Normality Test and Variance Homogeneity Presentation of Tables and Pictures

Proceedings of MatricesFor IITTEP – ICoMaNSEd 2015 ISBN: 978-602-74204-0-3 Keynote Invited Papers Page 10

2.3. Population and Sample

Regarding to the procedure of data collecting, there are two kinds of research, which are experimental research and non-experimentalobservational research. Population and sample are acknowledged in non-experimental study whereas there is no such thing as real population in experimental research. The data will be generated through an experiment, different with the non-experimental design, the data already provided in observational unit. For this reason the population in experimental study is not mentioned. It is a mistake to mention the population in experimental study, instead the term used should be observational unit or research subject. In many researches they do not put the reason to get the sample and not so rarely the relatively small population is still taken as a sample. Ideally a research should be done on population not on sample. If a research using a population then there is no need statistical tests t-test statistics, analysis of variance, regressioncorrelation analysis that match with the hypothesis formulation. In sampling, many of studies in determining the sample size using the formulation that not fit with the problem formulation or the goal of the research. If the problem formulation or the goal of the research is comparing two group then the formula use for determining the sample size is Dawson and Trapp, 2001; Mulyatiningsih, 2012:   2 2 1 2             s Z Z n a And if the goal of the research is to analyze the relationship between two variables, then the formula for the sample size is Dawson and Trapp, 2001:   2 3 1 0, 5ln 1 a Z Z n r r                         Many articles using the formula of sample size as Sugiono, 2007; Sukardi, 2010; Mulyatiningsih, 2012: � = � 2 �� 1 − � � 2 � − 1 + � 2 � 1 − � The latter formula is use if the goal of the research is to test the proportion of one population.

2.4. Normality Test and Variance Homogeneity

In educational research, the normality test aimed to test the condition in Pearson’s correlational analysis or comparison of means test. Many of the researchers do not understand why the normality test should be applied in the formulation of correlational research or comparison of means test. In the case if this condition unfulfilled they change the data instead of transform the data or using the Spearman correlation or nonparametric test. In research it is acceptable to mention “the data assumed to be normally distributed” in the datastatistical analysis section, so that the data normal test is unnecessary. The test of variance homogeneity is frequently misused, as in the case of the researcher cannot determine the difference which one is the homogeneity of variance test in the problem of comparison of two or more populations and which one is the regression analysis. The homogeneity of variance can be assumed, so the test is unnecessary. Proceedings of MatricesFor IITTEP – ICoMaNSEd 2015 ISBN: 978-602-74204-0-3 Keynote Invited Papers Page 11

2.5. Presentation of Tables and Pictures

The purpose of presentations of tables and pictures in a scientific article is for simplification and informative. But, in many articles the presentation of tables or pictures do not simplified the information, in the contrary they adding confusions. That because the title of is too long or contain irrelevant information. In some cases, some of the articles shared the same tables and pictures in the different sections for the same repeated information. In the worst case the information that already represented by tables or pictures all re-presented in the form of narrations. In one of the re reviewed article is found that the result of data analysis presented in tables just use the copy and paste of the same contents which showed that the unnecessary columns and rows in the table.

2.6. Bar Chart and Histogram Presentation