A QUICK INTRODUCTION TO MINITAB 16

Amonng other winndows are thhese:Project managermwinndow The project mannager windoww shows eithhera schemmatic layoutt of all the wwork currentlly active (cllick on ShowwWorkssheets Foldeer, the icon wwith cascadiing windowss)orrdetailss about all thhe columns inn the active worksheet (cclick onthe Show Info iconn),Graph winndows will beb created ass needed.The Histoory windoww can be acceessed by cliccking on the Show historry icon onthe toolbaar. This is thhe chevron symbol on a spiral noteboook. This wwindowprovides a log of the activities,aannd it will inc lude the commmand formms equivalenttto all the mousemactionns that you havehdone.The RepoortPad winddow provides a place in wwhich you ccan composee and editdetails about all the work.wYou caan reach thiss by clickingg on the ShoowReportPaad icon, whiich is the redd letter A on a spiral noteebook. You can cut andpaste fromm the sessionn window to the Reportppad, and youu can also paaste graphicaaloutput intto this locatioon.You mightmlike too explore somme of the othher features on the toolbbar.w any MS--Windows appplications, youy can use the double-hheaded cursoor arrows toAs withresizee these winddows. Click on the Winddow commaand to get a llist of availabble windowss.Wipping Minnitab CleaanWhenn you first sttart up Minittab, all its woork areas wiill be empty. If you havee finishedworkking on a parrticular task anda wish to start anotherr, first be surre that you hhave savedthe reesults in a foorm that you want. Thenn do File NNew and sellect Minitabb Project.This act will emppty out all the working arreas.This actioon will alwayys get the quuestion “Sav e changes too the project ‘Untitled’before cloosing?” evenn if there is nothingnto bee saved.Reaading Proojects intto MinitaabIf youu wish to usee a project previouslypasssembled by Minitab, use command File Openn Project. ThereTare no choices or optionsoin reading projeccts, so this prrocess is verrysimplle.Paage 6

Reaading Woorksheet Data intoo Minitaabu would like to read a sinngle worksheeet into Miniitab. Single worksheetsSuppose that you(of thhe type *.mtpp and *.mtww) were previiously assemmbled by Minnitab and aree simple toread. Files assemmbled outsidde of Minitabb require somme care to reead properlyy. The mostcommmon types arre *.dat, *.txt, *.csv, *.xlls, and *.xlsxx.First note whether a projectpis currrently activee. If so, anyy worksheet tthat youbring in willw become partp of this project.pThiss may, at timmes, be exacttly what youuwant. If youy want thee new file separated fromm other data,, please see tthe sectionWiping MinitabMCleaan.Beginn with File Open Woorksheet.f type musst show in thhe Files of tyype panel. IIf the file typpe does not show, clickThe fileon the drop-downn symbol tot display alll the file typpes.Opannel becomes available whhen ask to reead a spreadssheet of anyy variety otheerThe Optionsthan *.mtp or *.mmtw. In suchh cases, youu will need too provide infformation soo that Minitaabws how to reaad it.knowIf you aree unsure abouut the formaatting of yourr file, you mmight find it hhelpful toexamine iti in a text prrocessor such as WORDD or WordPad or NotePaad.Clickking on Optiions will lead you to thiss panel:Paage 7

You will need to examine the files to see what formatting details need to besupplied to Minitab. Watch out for*whether or not the file has a row of variable namesThis is related to the Variable names section.Select None if there are no names and the datavalues begin in row 1.Select Automatic if there is a header row withthe names of the variables.***the strategy by which the file separates fields (open space for “freeformat,” tabs, commas, or other characters)This is related to the Field definition section.Select Free format if data values are separated byone or more blanks spaces and rows endwith hard carriage returns.Select Single character separator if a particularcharacter is used to separate values. Themost commonly used separators are tabs,commas, and single spaces.whether or not alphabetic information contains internal blanks thatmight conflict with spaces used as separatorsThis is related to the Text delimiter section. You mightneed to put city name “Los Angeles” in quotes so that itwill be interpreted as a single item of information. (It isusually easier to use tab-separated fields to prevent thisproblem.)the technique that is used to denote missing dataThis is related to the Convert missing section. Minitabuses the asterisk * as its missing data marker. If, forexample, your input file uses NA for missing data, youshould ask to convert NA to *.Now click OK and then supply the file name. You can use Preview to see if the file willbe read as anticipated.If you read a file without variable names, you should supply names based on yourunderstanding of the data. Save the file in storage space that you control.At times, the data read this way can end up as the wrong type (text rather than numeric).This issue is addressed as the final item on page 19.Page 8

Managing the Minitab ScreenMinitab always has its two basic windows, but it often creates many other windows aswell. Screen management is highly idiosyncratic, in that computer users seem to havevery different opinions about what a screen layout should look like. Most users arereasonably familiar with windows management.Minitab, like other MS-Windows programs, can arrange its own screen. Click onWindow at the top of the screen and see what happens when you select any of thechoices Cascade, Tile, or Minimize All.The two basic Minitab windows can be made full screen, partial screen, or reduced toicons, but they cannot be closed. These windows are only closed when you quitMinitab.Use the Show graphs folder icon on the toolbar to manage your graphs. This symbolis the pale-colored set of cascading windows. The sequence Window Close AllGraphs will let you eliminate all graph windows at once.Saving Minitab’s OutputYou will of course want to save the results of your work. The usual methods are to cutand paste to another MS-Windows application or to save to file. Minitab also allowsyou to save to the Minitab ReportPad.Text output is moved through the copy or cut-and-paste methods common toMS-Windows applications. If you are working with a project file (*.mpj) then you canalso paste the selected text in the ReportPad. This is accessed through the icon which isthe red letter A on a spiral notepad.If a graph window is active, you can use Edit Copy graph or Ctrl C to place theimage on the MS-Windows clipboard. The image can than be pasted in anotherdocument or placed on the Minitab ReportPad.Pressing Alt PrtScr copies a bitmap of the active window onto the MS-Windowsclipboard. This can be useful if you wish to describe how you did something in Minitab.The figures in this document were produced in exactly this way.If the active window is a dialog box, Alt PrtScr copies just the dialog box to theClipboard. If the active window is Minitab itself, Alt PrtScr copies the Minitab framewindow and all visible subwindows (such as the Data window and Graph windows) tothe clipboard.Page 9

Preparing a Minitab GraphThe data in this illustration are from Exercise 3.69 of Hildebrand, Ott, and Gray, fileEx0369.mtw. The data set notes the numbers of tax forms making various numbers ofdeductions, ranging from 0 to 12. We’d like our graph to have deductions on thehorizontal axis and frequencies on the vertical axis. Use Graph Scatterplot, thenselect Simple. and then put this into the resulting panel:If you take the default settings, you’ll get this:Scatterplot of Frequenc vs e are many ways in which Minitab will allow you to change the appearance of thisgraph. When the graph window is active, use the mouse to double-click on any featurethat you would like to edit. If you double-click on a symbol (meaning the circles here),you can change the shape, size, and color of the symbols. You can similarly editcaptions and scales, and you can insert grid lines.You can get additional editing power through Tools Toolbars Graph AnnotationTools. Editing features such as text insertion, lines, circles, ellipses, and so on will thenbe available on the Minitab toolbar when graph windows are active.Page 10

ng Minitab to do a RegresssionUsinLet’s illustrate thhe use of regrression throuugh the file XXM1216.mttw from Hilddebrand, Ott,and Gray.GThis is a Minitab worksheetwfiile, and it cann be read eaasily by Miniitab. Pleasesee thhe instructionns given in thet section ReadingRWorrksheet Datata into Minittab.The ProjectPManaager windoww (obtained byb clicking oontoolbar) will havee this informmation:, the Shhow info icoon on theThis is summary informationn about the data set. It giives the variiable names, thecorresponding MinitabMcolummns, number of values inn each colummn (and how many ofthosee are missingg) and also thhe variable tyype. Here N stands for nnumeric. Coolumn C1 issnameed s (for salees) and this willw be the deependent varriable in ourr example regression.To start the regreession, do Stat Regreession Reegression. SSet up the neext panel toshoww this:k on Graphss, and then seelect Residuuals versus ffits and Resiiduals versuus order.ClickReturrn by clickinng OK, and thentclick onn Options; cchoose Variiance inflatiion factorsand Durbin-WatDtson statisticc. Click OKK. This will return you tto the main rregressionpanell, and you cllick OK agaiin.vwithh the other thhree variablees used asThis action selectts C1 as the dependent variable,indeppendent variaables. This hash also askeed to see twoo graphs andd some inciddentalcalcuulations as well.wThe numericalnouutput will appear in the sessionswinddow, and youu will also geet twographhs. A little screensmanaggement will be needed too examine all of these.Paage 11

Regression is a fairly intense operation, and there are many secondary calculations anddisplays that are of some interest. We’ll discuss a few of these.Suppose that you wanted to list the individual residuals for your data set. The initialprocedure would include an additional step from the panel reached by Stat Regression Regression. That step consists of clicking on Storage and then selectingResiduals.When the regression is completed, a new column RESI1 will appear in the data window.(Subsequent uses will create RESI2, RESI3, and so on.)Suppose that your work with the data set has indicated that you might want to omitcertain data points (rows) from the analysis. Perhaps such data points were destructiveoutliers or were simply miscoded. You’d prefer not to delete the rows completely, sinceyou might want to restore them later. A plausible solution (but not the only solution) is tocreate a copy of the worksheet so that you can delete selected rows in the copy. Onceyou’ve created multiple worksheets, the work should be saved in project *.mpj format.You can make a copy by doing File New Minitab worksheet. At this point ablank worksheet will appear. Now go to the original worksheet and place the cursor inthe name box for the first column, C1. Hold down the shift key and use the arrow keys tomark the entire worksheet as a block. Go to the blank new worksheet, place the cursor inthe name box for the first column and press Ctrl V. A copy will now appear in the newworksheet, and you can edit it as appropriate. You can give a name to this copy; accessthe name by clicking on the Show Worksheets Folder symbol on the toolbar. Thenclick with the right mouse button on the icon, and select Rename. It is stronglyrecommended that you leave yourself a description as to exactly how this new worksheetdiffers from the original. Use Set Description to write notes about this worksheet.If you wish to delete a row in the copy, you use Data Delete Rows. Be extremelycareful if you are deleting more than one row. Each use of Data Delete Rows willcause the remaining rows to be renumbered! If you want to delete multiple rows, it’srecommended that the operation be done all-at-once, rather than one row at a time.Creating New Variables through TransformationsThere will be many occasions in which you need to create new variables which arefunctions of other variables. For example, you might need to create log(SALES), thebase-e logarithm of the variable SALES, which is in C1. Click on Calc Calculator.In the box next to Store result in variable type in the name Logsales. (If you type C1 inthis spot, the calculation will overwrite the original values.) In the box for Expressiontype in LOGE(Sales) or type in LOGE(C1) and click on OK.Page 12

As a result of this process, a new column Logsales will appear in the data window. Youcan of course overwrite an existing column, but then the original information in thatcolumn will be lost.This use of the calculator option is simple, but there are some subtleties:You need not remember function identifications such as LOGE for base-elogarithms. The Calculator has a Functions panel that lists the available choices.If you make an entry-by-entry calculation, as in the example above, then the resultis a new column in the worksheet. If you calculate a single-valued function of acolumn (such as the median), then your transformation will create just onenumber, occupying one position in the data window worksheet.Logical operations take the values 0 and 1, so that you can use these to maketransformations of the “if-then-else” form. Suppose, for example, that you’dlike variable C3 to be defined as follows:C3 RS 5TC2if C2 5if C2 5The expression that makes this work is5*(C2 5) C2*(C2 5)The yes-or-no logical statement (C2 5) will either take the value 0 or thevalue 1, as will the yes-or-no logical statement (C2 5).Missing values get passed along as missing. If you are creating a new column C6from columns C2 and C3, and if a value is missing in row 28 for either C2 or C3,then the row 28 entry for C6 will also be missing, shown as * in the data window.Invalid calculations end up as missing. Thus LOGE(0) will result in a missingvalue, as will LOGE(negative number). These will be shown as * in the datawindow. Minitab gives you a warning when you create missing values in thisstyle, and thus you get an opportunity to reconsider the appropriateness of thetransformation. (Releases 13 and previous did not give the warning.)Be sure to see also the section on Formulas, page 16.Page 13

Usinng Minitab to Finnd Probaabilitiesc use Minnitab to find manymprobabbility calculaations. For iinstance, if yyou want toYou canfind thet probabiliity that a binnomial randoom variable wwith n 23 and p 0.733 takes thevaluee 15, you calll up Calc Probabilitty Distributiions Binoomial. Thenn set up theresultting panel ass follows:Whenn you click OK,O the sesssion windoww will then shhow the folloowing:Prrobability DensityDFuncctionBiinomial withh n 23 and p 0.73x155P( X x )0.1233775This indicates thaat your desirred probabiliity is 0.1233 75.u had wantedd to find thee probability of a value leess than or eequal to 15, yyou’d doIf youexacttly the same as the abovee, except thaat you wouldd click on thee CumulativveProbbability buttoon.That result showss up in the seession windoow asCuumulative Distribution Fuunctioninomial witth n 23 and p 0.73Bix155P( X x )0.2655051This gives your probabilitypas 0.265051.It’s also possible to give the completecproobability disstribution. SSuppose that you wanteddto seee the whole sets of probabbilities for thhe binomial random variiable with n 23 andPaage 14

p 0.73. This is done by setting up a datta column foor the integerrs 0, 1, 2, , 23. Clickon Caalc Makee Patterned Data Simmple Set of Numbers. Set up the reesultingpanell as follows:There is another wayw to do thiss. In columnn C1, type 0 in position(C1, roow 1) and 1 in position ((C1, row 2). Mark thesee two boxesand theen place the mouse cursoor at the lowwer right position until thhesymbool so that the symbols aappears. Drrag downwaard and thecolumnn will fill upp with conseccutive integeers. This is the styleused inn Excel. Thhe Calc MMake Patterrned Data Simple Seetof Nummbers sequeence is needeed for more exotic numbber patternssuch as 1 1 2 2 3 3 4 4 .Thenn with Calc Probabiliity Distribuutions Binnomial, set uup the panel as indicatedd:Paage 15

This illustration assumes that columns C1 and C2 are not being used for other purposes.At the completion of this command, the worksheet in the data window would havecolumns C1 and C2 showing the following information (slightly rearranged 90630.0643820.0248670.0061120.000718Here we can see, for example, that the probability of exactly 15 successes is 0.123375.The feature Calc Probability Distributions will allow you to deal with manydifferent probability distributions.FormulasEmbedded formulas were introduced with release 15. These are not available in StudentMinitab or in Minitab 14 or any earlier releases.An embedded formula is an Excel-type task, and Minitab and Excel are used in differentstyles. So be careful. Please be aware that, as noted in Minitab help, “In addition,formulas cannot be imported or exported to or from Excel or other programs.”You can assign formulas only to columns or to constants. (Constants are not discussed inthis document.) The formulas are precisely those of Calc Calculator. Thus, youcould assign column C9 as SQRT(C8).You cannot assign C9 as the residuals from a regression.Column assignments must be based on all rows.Let’s say that you want C3 to be the sum C1 C2. With the cursor in any cell of columnC3, use Editor Formulas Assign Formula to Column. A small will appear inthe header cell for column C3.Missing data positions in C1 or in C2 become missing positions in C3. If columns C1and C2 have different lengths, the length of C3 will be the smaller value.You generally want formulas to be updated automatically. In this example, you wouldlike a change in either C1 or C2 to be reflected instantly in C3. (Here “instantly” meansPage 16

“as soon as the cursor leaves the cell that was changed.”) The instant change isMinitab’s default strategy, as noted in Editor Formulas Calculate All FormulasAutomatically.If you turn this feature off, then*A small – will appear in the header cell for a column which needs anupdate.*Updating can be done through Editor Formulas Calculate AllFormulas Now. This does all updating; you cannot do updating for justone formula column.You can convert a formula column to an ordinary column. Use Editor Formulas Remove Formulas from Selected Columns.Formulas are very effective if you might have to correct errors in the input data, as anycorrected values will be transformed automatically. Suppose that you have a new dataset, with 5 columns and 244 rows. You want to develop a multiple regression of C1 on(C2, loge(C3), C4, C5). Assign formula LOGE(C3) to column C6. If you repair inputerrors in C3, the corresponding corrections will be made automatically to the logarithmin C6.If you copy all or part of a worksheet to a second worksheet, then the columns withformulas become ordinary columns in the second worksheet.Graph UpdatingIn certain cases you might want Minitab’s graphical output to be revised if data values inthe worksheet are changed. The symbol at the left end of a graph’s title bar will revealits status with regard to updating.A green cross in a white circle indicates that the graph is up to date.A yellow circle indicates that the worksheet values were changed and that thegraph is not up to date.A red X in a white circle indicates that the worksheet h

Sep 02, 2010 · created in Minitab 16 or in Minitab 15 cannot be read by previous releases of Minitab and cannot be read by Student Minitab. See page 3, point [4], for the work-around solution. In addition, Minitab can read and write Exc