MINITAB: AN OVERVIEW - IASRI

1 minitab : AN OVERVIEW Rajender Parsad , Library Avenue, New Delhi 110 012 The functionality of minitab is accessible through interactive windows and menus, or through a command language called session commands. There are three windows viz. Data window, Session window and Project Manager. Data window is a worksheet in a spreadsheet format, with rows and columns that intersect to form individual cells. A worksheet can contain up to 4000 columns, 1000 constants, and up to 10,000,000 rows depending on memory of the computer. The text output generated by the analyses is displayed in Session window.

2 The Project Manager contains folders that allow one to navigate, view, and manipulate various parts of the project. minitab has the advanced Design of Experiments (DOE) capabilities. One can screen the factors to determine which are important for explaining process variation. It can generate two-level full and fractional factorial designs, and Plackett-Burman designs, Box-Behnken and central composite designs, simplex centroid and simplex lattice designs and Taguchi orthogonal array designs. It also allows one to perform one way analysis of variance, two-way analysis of variance for balanced data, test for equality of variances, and generate various plots.

3 Balanced ANOVA models with crossed or nested and fixed or random factors can also be analyzed. The option General MANOVA analyzes balanced or unbalanced MANOVA models with crossed or nested and fixed or random factors. The analysis of covariance is also possible with option General MANOVA. For initiatinfg the work on minitab . From thw Windows Taskbar, choose Start Programs minitab 14 ( minitab SOLUTIONS) minitab 14 ( minitab 15). minitab opens with two main windows viz. Session Window and Data Window. The first screen of minitab are shown as minitab : An OVERVIEW I-180 Under the Data Menu.

4 The following options are available Subset Worksheet - copies specified rows from the active worksheet to the new worksheet Split Worksheet - splits or unstacks the active worksheet into two or more new worksheets based on one or more "By" variables Merge Worksheets - combines two worksheets into one new worksheet Sort - sorts one or more columns of data Rank - assigns rank scores to values in a column Delete Rows - deletes specified rows from columns in the worksheet Erase Variables - erases any combination of columns, stored constants and matrices Copy - copies selections from one position in the worksheet to another; can copy entire selections or a subset Stack - stacks columns on top of each other to make longer columns Unstack - unstacks (or splits) columns into shorter columns Transpose Columns - switches columns to rows Concatenate - combines two or more text columns side by side into one new column Code - recode values in columns Change Data Type - changes columns from one data type (such as numeric, text, or date/time)

5 To another Display Data - displays data from the current worksheet in the Session window Extract from Date/Time to Numeric/Text - extracts one or more parts of a date/time column, such as the year, the quarter, or the hour, and saves that data in a numeric or a text column. In the worksheet, one can enter the data in columns numbered as C1, C2, .. The names of the variables can be written in the row below the row cotaining column numbers C1, C2, .. Calc Menu has the following sub-options Calculator - does arithmetic using an algebraic expression, which may contain arithmetic operations, comparison operations, logical operations, and functions Column Statistics - calculates various statistics based on a column you select Row Statistics - calculates various statistics for each row of the columns you select Standardize - centers and scales columns of data Make Patterned Data - provides an easy way to fill a column with numbers or date/time values that follow a pattern.

6 See also Generating Patterned Data OVERVIEW for related information. Make Mesh Data - creates a regular (x,y) mesh to use for drawing contour, 3D surface and wireframe plots, with the option to create the z-variable as well minitab : An OVERVIEW I-181 Make Indicator Variables - creates indicator (dummy) variables that you can use in regression analysis. See also Generating Patterned Data OVERVIEW for related information. Set Base - fixes a starting point for minitab 's random number generator Random Data - displays commands for generating a random sample of numbers, sampled either from columns of the worksheet or from a variety of distributions Probability Distributions - displays commands that allow you to compute probabilities, probability densities, cumulative probabilities, and inverse cumulative probabilities for continuous and discrete distributions Matrices - displays commands for doing matrix operations The main menu for statistical data analysis Stat.

7 Under this option, following suboptions are available: Basic Statistics Regression ANOVA (Analysis of Variance) DOE (Design of Experiments) Control Charts Quality Tools Reliability/Survival Multivariate Time Series Tables Nonparametrics EDA (Exploratory Data Analysis) Power and Sample Size In Basic statistics, following sub-options can be used through selecting Stat > Basic Statistics Select one of the following commands: Display Descriptive Statistics , Store Descriptive Statistics , Graphical Summary, 1-Sample Z, 1-Sample t, 2-Sample t, Paired t, 1 Proportion, 2 Proportions, 1-Sample Poisson Rate, 2-Sample Poisson Rate, 1 Variance, 2 Variances, Correlation, Covariance, Normality Test, Goodness-of-Fit Test for Poisson.

8 Then further sub-sub options can be used. For performing regression analysis, from the menus choose Stat > Regression and then select one of the following commands to fit a model relating a response to one or more predictors : Regression - does simple, multiple and polynomial regression Stepwise - does stepwise regression, forward selection, and backward elimination Best Subsets - does best subsets regression Fitted Line Plot - fits a simple linear or polynomial regression model and plots the regression line through the actual data or the log10 of the data Partial Least Squares - does partial least squares regression Binary Logistic Regression - does logistic regression for a binary response variable Ordinal Logistic Regression - does logistic regression for an ordinal response variable Nominal Logistic Regression - does logistic regression for a nominal response variable minitab : An OVERVIEW I-182 For performing Analysis of variance, Choose: Stat > ANOVA.

9 This option allows to perform analysis of variance, test for equality of variances, and generate various plots. The analysis can be carried out, using the suitable sub-option. One-Way - performs a one-way analysis of variance, with the response in one column, subscripts in another and performs multiple comparisons of means One-Way (Unstacked) - performs a one-way analysis of variance, with each group in a separate column Two-way - performs a two-way analysis of variance for balanced data Analysis of Means - displays an Analysis of Means chart for normal, binomial, or Poisson data Balanced ANOVA - analyzes balanced ANOVA models with crossed or nested and fixed or random factors General Linear Model - analyzes balanced or unbalanced ANOVA models with crossed or nested and fixed or random factors.

10 You can include covariates and perform multiple comparisons of means. Fully Nested ANOVA - analyzes fully nested ANOVA models and estimates variance components Balanced MANOVA - analyzes balanced MANOVA models with crossed or nested and fixed or random factors General MANOVA - analyzes balanced or unbalanced MANOVA models with crossed or nested and fixed or random factors. You can also include covariates. Test for Equal Variances - performs Bartlett's and Levene's tests for equality of variances Interval Plot - produces graphs that show the variation of group means by plotting standard error bars or confidence intervals Main Effects Plot - generates a plot of response main effects Interactions Plot - generates an interaction plots (or matrix of plots) minitab can also be used for generating the layout of designs for two-level full and fractional factorial designs using Stat > DOE > Factorial.