Minitab Demystified

Introduction

If you need to use Minitab for school or work, and you’re worried about it, this book is for you. In recent years, statistics has become a required skill for students and professionals in all fields. For many of those people, math and statistics are not their favorite activities.

The good news is that Minitab handles the hard math of statistical analysis quickly and easily. The better news is that new tools in Minitab 16 help explain and interpret those results with color-coded graphs, sentences in plain language, and neatly formatted reports. The best news is that this book will guide you gently and quickly to find the best features of Minitab and apply them to solve your problems.

This book is organized informally into four parts:

• Chapters 1-3 are for newcomers to Minitab who want to quickly come up to speed. Chapter 1 introduces common Minitab features using one extended example. Chapter 2 explores many types of Minitab graphs. Chapter 3 provides a thorough tour of the Minitab interface.

• Chapters 4-7 describe tools to make better decisions from data. These chapters cover distribution models, intervals for making decisions, hypothesis testing, and sample size calculations.

• Chapters 8-9 are for modelers and experimenters. Chapter 8 discusses regression tools. Chapter 9 covers the powerful design of experiments (DOE) toolbox in Minitab, including factorial, response surface, and mixture designs, plus an important new feature of Minitab 16 that handles split-plot designs.

• Chapters 10-14 introduce the most important tools used in process control, quality improvement, and Six Sigma initiatives. These tools include measurement systems analysis, control charts, capability analysis, acceptance sampling, and reliability analysis.

In release 16, Minitab introduced an impressive array of new and improved features. The revolutionary Minitab assistant offers sound advice about collecting and organizing data, selecting the right analysis tool, and interpreting analysis results with plain language and consistently formatted reports. Throughout this book, new features of Minitab 16 are flagged with the distinctive logo shown here.

How to Get the Most from This Book

Feel free to jump around in this book, as you need to. Each chapter is written without assuming a thorough familiarity with preceding chapters. The first example of each chapter is explained down to the keystroke level. When you need to solve a particular type of problem, jump to that chapter and start reading. Here are some other tips for readers:

• Work the examples. Almost all the datasets used in this book are included with Minitab in the Sample data folder. Follow along with the examples, and explore the results you see on your computer. You will learn much more by working on your computer than by merely reading a book.

• Refer to other books. This book explains Minitab, with explanations of statistical terms and concepts required to understand Minitab reports. Other books have more space to define and illustrate statistical tools more thoroughly. Most chapters in this book list good references in the Find Out More section.

• Do the quizzes. Each chapter ends with a 10-question quiz. The quiz reinforces concepts in the chapter and sometimes introduces new variations on Minitab tools. For most questions, the answer is explained at the end of the book.

• Work your own problems. After reading the examples here, try them on your own data, and use Minitab to solve your problems.

• Take the final exam. After some time passes, come back and take the final exam. The exam has several questions relating to each chapter, more or less in order. These questions provide more opportunities to practice your statistical and Minitab skills, using Minitab example data files. Answers to most questions are explained.

Not every Minitab tool and function is covered here. Most people will find their first, second, and hopefully third set of Minitab tools explained in this book. By that time, the Minitab help files will hopefully suffice for learning more Minitab skills.

Trademarks

MINITAB® and all other trademarks and logos for the Company’s products and services are the exclusive property of Minitab Inc. All other marks referenced remain the property of their respective owners. See www.minitab.com for more information on Minitab products.

Microsoft®, Windows®, Word®, Excel®, and PowerPoint® are registered trademarks of the Microsoft group of companies.

chapter 1

Getting Started with Minitab Statistical Software

In this chapter, we will explore a dataset by using Minitab® statistical software to create graphs, statistical summaries, and a report. If you are new to Minitab, working through the sections of this chapter will guide you through a variety of commonly used functions, including some of the exciting, new features of Minitab 16.

CHAPTER OBJECTIVES

Here’s what you’ll learn in this chapter:

• How to create, open, and save Minitab projects and worksheets

• How to create histograms

• How to combine multiple graphs in panels of a single graph

• How to calculate descriptive statistics

• How to reveal relationships between variables with scatterplots and matrix plots

• How to use brushing to select and analyze certain data values from a graph

• How to organize a Minitab project and navigate through it using the Project Manager

• How to quickly create a printed report from a Minitab analysis

1.1 Opening a New Minitab Project

To start Minitab statistical software, either double-click the Minitab icon on the Windows® desktop, or select Minitab > Minitab 16 Statistical Software from the Start menu. Minitab opens a window displaying a new, empty Minitab project, as shown in Fig. 1-1.

FIGURE 1-1 • New, empty Minitab project.

In the Minitab window, below the menus and toolbars, are three parts of this new Minitab project:

• The Session window displays messages and reports from all Minitab commands as you execute them.

• The data window named Worksheet 1 *** can hold and display data of many types, arranged into columns. Data values may be entered directly into the data window, imported from other files, or copied and pasted through the Windows clipboard. Navigation around the data window is similar to navigation in a Microsoft® Office Excel® spreadsheet. A Minitab project may hold any number of worksheets, but this new project holds only one. The *** indicates that Worksheet 1 is the current worksheet. If many worksheets are open in the project, all Minitab commands operate only on data contained in the current worksheet.

• The Project Manager window is minimized at the bottom of the Minitab window. The Project Manager window and toolbar allow quick navigation between the components of a Minitab project.

Chapter 3 provides more details about the windows, toolbars, features, and capabilities of the Minitab interface.

At this point, if we wanted to continue working with a saved Minitab project, we could use the File > Open Project menu command. Since we have no project to continue, we will build up a new project from this empty template.

1.2 Loading Data into the Project

In the Minitab File menu, select Open Worksheet. This command is located below the group of commands for Minitab projects in the File menu. The Open Worksheet dialog appears, as shown in Fig. 1-2. Near the bottom of the dialog, click the icon, labeled Look in Minitab Sample Data Folder. Since most of the datasets used in this book are stored in the Minitab sample data folder, this is a convenient shortcut.

FIGURE 1-2 • Open Worksheet form.

Open the Student14 folder in the Sample Data folder, select the worksheet file named OldFaithful, and click Open. A box appears to advise that a copy of the data in the worksheet file will be added to the current project. Click OK in the warning, and the dataset appears in a new data window named OldFaithful.MTW ***. MTW is the file extension for Minitab worksheet files, and the *** is a reminder that this is now the current worksheet. The empty Worksheet 1 is still in the project, but it is no longer current.

Figure 1-3 shows the data window for the OldFaithful.MTW worksheet. The worksheet contains three columns with names Duration, Interval, and Height. Columns in Minitab worksheets always have identifiers C1, C2, and so on, but entering more descriptive names in the name row below C1, C2, …, is always a good idea.

FIGURE 1-3 • OldFaithful.MTW data window.

How many numbers are in each column of data? Scrolling through the data window is one way to see the size of a dataset, but Minitab provides an easier way to display information about all the columns in the worksheet. In the toolbars, look for the button, and click it. The Project Manager window appears on the left, tiled with the worksheet data window on the right. Figure 1-4 shows the Project Manager information for this worksheet, listing column names, ids, counts, the number of missing values, and data types.

FIGURE 1-4 • Project Manager information on OldFaithful.MTW.

If you cannot find the button, this could be because the Project Manager toolbar is hidden. To see a menu of all the available toolbars, move the cursor over any of the menus or toolbars, then right-click the mouse. When there is a check next to Project Manager in the toolbar menu, the Project Manager toolbar should be visible. This toolbar contains many other buttons for quickly navigating to different parts of the Minitab project. An alternative to clicking the button is to use the Ctrl+Alt+I keyboard shortcut.

In statistical analysis, it is important to identify missing data values. These are different from the empty cells in the data window below the end of the data in each column. Minitab identifies missing data values with the * symbol in the data window. In the OldFaithful.MTW data window, click on any cell and delete the value with the delete key, followed by enter. Now, the deleted value is replaced by *, and the Project Manager reports one missing value. Restore the missing value with Edit > Undo or use the keyboard shortcut Ctrl+Z.

At this point, it is a good idea to save the new Minitab project using File > Save Project As from the Minitab menu. Choose any convenient file name and location. The file will be saved with the MPJ extension reserved for Minitab projects. Note that the worksheet loaded from the OldFaithful.MTW file is only a copy of the data, and this copy will be saved as part of the new MPJ file. The original MTW file is unchanged unless the Save Current Worksheet command in the File menu is used. It is good practice to preserve any datasets in their original, unchanged form, in case the analysis needs to be repeated later.

Many users of Microsoft Office have come to expect programs to automatically save their work at regular intervals and to recover unsaved files in the event of a crash. Autosave and recovery are convenient features of Microsoft Office, but these features do not exist in Minitab. Also, like many other complicated Windows applications and Windows itself, Minitab does occasionally crash. For these reasons, savvy Minitab users save their work frequently. If they don’t, they will soon learn why they should.

1.3 Creating a Graph

Now it is time to explore this data by creating some graphs. The OldFaithful.MTW dataset contains measurements of duration in seconds, interval in minutes, and height in feet for 50 eruptions of the Old Faithful geyser in Yellowstone National Park. How are duration, interval, and height distributed?

In the Minitab menu, click Assistant > Graphical Analysis. The assistant appears as shown in Fig. 1-5. Starting in version 16, Minitab provides a variety of assistants to help people choose the most appropriate tool for their situation and then interpret the results. The graphical analysis assistant offers three categories of graphs: to graph the distribution of data, to graph variables over time, and to graph relationships between variables.

FIGURE 1-5 • Graphical analysis assistant.

Clicking on any of these categories shows a flow chart with decisions to help identify the best tool. Clicking on any diamond-shaped decision box in the flow chart displays examples and explanations for this decision. Next to each tool in the flow chart is a more… hyperlink. Clicking more… shows guidelines for collecting data and interpreting the results of any tool you choose.

For the OldFaithful.MTW worksheet, many of these graphs could be appropriate. For a first graph, select a histogram from the assistant. Figure 1-6 shows the histogram assistant dialog. This dialog is only used to create a simple histogram of one variable. In this dialog, click on the field labeled Y column. The column listbox is the large white space near the left edge of the dialog. This listbox lists all the columns in the current worksheet which may be selected. If this listbox is empty, be sure to click on the Y column field first. Double-click on C1 Duration in the column listbox, and Duration appears in the Y column field. Alternately, you may directly enter C1 or Duration into the Y column field. Click OK, and the histogram appears, as shown in Fig. 1-7.

FIGURE 1-6 • Histogram assistant.

FIGURE 1-7 • Histogram of eruption duration.

The histogram shows what is often called a bimodal distribution for duration. Some eruptions have a short duration, around two minutes (120 seconds), while others have a long duration, around four minutes. Between these two groups of eruptions is a gap. There are no eruptions in this dataset with a duration of around three minutes.

Statistics offers many tools to describe and summarize datasets with numerical values. But none of these numerical tools can adequately convey the important fact of this bimodal distribution as effectively as this one simple graph.

The OldFaithful.MTW dataset contains three variables. With a little more effort, we can create a single graph displaying the distribution of several variables. This time, select Graph > Histogram in the Minitab menus. Compared to the Assistant menu, the Histogram function in the Graph menu offers more options and flexibility.

The first Histograms dialog is a gallery of available histogram styles. A Simple histogram is already selected by default. Click OK to continue.

In the Histogram - Simple dialog, select all three columns. One way to do this is to click first on C1 Duration in the column listbox, then shift-double-click on C3 Height. This will enter Duration-Height into the Graph variables field of the dialog. Clicking OK now will produce three separate histograms. Try this.

Now, look for the button in the Standard toolbar and click it. This is a shortcut to recall the last dialog used, which is the Histogram - Simple dialog. Two other ways to access the last dialog used are to click Edit > Edit Last Dialog, or to use the keyboard shortcut Ctrl + E. For many people, this is their favorite Minitab function!

Back in the Histogram - Simple dialog, click the Multiple Graphs button. Under Show Graph Variables, select the option labeled In separate panels of the same graph. Click OK in the two open dialogs. Figure 1-8 shows the resulting histogram graph with three panels, one for each column of data.

FIGURE 1-8 • Histogram with panels.

The default arrangement of panels in a paneled graph may or may not be the most appropriate. To try a different arrangement of panels, right-click on the graph, and select Panels. In the Edit Panels dialog, select the Arrangement tab. Here, we can override the automatic arrangement and specify any number of rows and columns for the panels. Alternate arrangements for this example include one row with three columns and three rows with one column. Try alternate arrangements for the panels to find the one you like best.

Many graphs are possible, but which one is best? In general, the best graph makes it easy to see the most important features or stories of the dataset with a minimum of clutter and distractions. This comes down to a set of judgments made by the analyst. The first judgment is to decide what the big story in the data is, and next, to decide which graphs best reveal that story to the viewer.

As a reminder, now is a good time to save the Minitab project file by clicking the button in the Standard toolbar, or by using the Ctrl + S keyboard shortcut.

1.4 Calculating Descriptive Statistics

A typical statistical report on a dataset includes both graphs and descriptive statistics to summarize the data. To prepare a table of descriptive statistics, click Stat > Basic Statistics > Display Descriptive Statistics. In the Display Descriptive Statistics dialog, select all three variables using the column listbox as before, or simply enter C1-C3 in the Variables field.

Click OK, and the following report appears in the Session window:

Descriptive Statistics: Duration, Interval, Height

This table lists selected descriptive statistics for the three variables in the dataset. This table is helpful, but perhaps not ideal for a report. The overflow onto a second row is awkward, and perhaps other statistics are needed, such as coefficient of variation. The coefficient of variation, defined to be standard deviation divided by the mean and expressed as a percentage, is a useful way of comparing variation between variables with very different mean values.

Recall the Display Descriptive Statistics dialog with the button or the Ctrl + E keyboard shortcut. Click on the Statistics button to display a dialog listing the available statistics. To make room in the report, clear the checkboxes for SE of mean and N missing. Set the checkbox for Coefficient of variation. The dialog should now look like Fig. 1-9. Click OK in the two open dialogs, and the following report appears in the Session window.

FIGURE 1-9 • Descriptive statistics.

Descriptive Statistics: Duration, Interval, Height

The Display Descriptive Statistics command is convenient for generating a neat table for a report. However, reports of statistical summaries often require greater flexibility. When more significant digits are required, or fewer, these options are not available in this Session window report. Also, the fixed-width Courier New font of the Session window helps to align columns of numbers in the table, but it looks old-fashioned.

The Stat > Basic Statistics > Store Descriptive Statistics command is similar to Display Descriptive Statistics, except that the statistics are stored in new columns of the current worksheet, with all available significant digits. From here, the statistics may be copied and pasted into a report for further formatting.

Still Struggling

If you ever have questions about Minitab features or functions, the Minitab help files should be your first reference. Now is a good time to explore the Minitab help features. Most dialogs have a Help button leading to explanations of the available features and functions. Back in the Display Descriptive Statistics - Statistics dialog shown in Fig. 1-9, click the Help button. This displays a Minitab Help window similar to Fig. 1-10. In the right pane, each statistic listed has an underlined hyperlink, which, when clicked, displays a definition and explanation of that statistic. The left pane shows the outline structure of the Minitab Help file, with quick links to help on almost any Minitab topic.

FIGURE 1-10 • Minitab help for descriptive statistics.

Before proceeding, now is a good time to save the Minitab project file.

1.5 Exploring Relationships between Variables

The histograms and descriptive statistics created in the preceding sections consider each column as a separate, disconnected dataset. This section continues the exploration of data in the Minitab sample dataset OldFaithful.MTW by examining connections and relationships between the three variables.

The values in each row of this dataset are connected because of the way the data were collected. Each row lists the duration, interval before, and height of the same eruption. We can explore the possibility of relationships between these variables using scatterplots and correlation coefficients. In particular, we might wonder if the duration or height of each eruption can be predicted by measuring the interval before the eruption.

The scatterplot is the fundamental graph for visualizing relationships between two variables. Start the graphical analysis assistant shown in Fig. 1-5, by selecting Assistant > Graphical Analysis, and then click on Scatterplot. In the Scatterplot dialog, select Duration for the Y column and Interval for the X column. The quickest way to do this is to double-click first on C1 Duration and next on C2 Interval in the column listbox. Click OK, and the scatterplot in Fig. 1-11 appears.

FIGURE 1-11 • Scatterplot of eruption duration vs. interval.

In a scatterplot, each dot represents one row of values in the worksheet. In this example, each dot represents the interval before and the duration of one geyser eruption. The dots form two obvious clusters. One cluster of eruptions has shorter intervals and shorter durations; the other cluster has longer intervals and longer durations.

Visually, it appears that interval is a good predictor of duration. If an eruption starts in less than 75 minutes after the preceding eruption, it will probably be a two-minute eruption. If the interval is longer than 75 minutes, the eruption will probably be a four-minute eruption.

With three variables in the dataset, six different scatterplots are possible by choosing different variables for the Y column and the X column. Each choice of variables provides a different view of potential relationships between variables. For example, to explore whether the height of an eruption can be predicted by measuring the interval before the eruption, a useful scatterplot would plot Height on the Y axis versus Interval on the X axis.

In Minitab, a matrix plot can display many scatterplots in a single graph. To create a matrix plot, select Graph > Matrix Plot from the Minitab menu. The first dialog to appear is the Matrix Plots gallery, as seen in Fig. 1-12. This dialog offers six versions of matrix plots. The top row, labeled Matrix of plots, produces a matrix of scatterplots containing all possible pairs of variables. The second row, labeled Each Y versus each X, produces a matrix with any number of Y variables and any number of X variables.

FIGURE 1-12 • Matrix plots gallery.

Select the Simple matrix plot in the top row of the dialog, and click OK. Then select all three variables Duration-Height in the Graph variables field of the following dialog. After clicking OK, the matrix plot in Fig. 1-13 appears. This matrix shows all six possible scatterplots made from the three variables in this dataset.

Do the scatterplots we have created provide any evidence of a relationship between interval and eruption height? In Fig. 1-13, the bottom-center panel shows Height versus Interval, and the center-right panel shows Interval versus Height. Both of these scatterplots show a square, random pattern of points, with no evidence of correlation or trend. Visually, it is easy to see that eruption height cannot be predicted by interval, or vice-versa.

FIGURE 1-13 • Matrix plot.

Correlation coefficients are measures of linear relationships between pairs of variables; the most common type of correlation coefficient is known as a Pearson correlation. In an Excel worksheet, this is easily calculated using either the CORREL or PEARSON worksheet functions. Minitab can also calculate and display Pearson correlations for many variables using the Stat > Basic Statistics > Correlation function in the Minitab menu.

To try this Minitab tool, select Stat > Basic Statistics > Correlation. Select all three columns in the OldFaithful.MTW worksheet and click OK. The following report appears in the Session window:

Correlations: Duration, Interval, Height

This table displays two numbers for each pair of variables. For Duration and Interval, the first number is 0.870, the Pearson correlation. The second number, 0.000, is a p-value for a test of significance. Since this p-value is small, less than 0.05, we can be very confident that these two variables are positively correlated. This is consistent with the appearance of the scatterplot.

The other two pairs of variables have a Pearson correlation close to zero, and a high p-value, indicating no evidence of a linear relationship between them.

Scatterplots reveal relationships between variables. If these relationships are linear, linear correlation coefficients, such as the Pearson correlation, measure the strength of that relationship. The Pearson correlation is a number between -1 and +1. When the Pearson correlation is positive, higher values of the two variables tend to be associated with each other. When the correlation is negative, higher values of one variable are associated with smaller values of the other. Extreme values of -1 or +1 indicate that all points in the dataset fall along a straight line of negative or positive slope, respectively.

But in real life, we usually want to know something about cause and effect. Do changes in X cause changes in Y? Correlation does not answer this question. The correlation between eruption interval and duration does not say which variable causes the other variable to be more or less. It is possible that an unseen third variable causes both variables to change together.

Correlation alone is not enough to prove cause and effect. Only when correlation is a result of a carefully designed experiment, in which other variables are controlled and measured, is a conclusion about causality justifiable. Regression and experimental tools, which can detect and prove a causal relationship, are discussed in Chapters 8 and 9.

1.6 Brushing Data

This section introduces brushing, a powerful technique to identify a subset of data in one plot and explore features of that subset in additional plots and analysis steps. The example in this section uses scatterplots created from the OldFaithful.MTW dataset in the preceding section.

Using the OldFaithful.MTW dataset, find or create a scatterplot of Duration versus Interval as shown in Fig. 1-11. Can we learn something about the group of 12 eruptions with short, two-minute durations? To find out, look for the brush icon in the Graph Editing toolbar, and click it. Alternatively, select Editor > Brush from the Minitab menu.

The brush function is only available if a Minitab graph window containing certain types of graphs is active. If the brush icon is gray or unavailable, click anywhere in a Minitab scatterplot. This should activate the graph and enable brushing.

When brushing is on, and the cursor is over a scatterplot, the cursor becomes a pointing hand . To select or brush a data point, point the hand to any data symbol on the graph, usually a red dot, and click it. The data symbol changes to a different color, and the brushed data value is also highlighted in the data window and in other graphs, where brushing is on.

To brush a cluster of data on a graph, click the mouse and drag to define a rectangular region, as shown in Fig. 1-14. After releasing the mouse button, all data symbols in the box are selected, and they are said to be brushed. Now, when the cursor is over a region of brushed data, the cursor changes to a flat hand . By clicking and dragging, the region can be moved to brush a different set of data values.

FIGURE 1-14 • Scatterplot with data brushed.

When brushing is on for the current graph, a small window appears called the Brushing Palette. This can be seen at the left of Fig. 1-15. This figure shows several Minitab windows after brushing a subset of data values. The Brushing Palette is the top left window in Fig. 1-15, and lists the row numbers of the brushed data points in the worksheet. The data window includes • symbols to highlight brushed rows.

FIGURE 1-15 • Brushed data highlighted in many windows.

One of the most powerful features of brushing is the ability to see the same brushed points highlighted in every plot which shows data points with individual symbols. The bottom graph in Fig. 1-15 is a time series plot displaying Duration measurements in time order. This important type of graph can be selected from the graphical analysis assistant, or by using the menus and selecting Stat > Time Series > Time Series Plot.

In Fig. 1-15, both the scatterplot and the time series plot show the same set of brushed data points highlighted. To highlight brushed points in multiple plots, brushing must be turned on for each graph separately by selecting each graph and clicking the button. After this is done, points brushed in one graph are highlighted in all graphs.

In this example, brushing allows us to quickly see that the shorter two-minute eruptions are not clustered together in time, but are scattered throughout the dataset. Further, the time series plot shows that a two-minute eruption was always preceded and followed by a four-minute eruption, at least during the time this dataset was collected.

After brushing a subset of data, it is useful to see values of certain variables only for that subset. For a large dataset, it is not practical to scroll through a large data window, picking out only the brushed values. Another way is to select what Minitab calls ID variables, which will appear in the Brushing Palette.

With a graph window active, select Editor > Set ID Variables from the Minitab menu. Then choose any or all columns and click OK. Up to ten columns may be chosen as ID variables. Figure 1-16 shows the Brushing Palette with ID variables Interval and Duration.

FIGURE 1-16 • Brushing palette with ID variables.

Suppose we want to calculate the mean, standard deviation, and other descriptive statistics separately for the short eruptions and for the long eruptions. To do this quickly, first brush the short eruptions in a scatterplot graph. Then, with the scatterplot still active, select Editor > Create Indicator Variable from the Minitab menu. In the Column field, enter a name for a new column, such as Short. Then, select the Update now option, and click OK.

Now look in the data window, and notice there is a new column named Short containing 0s and 1s. In this column, 1 indicates the short eruptions from the brushed data, and 0 indicates the long eruptions.

Now select Stat > Basic Statistics > Display Descriptive Statistics from the Minitab menu. In the Variables field, enter Duration-Height, and in the By variables field, enter Short. If you wish, click the Statistics button and select which statistics to display. After clicking OK, a table like this appears in the Session window.

Descriptive Statistics: Duration, Interval, Height

In this table, the rows where Short = 1 display the descriptive statistics for the short eruptions with a mean duration of 118.67, and the rows where Short = 0 correspond to the long eruptions with a mean duration of 247.58.

This section illustrated the use of brushing with a scatterplot and a time series plot. In Minitab, brushing is available for any type of graph where one data symbol represents one row in the data window. Graph types where one data symbol represents many data values, such as a histogram, do not have brushing available.

1.7 Navigating with the Project Manager

After creating many graphs and analyzing data in many ways, a Minitab project can become quite large. With dozens or hundreds of graphs and analysis results, it can become difficult to find one particular graph or to remember what options were chosen for a particular analysis. The Project Manager makes these navigation and organization tasks much easier.

The fastest way to use the Project Manager is through the Project Manager toolbar, shown in Fig. 1-17. If this toolbar is not visible, place the cursor over any menu or toolbar and right-click. In the toolbar menu, be sure that Project Manager is selected. Here is a quick description of the buttons on this toolbar, along with the associated keyboard shortcuts.

FIGURE 1-17 • Project Manager toolbar.

• The button or the Ctrl+Alt+M keyboard shortcut shows the Session window, which contains a list of graphs and analytical reports in the order they were created. In this view, the Project Manager window shows an outline view of the Session window, listing only the title lines of sections in the Session window. Click any of these titles to jump directly to that section. Right-click on any section title to append it to the Minitab project report or to send that section directly to Microsoft Office Word® word-processing software or PowerPoint® presentation software.

• The button (Ctrl+Alt+D) shows the data window of the current worksheet with the Worksheets folder of the Project Manager listing all the worksheets in the project. Right-click on any worksheet in this window to rename it, bring it to the front, or to enter a description for the worksheet.

• The button (Ctrl+Alt+G) shows the most recently viewed graph with the Graphs folder of the Project Manager listing all the graphs in the project. Right-click on any graph title in this window to append it to the Minitab project report or to send it directly to Microsoft Word or PowerPoint.

• The button (Ctrl+Alt+I) displays the data window of the current worksheet, along with information about each column in the worksheet. Right-click on any column name in the Project Manager to change its name, width, or format.

• The button (Ctrl+Alt+H) displays the History folder, containing all the Minitab session commands for the project. This command history is normally hidden from view, but it documents all actions performed by the user since the project was created. The command language seen in this window is used to write macros to automate Minitab tasks. When viewing the History folder, the left pane displays all folders of the project in outline format.

• The button (Ctrl+Alt+R) displays the ReportPad™, a simple word processor designed to hold the Minitab project report. The ReportPad is a convenient place to organize graphs and analysis results into a report without using any other software. Appending an analysis report from the Session window or a graph from the Graphs folder adds these items to the Report-Pad. From here, the report can be saved in rich text format (.RTF extension) or as an HTML web page.

• The button (Ctrl+Alt+L) displays the Related Documents folder. To add a link to files or web pages relevant to the project, right-click in the folder and select Add Link.

• The button (Ctrl+Alt+E) displays information about the experimental design attached to the current worksheet. This option is grayed out until a design has been defined using the Stat > DOE menu. These features are covered in Chapter 9.

• The button (Ctrl+M) brings the Session window to the front, without resizing or tiling any windows.

• The button (Ctrl+D) brings the data window of the current worksheet to the front.

• The button (Ctrl+I) brings the Project Manager window to the front.

• The button will close all graph windows, after asking for confirmation.

1.8 Creating a Report

After analyzing data and creating graphs, Minitab makes it easy to create a report. If you use Microsoft Office Word or PowerPoint, Minitab can export graphs and analysis reports directly to those programs using the Project Manager shortcuts mentioned in the preceding section. Even without Microsoft Office or any other program, Minitab’s ReportPad is a simple word processor for preparing reports.

Continuing the analysis of the OldFaithful.MTW worksheet featured in this chapter, follow these steps to build a report, using ReportPad.

First, click the button in the Project Manager toolbar to see the ReportPad window. If the report is empty, it will only contain this default title:

Minitab Project Report

Edit the title to something descriptive, and add some brief introductory text, such as this:

Eruption Characteristics of the Old Faithful Geyser

This is an analysis of measurements of eruption duration (seconds), interval before eruption (minutes), and height (feet) of 50 eruptions of the Old Faithful geyser in Yellowstone National Park.

The title of the report is centered by default. To left-justify the text below the title, select Editor > Align Left from the Minitab menu.

Start the report with a histogram of the three variables, either as three histograms or as one histogram with three panels. After creating and adjusting the histogram to be satisfactory, open the Graphs folder of the Project Manager by clicking the button. Then right-click the title of the histogram and select Append to Report. Swich back to ReportPad by clicking the button, and the histogram is there.

Now add a table of descriptive statistics. Using Stat > Basic Statistics > Display Descriptive Statistics, create a table of selected statistics of the three columns in the worksheet. Open the Session folder of the Project Manager by clicking the button. Right-click the Descriptive Statistics section title, and select Append to Report.

Next, add a scatterplot matrix showing relationships between the three variables. If not already part of the project, this can be created with the Graph > Matrix Plot tool. Right-click the graph and select Append Graph to Report. Notice that the report can be built either from the Project Manager window or directly from the graph windows.

Next, add a table of Pearson correlation coefficients, prepared with the Stat > Basic Statistics > Correlation function. In the Session window, right-click the correlation table and select Append Section to Report.

A time series plot provides another interesting view of this data. Figure 1-18 shows a plot made with the Graph > Time Series Plot function. In the gallery, select a Simple time series plot. In the next dialog, select all three variables Duration-Height in the Series field. Click the Multiple Graphs button, and select In separate panels of the same graph. After clicking OK, the plot shown in Fig. 1-18 appears. Append this new graph to the report.

FIGURE 1-18 • Time series plot with three panels.

If printed, the report created so far would look something like Fig. 1-19. Using ReportPad, the report can be enhanced with additional graphs, analysis or blocks of text discussing the findings.

FIGURE 1-19 • Example report.

From ReportPad, Minitab offers options to share the report with the rest of the world, including these:

• Print the report directly with File > Print Report.

• Save the report in rich text file (.RTF) format, or as a web page (.HTM)

• Copy the report to the default word processor. In the Project Manager window, right-click ReportPad, and select Copy to Word Processor.

As always, save the Minitab project often. The Minitab project file includes all worksheets, graphs, history, ReportPad, and other objects together in one .MPJ file. It is not necessary to save worksheets and graphs in separate files, if they are already included in a project file.

1.9 Find Out More

To learn more about any topic in this chapter, keep reading. The following chapters explore many popular types of Minitab functions in depth. The documentation that comes with Minitab software is also quite thorough and useful. Here are the major parts of the online documentation package:

• Meet Minitab is an introductory tutorial which guides the reader through a variety of simple Minitab tasks, using a single case study which evolves through several chapters. This guide can be downloaded in .PDF format for free from the Minitab Web site at www.minitab.com.

• Minitab Help provides instructions for every Minitab function, including a glossary, examples, methods, formulas, and references. Click the button or Help > Help to access the help files using the Windows help viewer. Or, from almost any Minitab dialog, click the Help button to jump directly to the relevant page in the help file. The Minitab assistant dialogs are supposed to be self-explanatory, so these dialogs have no help files.

• Minitab StatGuide™ provides guidance about when, why, and how to apply the tools in Minitab, and how to interpret the results of Minitab analysis. This guide goes beyond how to use the software and explains some statistical concepts in concise language. In Minitab, click the button or Help > StatGuide to access the StatGuide using the Windows help viewer. The button is context-sensitive, and is only available when a graph or Session window report with an associated topic in the StatGuide is active. Clicking will open the StatGuide directly at the relevant page.

• Methods and Formulas is a separate help file containing all the methods and formulas used by Minitab functions. This can be accessed directly through the Help menu. Or, from the help page for any function, click the see also link at the top, and then click Methods and formulas.

In this chapter, we explored many Minitab features by analyzing an example dataset. We created graphs using both the new graphical analysis assistant and the traditional menus. The assistant provides a friendly interface for selecting the most appropriate Minitab graph, while the traditional menus offer more flexibility and options. The next chapter introduces many other types of Minitab graphs and illustrates the powerful features of the Minitab graphical engine. Chapter 3 provides an in-depth survey of the Minitab environment, including windows, toolbars, and customization features.

QUIZ

1. Which Minitab window holds analytical reports and tables generated by functions such as Display Descriptive Statistics?

A.