LIR 832: MINITAB WORKSHOPSEPTEMBER ?, 2006Making Dummy (or Indicator) VariablesLIR 832: MINITAB WORKSHOPSEPTEMBER ?, 2006Opening MinitabMinitab will be in the Start Menu under “Net Apps”.Opening the DataGo to the following web site:http://www.msu.edu/course/lir/832/Datasets.htmRight-click and “Save Target As…” on the link for “20 Percent Subsample” (Save it to the desktop)In Minitab, select “File” “Open Worksheet”  and find “lir832-managers-and-professionals-2000-20pct-sample.MTW”.Learning the Variable NamesThis data set includes 10,968 observations on a number of characteristics of specific professional and managerial employees. See the corresponding codebook for definitions.region state age mstatus gender edattain race hispanic esr uhour1 workcls1 psic1 pocc1 weekearn wage3 parttime cbc2 yearsedBasic StructureMinitab saves all materials related to a session as a “Project”. This includes all graphs, datasets, results, commands, etc. Within this project is a “worksheet”, which is the raw data we use here.Variable Names: You can add or edit variable names by clicking on the names above the columns. Note that each column will either feature numeric or textual (non-numbers) data. In those textual columns, the heading will end with “-T” automatically. Results appear hereData here (the worksheet)Statistical CommandsVariable NamesAdding Data: You can add data in a number of ways. You can either hand-enter data…… or you can make patterned data under the “Calc” “Make Patterned Data”  “Simple Set of Numbers” (or other options) Notice: Empty ColumnMaking Dummy (or Indicator) VariablesSuppose you want to convert the “Sector” column into a series of 0/1 variables in order toinclude them in a regression equation (you can’t just put the “Sector” variable in the equation since it is non-numerical). Go to “Calc””Make Indicator Variables…”The variable “Race” has 4 values. Thus, we must create four dummy variables… which needs four empty columns. Thus, we input C20-C23 and we find the following…:You can then label each column based on what it represents.Where to put your new variableVariable to examine4 New Variables – One Dummy Variable for Each SectorDescriptive StatisticsWe can quickly calculate the descriptive statistics for each numeric variable. Follow the following steps:Click on appropriate variables (text variableswill not appear).After clicking on the “Statistics…” tab, the following menu pops up, allowing you to select all the statistics you will need for your analysis:Upon hitting “OK” on your main selection menu, your results should look like…:Results for: lir832-managers-and-professionals-2000-20pct-sample.MTW Descriptive Statistics: age, weekearn Variable N N* Mean SE Mean StDev Minimum Q1 Medianage 10968 0 42.142 0.111 11.606 15.000 33.000 42.000weekearn 9506 1462 895.59 5.71 556.81 0.01 519.00 769.23Variable Q3 Maximumage 50.000 89.000weekearn 1153.84 2884.61Descriptive Statistics By Category: You can also calculate descriptive statistics by sub-category. For example, if we want the descriptive statistics by “Gender”, we can do the following:Descriptive Statistics: age, weekearn Variable gender N N* Mean SE Mean StDev Minimum Q1 Medianage 1 5507 0 43.015 0.159 11.833 15.000 34.000 43.000 2 5461 0 41.261 0.153 11.306 15.000 32.000 41.000weekearn 1 4533 974 1077.3 9.02 607.4 0.03 653.8 961.5 2 4973 488 729.99 6.33 446.04 0.01 442.00 653.84Variable gender Q3 Maximumage 1 52.000 89.000 2 49.000 89.000weekearn 1 1384.6 2884.6 2 925.00 2884.61Will give usdescriptive statistics foreach genderUsing Evocative Names for the Data:1 and 2 are not very evocative for male and female, so we want to recode our gender data into men ( = 1) and women ( =2).Now Generate the Descriptive Statistics Replacing ‘Gender’ with ‘Sex’Basic GraphingMinitab allows you to quickly and easily create graphs of different variables. If you look under the “Graphs” command, you’ll find:SCATTERPLOT: There are many options one can use with the scatterplot command, including asking Minitab to include the regression line. SIMPLE SCATTERPLOT OF WEEKLY EARNING (“WEEKEARN”) (Y) vs. YEARS OF EDUCATION (“YEARS ED”) (X):years edweekearn22.520.017.515.012.510.07.55.0300025002000150010005000Scatterplot of weekearn vs years edOR, WITH ADDING THE “WITH REGRESSION” OPTION, WE’D FIND:years edweekearn22.520.017.515.012.510.07.55.0300025002000150010005000Scatterplot of weekearn vs years edHISTOGRAM (OF “Weekly Earnings”):weekearnFrequency2800240020001600120080040006005004003002001000Histogram of weekearnBOX-AND-WHISKER PLOT (“BOXPLOT”) OF WEEKLY EARNINGS:weekearn300025002000150010005000Boxplot of weekearnEDITING TITLES: On your graph, right-click on the title “Boxplot of weekearn” to change the title text, font, etc.Now, change the title to “Weekly Earnings Among Americans”, and make the title text 18-pt font, in Times New Roman, in the color red.weekearn300025002000150010005000Weekly Earnings Among AmericansNote: You can also change the X-axis and Y-axis labels in similar fashions.Comparing the Weekly Earnings of Male and Female Managers (splitting our box plot by sex)The MINITAB SYSTEM: Help is always availableHelp is always a click awayHistory Radio Button: HistoryHistory button gives you access to prior commandsand other materials.LESSON CHECK – EXERCISESWith the techniques learned in this lesson, complete the following exercises:1. Change the variable name of “esr” to “Labor Status”.2. Calculate descriptive statistics for individuals by race (1=white, 2=black, 3=asian-american, 4=native-american).3. Create a histogram of wage3.4. Create a scatterplot of “years ed” (x) vs. wage3 (y).5. Now create a scatterplot with a regression line of “years ed” (x) vs. wage3 (y).6. Retitle this graph, and change all the labels and titles to be in blue