eng TQMP Tutorials in Quantitative Methods for Psychology 1913-4126 2009-09-01 5 2 40 50 article Fernando Marmolejo-Ramos fernando.marmolejoramos@ adelaide.edu.au 1 Masaki Matsunaga matsunaga@aoni.waseda.jp 2 The University of Adelaide Waseda University Most psychological research employs tables to report descriptive and inferential statistics. Unfortunately, those tables often misrepresent critical information on the shape and variability of the data’s distribution. In addition, certain information such as the modality and score probability density is hard to report succinctly in tables and, indeed, not reported typically in published research. This paper discusses the importance of using graphical techniques not only to explore data but also to report it effectively. In so doing, the role of exploratory data analysis in detecting Type I and Type II errors is considered. A small data set resembling a Type II error is simulated to demonstrate this procedure, using a conventional parametric test. A potential analysis routine to explore data is also presented. The paper proposes that essential summary statistics and information about the shape and variability of data should be reported via graphical techniques. http://www.tqmp.org/Content/vol05-2/p040/p040.pdf Statistics Exploratory data analysis Box plot Confidence intervals Violin plot fre TQMP Tutorials in Quantitative Methods for Psychology 1913-4126 2009-09-01 5 2 51 58 article Louis Laurencelle louis.laurencelle@uqtr.ca 1 Université du Québec à Trois-Rivières Dans le langage de tous les jours, l’expression « corrélation entre deux variables » est entendue et bien comprise de manière générale : c’est un lien, un rapport de correspondance grâce auquel la variation d’un attribut peut être associée à la variation d’un autre attribut... http://www.tqmp.org/Content/vol05-2/p051/p051.pdf Statistics Correlation Ordinal variables Categorical variables eng TQMP Tutorials in Quantitative Methods for Psychology 1913-4126 2009-09-01 5 2 59 67 article Denis Cousineau denis.cousineau@umontreal.ca 1 Université de Montréal When programming complex experiments (for example, involving the generation of stimuli online), the traditional experiment programming software are not well equipped. One solution is to give up entirely the use of such software in favor of a low-level programming language. Here we show how E-Prime can be connected to Mathematica so that the easiness and reliability of this software can be preserved while at the same time granting it the full computational power of a high-level programming language. As an example, we show how to generate noisy images with noise proportional to the rate of success of the participants with as few as 12 lines of codes in E-Prime. http://www.tqmp.org/Content/vol05-2/p059/p059.pdf Experiment programming eng TQMP Tutorials in Quantitative Methods for Psychology 1913-4126 2009-09-01 5 2 68 76 article Laurence Richard laurencerichard2@gmail.com 1 Dominic Charbonneau dominic.charbonneau@umontreal.ca 2 Miami University Université de Montréal When running an experiment, precision is essential to ensure results are as exact as possible. Thus, computers, which offer endless accuracy, have become an inevitable tool to design experiments. To avoid programming from scratch for each new situation, a program, E-Prime, has been created to ease the conception of experiments. E-Prime, developed by PSTNet, offers a user-friendly interface that makes typical experiments easy to create. This paper shows how to effortlessly create an experiment with E-Prime, followed by a simple example. http://www.tqmp.org/Content/vol05-2/p068/p068.pdf Experiment programming