SPSS for the Classroom - Practical Exercises

Analyze - Descriptive Statistics - Frequencies Select one or more variables in the selection list on the left, and move them into the analysis list on the right by clicking on the arrow in between. Then click OK. Syntax: frequencies variables = gender minority. Output: Note that this is one of the few tables where missing values (whether system missing "." or user designated missing) show up in the default output table (however, not in this particular example). Descriptives: Means and Standard Deviations The mean and standard deviation of a variable are such fundamental quantities in statistics, that there are many SPSS commands that will report them to you. The most straightforward command to use is Descriptives. Two other useful commands are Frequencies (in the dialog box, click on the Statisticsbutton), when you want to see counts as well as means and standard deviations (perhaps for Likert scales), and Explore, which gives you such additional statistics as the median and interquartile range as well as a variety of graphs. Menu and Dialog Box: Analyze - Descriptive Statistics - Descriptives Select one or more variables in the selection list on the left, and move them into the analysis list on the right by clicking on the arrow in between. Then click OK. Syntax: descriptives variables=educ salary. Output: Histograms SPSS has three different sets of commands for producing graphs. The easiest to learn and use are the oldest "legacy" graphing commands. They give you graphs with a default visual style (colors used, weight of lines, size of type, etc) that can be customized by hand. Histograms are vexing because they can be alternately informative or deceptive, depending upon how the bins (the bar boundaries) are chosen. They are useful and popular because they are conceptually very simple, easy to draw and interpret, and if drawn well they can give a good visual representation of the distribution of values of a variable. Menu and Dialog Box: Graphs - Legacy Dialogs - Histogram The basic histogram command works with one variable at a time, so pick one variable from the selection list on the left and move it into the Variable box. (A useful option if you expect your variable to have a normal distribution is to Display normal curve.) Syntax: graph /histogram(normal) = prevexp. Output: In this example, the distribution of the data is nothing like a normal distribution! To edit colors, titles, scales, etc. double-click on the graph in the Output Viewer, then double-click on the graph element you want to change. Crosstabs: Counts by Group The basic crosstabs command just gives you counts by default. Typically it is useful to also look at either row-percents or columnpercents, which must be specified as options. Menu and Dialog Box: Analyze - Descriptive Statistics - Crosstabs Select one variable as the rows, another variable as the columns. Conventionally you might put an independent variable in the rows and a dependent variable in the columns, although mathematically it doesn't really matter. To get percents in your output, click on the Cells button and specify the kind of percents you want to see. Syntax: crosstabs /tables=jobcat by minority /cells=count row. In this command syntax (and the next one, the key word MEANS), you see BY used to specify a categorical variable that divides the data into groups. Output: Means by Group Menu and Dialog Box: Analyze - Compare Means - Means Select the variable(s) that you want means of, and move it to the Dependent List. Select the variable that divides the data into subsets (the "grouping" or "by" variable) and move it to the Independent List. You may have more than one variable in either/both lists, and SPSS processes them in pairs and produces separate tables. Syntax: means tables=salary by minority. Output: Bar Charts Similar to a histogram, the x axis is treated as a categorical variable, and the y axis represents one of a variety of summary statistics: counts (a.k.a. a histogram!), means, sums, etc. Menu and Dialog Box: Graphs - Legacy Dialogs - Bar This takes you through an initial dialog box, where you choose among several basic schemas for making bar charts, and then to the main dialog box. To graph means by groups, select Other statistic for what the bars represent, the variable for which you want to calculate means in the Variable box (means will be the default statistic), and the group in the Category Axisbox. Syntax: graph /bar=mean(salary) by jobcat. Output: Boxplots Menu and Dialog Box: Graphs - Legacy Dialogs - Boxplot As with bar charts, you first choose a specific boxplot schema from an initial dialog box, and then choose the analytical variable (the one you want to see medians and interquartile ranges for, the y axis), and the categorical variable (the x axis). Syntax: examine variables=salary by jobcat /plot=boxplot /statistics=none /nototal. Output: Correlation Menu and Dialog Box: Analyze - Correlate - Bivariate SPSS calculates bivariate correlations (the Pearson's r) for all pairs of variables in the list. Syntax: correlations /variables=educ salary prevexp. Output: Scatter Plots Both simple scatter plots and scatter plot matrixes are pretty easy to produce. Menu and Dialog Box: Graphs - Legacy Dialogs - Scatter/Dot Takes you through two dialog boxes. First you choose the scatter plot schema you want to work with, And then you specify the variables with the x and y coordinates of the points you wish to plot. Syntax: graph /scatterplot=salary with salbegin. graph /scatterplot(matrix)=salary salbegin prevexp. Output: T-tests T-test can be used in a variety of ways, and SPSS gives you quick access to three of them (univariate, grouped, and paired) through the Compare Means menu. They all access the same test command. Menu and Dialog Box: t- Analyze - Compare Means - Independent-Samples T Test When setting up an independent-samples (grouped) t-test, you not only specify the variable being tested and the grouping variable, but you also have to specify which data values represent the two groups you want compared (because in general the grouping variable might have an arbitrary number of categories, not just two). Use the Define Groups button, and type in the data values (not the value labels) that define the groups being compared. If you type in an invalid data value for any of the groups, SPSS will not catch your mistake until you actually run the command. You need to know what your data look like before you get to this dialog box, because SPSS will not let you browse your data set while a dialog box is open. Syntax: t-test groups=gender('f' 'm') /variables=educ. Output: Chi-square Tests Like t-tests, chi-square tests come up in a wide variety of circumstances, the most common of which is assessing the independence of two variables in a contingency table (a crosstab). So this chi-square test is specified as an option on a crosstab command. Menu and Dialog Box: Analyze - Descriptive Statistics - Crosstabs In the main dialog box, click on the Statistics button, then select Chi-square and Continue back to the main dialog box. Specify your variables and run. Syntax: crosstabs /tables=jobcat by minority /statistics=chisq /cells=count row. Output: ANOVA Tables and Tests ANOVA tables are a core concept in statistics, and they are produced by several different commands in SPSS, ONEWAY, GLM, and UNIANOVA. The UNIANOVA command is perhaps the easiest to use including overall, because it allows you to use string (character) variables as factors. (If you are doing a one-way ANOVA and your factor is coded in numeric form, then ONEWAY is even easier to use.) Menu and Dialog Box: Analyze - General Linear Model - Univariate For a simple ANOVA, your factors are considered Fixed Factors. If you have more than one factor and you do not want to include interactions in your model, you will need to specify that with the Model button. Syntax: unianova salary by jobcat. Output: Regression Menu and Dialog Box: Analyze - Regression - Linear Syntax: regression /dependent salary /method=enter salbegin. Output: "' •spss2.spv [Document2J • IBM SPSS S1a11S11cs V,e:wff [lle §_di в 'f.ll!!W Qata Iransloon /D!.nN� salary Тilll N:JlitDatast1 �Scatttrofsat �Loo 1!] Gtap/1 (cjTolle Notes �LOO Scatl:tr 01 lal l!J T•TISI ,ajrde eNotes At:J.нeDatastl (i Gtoop SlaH•i lndlpendtntS feLoo li (!1 Cross�Ds Tde eiNotes N.'JN9Data.set {.а, Сэsе Process, (iEm111Dym1ntC " Qtaphs ,(j]Toll1 1!] Gtapn е Note• в �е /.!tA!ISTICЗ COt:f'f' otrr! R A..�OVA �L"9 li 1!1 Gtaoh li Fomiat NJ/Nt Datast1 Qj Сопеlа ons �Loo eNolll Adw-e Oatasel WatnxolIOUC в ]nselt (i Cni--Square те !,'tilюes /CIUTtRtA•frN(. 0SI РОUТ ( .10) /NOORIGIN /Мl'lHOD=!:NYU salbeg1.n. + Regresslon (Dac.aSec.l) С: \P:rog:raa !'1.les\lВН\SPSS\Sc.ac.1.st1.ca\l9\Su.ples\�nglз.sh\t:aiplo Model Vari.-es Entered!RemcJwdЬ V1naЫ11 Entertd Beg1nnlng sa1ary VanaЫe1 Removed MelhOd Enter а "1l 11questedva11aыes entered ь DepenctentVar1aь1e· Current81lary Model вво• 115 774 Sumo1 Squares "' Std EltOf Of the Esllmate 10.115 356 а Prtdic10�. (ConstanQ, вeoln.nlng Salar; Model 1 Regresslon Resldual Tot.al �LOO � Untvatlatt Analylil 1 068Е11 З 109EIO 1 412 473 1 Э79Е11 Mean Square 1 068Е11 65658997.21 ' Slg 1622118 ООО' IGi Tdle Notll Аdм, Dataset (i В.tween-Suo, -Qj Ttsts о1 В.1.vе Loo r§J R1g,esslon е Not1s +,(] Tdle N:Jlr.e Oataset vanaьt11En • � Ltodtl Summa {IIЖOVA � Cotllleltnts coemcten1s• Modet 1 Unstandardaed Сое ctents (ConstanQ Beginning Зalary в 1928.106 1 909 а Dependenl Vanaыe· Current Salary 1 1 Std Error 888 680 047 Standarcttzed Сое cIents в.ь, 880 1 2 170 40 276 Sig 031 ООО