To write a new Syntax program, do the following:

Figure 15.1. This is where you write Syntax programs.

Syntax programs are tied tightly to the variable definitions in the currently loaded dataset. Syntax programs use the dataset's variable names, often in such a way that the type of the variable can be important. Thus the first instruction in a Syntax program is usually to load the data file.

You can load a file by choosing File

GET FILE='C:Program FilesSPSSInc
  PASWStatistics18SamplesEnglishEmployee data.sav'.

The following is a program with a simple GRAPH command, using the salary and job category information of the loaded data:

GRAPH TITLE = "Means of Salaries"
  /SUBTITLE = "separated by job category"
  /BAR = MEAN(salary) BY jobcat.

The resulting display in SPSS Viewer is shown in Figure 15-2.

Figure 15.2. A bar graph produced by a Syntax command.

To load a Syntax program from disk, choose File

To save your program, first decide where you want to save it and what you want to call it. Then, in the Syntax Editor dialog box, choose File

Often you'll want to leave your original program as it is and create a new one by making changes to the original. In that case, load the original program from disk and then choose File

Every SPSS menu selection is nothing more than a command to execute a Syntax program. Adding a new item to the menu is a matter of adding a new menu button and assigning a task to it.

You can add new menu selections to customize SPSS and make it easier to do your common tasks. For example, if you're working on a data file and loading it regularly, you could define a new menu button to load the file for you. If you have an analysis or a report-generating program that you run regularly, you could define a menu button that runs it with your set of parameters. Or you could set up a button to export data in your preferred format.

To define terms clearly, a menu consists of the menu bar (the part that's always visible at the top of the window) and its row of pull-down lists. Each list is made up of clickable buttons. Each button can be set to execute a particular Syntax command or to display another list of buttons. You can modify a menu by adding a new pull-down list or by adding a single button to an existing list. You can delete existing menu items, but usually there's no need to do that — your modifications will almost always be for the purpose of adding buttons that perform your own tasks.

The following steps take you through the process of adding a new pull-down list with one button. The button executes the Syntax program named loadfile.sps, which is a program consisting of one GET statement to load a file:

Figure 15.4. The MyFile selection has been added to the Open menu.

If you want to add this same option to the menus of other dialog boxes in SPSS, you have to follow the same procedure for each one. Adding a menu item takes only a small amount of work and can prevent many repeated steps. For example, if you're in the process of entering and correcting data, a simple menu item to load the file would keep you from hunting for the file every time you need to load it. Also, if you have a group of analyses that you run repeatedly, you could include them all in one Syntax program and have them all run for you at the click of a button. The program could be made to load the file at its startup, so one click of a button is all you need to do all that work.

There are lots of Syntax commands, and they all have lots of options. If you have something you want to do, and you want to find the Syntax command to do it, you have three basic approaches.

One way is to use the menu system to command SPSS to do whatever it is you would like it to do. In SPSS Viewer, you will be able to see the text of the Syntax commands that generated the output. Highlight that text, choose Edit

Another way to find commands is to select Help

A third approach is to select Help

You can write a Syntax program to do more than one thing. All the commands in such a script are executed one after the other. And because one Syntax command can do quite a bit, you don't have to write much of a program to do lots of processing. For example, the following four-line program named makeplot.sps performs four separate tasks:

GET FILE='C:Program FilesSPSSIncPASWStatistics18
  SamplesEnglishCars.sav'.
DATASET NAME DataSet1 WINDOW=FRONT.
GRAPH LINE=MEAN (HORSE) BY YEAR.
GRAPH BAR=MEAN (MPG) BY ACCEL.

The first two lines load the SPSS data file named Cars.sav. The third line renames the dataset to DataSet1 and brings the window displaying it to the front. As a result, if data has already been loaded and named DataSet1, this new file assumes the name (the other will be closed). The last two lines draw graphs — one line graph and one bar graph, as shown in Figure 15-5 and 15-6.

Note how variable references are made on the GRAPH commands. Referring to a variable by its name specifies that all its values be used; using the word MEAN before the variable name in parentheses specifies that the mean of the variable's values be used. These commands are simple but the actions are complex.

Figure 15.5. A line graph displaying the mean horsepower for each year.

Figure 15.6. A bar chart displaying the mean acceleration for each mile-per-gallon rating.

The Syntax language contains the PPLOT command, which you can use to generate either of these types of plots:

The following program, named makeplot2.sps, contains commands to produce both.

GET FILE='C:Program FilesSPSSIncPASWStatistics18
  SamplesEnglishEmployee data.sav'.
DATASET NAME DataSet1 WINDOW=FRONT.
PPLOT SALARY /TYPE=Q-Q.
PPLOT SALARY /TYPE=P-P.

This program loads the dataset and then produces a plot of each type. The q-q plot is displayed in Figure 15-7.

Figure 15.7. A q-q plot produced by the PPLOT command.

Figure 15-8 displays the p-p plot produced from the program.

Figures 15-7 and 15-8 don't represent all the output you get from the PPLOT command. In particular, a detrended plot (a plot in which the actual values are plotted against deviations of the expected values) is also produced.

Figure 15.8. A p-p plot produced by the PPLOT command.

In this section we look at a program that loads a data file and counts the repetition of values in a certain variable. The program counts repetitions for all cases in a file, splits the file, and takes a count for each portion. The program is named splitfile.sps:

GET FILE='C:Program FilesSPSSIncPASWStatistics18
  SamplesEnglishEmployee data.sav'.
FREQUENCIES SALARY.
SORT CASES BY GENDER.
SPLIT FILE BY GENDER.
FREQUENCIES SALARY.

The first line of the program uses the GET command to load the file. The second line uses the FREQUENCIES command to generate the counts and percentages for the salary values. The top section of the table produced from this command is shown in Figure 15-9. As you can see, the table that the program generated includes five columns. A salary value is shown in the first column; a count of the total number of occurrences of the value appears in the second column. The Percent column holds the percent of the total number of cases (excluding cases with missing values in any variable) that contain this particular salary value. The Valid Percent column holds the percent of the total number of cases (including those with missing values in other variables) that contain this particular value. The Cumulative Percent is the number of cases with salaries less than or equal to the salary shown in the first column. For this example, the values displayed as Percent and Valid Percent are the same because no cases in the displayed portion contain a missing value for any variable.

Figure 15.9. A frequency table for the entire dataset.

Then comes the split. The program that sorts cases uses the SORT command to do the split. You must sort a dataset using the same variable as the key that's about to be used to split a file; variables of like values must be all together for the split to work properly. In this example, the SORT command groups all the female cases before the male cases.

The SPLIT command logically inserts a divider at each point where the value of the named variable changes. In this example, the value f is used for female and the value m for male, so a logical divider is placed between them. The divider is logical because the split refers only to the memory-resident form of the data — the split does not survive the data being written to a file.

The last line of the program builds a new set of counts and percentages, but this time the data is divided by gender, so the table is generated in two parts. The upper part of the table is shown in Figure 15-10. The headings of the table have the same meanings they had before, but you can see that the top of the table contains the numbers for the female cases. The bottom portion contains data from males. If the SPLIT command had used a variable with more values, the cases would have been split into more parts.

Figure 15.10. A portion of the separate frequency tables for females and males.

The EXAMINE command in the Syntax language may be the quickest way to look at data. For example, with the system data file named Cars.sav loaded into SPSS, a two-word Syntax program produces a graph of a variable. The two-word program is as follows:

EXAMINE MPG.

This command results in the box plot shown in Figure 15-11, which graphically displays the mean, the standard deviation, and the extreme values.

Figure 15.11. A box plot of miles per gallon.

But that's not the only way EXAMINE can show you data. You can include more than one variable, or you can change the plot style to a histogram. The following command generates more than one histogram:

EXAMINE ACCEL, HORSE /PLOT=HISTOGRAM.

This command produces a histogram for each of the two named variables. The histogram representing the acceleration values (ACCEL) is shown in Figure 15-12.

Figure 15.12. A histogram of acceleration rates.