of Walter Chrysler, Amelia Earhart, Dwight D. Eisenhower, Melissa Etheridge, and Langston Hughes.3 One of three states claiming to be the home of the largest ball of twine,4 a mostly Christian state,5 site of the school desegregation dispute leading to...»

«Exploring the Cold War through The Twilight Zone 39 Exploring the Cold War through The Twilight Zone: Five episodes in a journey to a dimension of sight, sound and mind Heather Lunney Masters, University of New England The Cold War between the United States and the Soviet Union was largely characterised by the threat of military conflict rather than its actuality.1 The ideological nature of the Cold War led both sides to recognise the critical importance of not only building up their military...»

«We.sternJournal ofCommunicalion, 60(2) (Spring 1996), 168-187 The Priestly Rhetoric of Neoconservatism KENNETH S. ZAGACKI The torrid conflicts of the 19608 produced a profound change in American politics. Nowhere was this change more evident than in the clash between radicals and liberals of the New Left, and the emerging neoconservative ll defenders of American culture and politics. Once members of the Old Left, neoconservatives adopted a priestly voice to attack contemporary liberalism and...»

«Aplicación de las licencias de Creative Commons en el ámbito educativo Ignasi Labastida i Juan Resumen: En este texto se presentan las licencias que ofrece la organización Creative Commons para gestionar los derechos de propiedad intelectual de una manera distinta al tradicional “todos los derechos reservados”. Estas licencias inspiradas en el movimiento del software libre se ofrecen gratuitamente para todos aquellos autores que quieran compartir sus obras permitiendo determinados usos...»

«Journal of Research in Reading, ISSN 0141-0423 Volume 23, Issue 2, 2000, pp 181±193 Computer-assisted learning to read and spell: results from two pilot studies Victor H.P. van Daal School of Psychology, University of Wales, Bangor, UK Pieter Reitsma Paedologisch Instituut, The Free University, Amsterdam ABSTRACT Because multimedia computer programs may provide promising opportunities for the training of initial reading and spelling skills, two small-scale pilot studies have been conducted...»

«Eλλην The Illuminati Formula Used to Create an Undetectable Total Mind Controlled Slave By Fritz Springmeier and Cisco Wheeler Ηττp://ellhn.e-e-e.gr TABLE OF CONTENTS Introduction The 12 Major Sciences of Monarch Mind Control I. The Selection & Preparation of the Victim A. Genetics & dissociative abilities B. Availability C. Physical & Mental requirements D. List of organizations carrying out programming E. 4 foundational steps for programming F. Step 1. Spiritual requirements, Moon...»

«Longarm And The Cottonwood Curse Under this the funds wo stick considered to be your willing receipts up any organizations/councils. Tell, them do to reconnect their firm home and however make you. They is financial sector research to pay a brother from the service, and paying as our time will enough have their genre people for you are his course and in a mind if you among the. The from the prices does such Longarm and the Cottonwood Curse to a genre structure separately inflated as of Loans...»

«The Sport Psychologist, 1997,11,257-276 O 1997 Human Kinetics Publishers, Inc.Burnout in Competitive Junior Tennis Players: 111. Individual Differences in the Burnout Experience Daniel Gould Suzanne Tuffey University of North Carolina U.S. Olympic Committee at Greensboro JamesLoehr Eileen Udry Loehr-Groppel Sport Science, Orlando, FL University of Oregon This is the third in a series of manuscripts reporting results from a research project designed to examine burnout in competitive junior...»

«Team Hand Book Thomassin & Grand Goave, Haiti Team Member, We are excited that you have chosen to join our team to Haiti! We look forward to sharing in this journey with you. We pray that along the way your heart will be awakened, your mind will be renewed, and that your life will be forever changed through your week in Haiti. The vision of a door to hope is to be the living love of Jesus in our own lives as well as in the country of Haiti. We yearn to be the Gospel by sharing Truth and being...»

<<  HOME   |    CONTACTS
2016 www.theses.xlibx.info - Theses, dissertations, documentation

Materials of this site are available for review, all rights belong to their respective owners.
If you do not agree with the fact that your material is placed on this site, please, email us, we will within 1-2 business days delete him.