SPSS

An SPSS program is a set of syntax statements, or "SPSS language," for executing a particular procedure, data transformation, or statistical formula.

Like many software packages, SPSS is menu-driven. That means users can execute statistical analyses, simple or complex, by clicking a series of pull-down menus and selecting the desired pre-programmed commands. However, many researchers or analysts may use certain statistical procedures that are not pre-programmed into SPSS. As a result, SPSS allows users to create custom programs, or to link together multiple pre-programmed operations to be applied in sequence.

An SPSS program enables a user to conduct the same procedure repeatedly, without having to remember which pull-down menus or commands to click and choose in order to set up the needed series of procedures. That saves time when organizing and analyzing data. Those programs also can be modified to run different statistical models, examine different variables, or access different data files.

To execute a program, just click on the syntax and drag to highlight it. After that, click the "execute" command, an arrow-shaped key in the syntax file menu.

Although an SPSS program appears complex, you do not have to be a master programmer or be fluent in SPSS syntax to use a program. It's not necessary to be able to write syntax, either, although a person skilled in SPSS syntax could write a program. To save a procedure as an SPSS program for future use, all you have to do after selecting the appropriate commands is to click the "Paste" command in SPSS instead of the "OK" command. The "Paste" command saves the syntax into a separate file, which you can save and modify for future use.

Users can create a wide variety of SPSS programs to meet their data management and analysis needs. They can save programs of SPSS syntax for exporting data from another source, such as an Excel spreadsheet. Other programs that can be saved as syntax include procedures for recoding variables, merging data files, or calculating values. Finally, users can save SPSS programs for the statistical procedures of which SPSS is capable. Those range from simple descriptive statistics to multivariate regressions.

SPSS stands for Statistical Package for the Social Sciences. When SPSS was first established in 1968, SPSS as a company and SPSS as a software package were one and the same. Since then, SPSS as a company has branched out into other areas of statistics-related software. Over time, the company came to be known as SPSS, Inc. to differentiate itself from its original product, which was known simply as SPSS, or SPSS software.

SPSS software was designed to perform statistical analysis on quantitative data. In plain English, SPSS software is used for complex calculations to analyze numerical data. SPSS software is used in nonprofit agencies, educational institutions and even in business to analyze numerical data. It performs functions such as regression, which is a form of predictive calculation used to determine the relative effect of a single factor on a situation.

In addition to the original SPSS, SPSS Inc. offers products in four major groupings: statistics, modeling, data collection and deployment. The statistical group of products includes SPSS. The modeling group is the name SPSS Inc. has assigned to data mining software, formerly called Clementine. The Data Collection group is the assigned name for SPSS, Inc. products designed to analyze surveys and other qualitative, or word-based data. The Deployment group is designed to bridge the gap between analysis and making active use of the data.

In 2009, SPSS, Inc. was acquired by IBM, and underwent the process of renaming several of its software programs. The new umbrella name for the entire portfolio of products offered by SPSS, Inc. is PASW, which stands for Predictive Analytics Software. The name change was designed to emphasize the position of SPSS, Inc. as a prominent player in predictive analysis -- software designed to assist the user in predicting future action or circumstances based on collected data.

The software packages offered by SPSS, Inc. include its survey analysis software, which is rivaled by NVivo and Ethnograph, among others. SAS (Statistical Analysis System) is a major competitor to SPSS software for statistical analysis. R is an open source software language and environment used for statistical computing and graphics. Nonetheless, no one company offers as wide an array of predictive software as does SPSS, Inc.0

Factor analysis is a data reduction and analysis technique that examines the extent to which underlying, unmeasured issues, or factors, influence a set of related measures, or variables. The technique is especially popular in survey research. Factor analysis requires two stages: factor extraction and factor rotation, both of which SPSS can conduct.

SPSS provides a user-friendly method for conducting factor analysis. To begin a factor analysis, you click the "Analyze" command in the top task bar of an SPSS spreadsheet. This opens a drop-down menu with a set of analysis options. Factor analysis is under the "Data Reduction" option. Selecting "Factor" will open the factor analysis box, enabling you to begin factor extraction. This step helps determine the number of underlying factors influencing the variables in your data set. This same box in SPSS can then rotate the extracted factors, examining patterns of correlation between factors and measured variables, such as individual items in a survey.

SPSS helps you decide the number of underlying factors during the factor extraction stage by retaining all factors with eigenvalues greater than one. An eigenvalue measures the amount of variance. For example, an analysis of influences on registered voters' political ideologies that revealed three eigenvalues greater than one would have three underlying factors.

You need not be a master computer programmer to conduct factor analysis with SPSS. The program's "Paste" command allows you to save a file of complex syntax commands, which enables you to run the same analysis without having to click the same commands again. With a saved syntax file, you can conduct the same factor analysis on different sets of data with only minor changes to the syntax, such as changing the name of the data file to analyze.

SPSS released its first program in 1968 after being developed by Norman Nie, a postgraduate in political science at Stanford University. Since then, SPSS has grown in popularity among researchers in the social sciences, public health, education, market research and more. In 2009, IBM acquired SPSS for more than $1 billion.

SPSS stands for Statistical Package for the Social Sciences, and it organizes quantitative research data into various statistical formats to determine the relevance of variables associated with the research topic. Developed primarily for researchers in the field of social science in the 1960s, SPSS has since been adapted by practically every profession commonly using research and quantitative data analysis, and it frequently serves as the standard data analysis software for undergraduate and graduate statistics students. In 2008, SPSS Inc. changed the SPSS Statistics brand to PASW, Predictive Analytics SoftWare, Statistics.

In quantitative research, the researcher must rely on the relevance of the dependent and independent variables to each other and to the population targeted by the research topic. SPSS/PASW helps researchers arrive at statistical relevance with efficiency and accuracy, and many scholars and researchers consider it a critical tool.

Few people outside academia ever encounter, use or even hear of SPSS, but it can run descriptive statistics, t-Tests, Chi-Square and other non-parametric tests, as well as linear regression models and analysis of variance, among other statistical analysis tests and procedures.

IBM recommends the SPSS/PASW Student Version for students starting their studies in statistics. IBM licenses the student version for 13 months, after which the program will cease to work on a particular machine. This version allows students to run data analysis for up to 50 variables and 1,500 cases. This version retails for about $100 in 2010.

IBM also offers two versions of SPSS/PASW for graduate students with four-year licenses. The GradPack Base includes the full version of SPSS Base. The full GradPack includes the SPSS Base plus SPSS Advanced Statistics, IBM SPSS Regression and Amos software. Amos (Analysis of Moment Structures) allows students to perform structural equation modeling. These versions retail for about $200 in 2010.

College students should consider learning SPSS/PASW or a similar statistical program. Many companies use SPSS to track and project trends like sales figures and polling data. Advanced academic quantitative studies also require statistical software packages to analyze data.

For statisticians and analysts in fields ranging from education and health care to business and public policy, SPSS 16 is a valuable tool for analyzing large sets of data and answering critical research questions. SPSS, a software company acquired by IBM, released SPSS 16.0 in 2008. Like most spreadsheets and statistical software, SPSS 16.0 can calculate basic descriptive statistics, such as frequency distributions, means, medians and standard deviations. Its Tables function enables you to display summarized data in rows and columns. For users with statistical knowledge and whose research questions require more rigorous analysis, SPSS' capabilities include a range of complex statistical procedures.

A favorite statistical method among economists and other social science researchers, linear regression analyzes the change in an effect, or dependent variable, resulting from changes in one or more predictors, or independent variables. An economist, for example, may study the extent to which additional years of education or work experience changes the average salary received by professionals in a particular industry. SPSS 16 offers the regression option in its menu of analytical procedures. To conduct a regression analysis, click the Analyze pull-down menu at the top of an SPSS spreadsheet, then select "Regression." This opens a regression window that allows you to enter the dependent and independent variables you want analyzed. SPSS can display results in tabular form and with graphics.

Logistic and probit regression procedures involve dependent variables or effects that are binary. Whether a person graduated high school or whether someone is male or female are examples of binary dependent variables. The Regression option under the "Analyze" menu in SPSS 16 enables you to conduct both types of regression analysis. Choosing the regression option under the Analyze menu will give a range of regression options, including logistic and probit, as well as linear regression.

Popular in survey and marketing research, factor analysis reduces a large set of data to identify underlying factors and how they influence a set of related measures. This procedure is beyond the scope of many spreadsheet programs, such as Microsoft Excel. SPSS 16.0, however, enables you to conduct factor analysis by clicking the "Analyze" pull-down menu and selecting "Data Reduction," then choosing "Factor Analysis." This opens a window that allows you to select the measures to analyze.

SPSS is statistical software produced by IBM. The software integrates with Microsoft Office applications, such as Excel. There are several different versions available depending on your needs. Choose from the standard, professional, premium or modules. While IBM releases a new version of SPSS every year, the actual differences between the programs each year are few. SPSS 15 was released five years after SPSS 10 - so there are several major differences between the two programs.

The chart builder is an interface that allows for chart creation. This interface was added in SPSS 14 and is not found in SPSS 10. The chart builder in SPSS 15 has been further expanded with new chart types and PDF export capabilities. The new design of chart builder provides more efficiency and tools to create polished and professional charts quickly.

With SPSS 15, it is now possible to use SPSS with Dimensions products. For example, you can import data that was collected for a survey in the Dimensions data model. The export to database wizard allows the user to create tables in SPSS 15 and export this table to a new database, or into an existing database.

Some of the new statistical analysis found in SPSS 15 include PLUM ordinal regression, which was previously only available from the advanced add-on module. Use PLUM to determine the ordinal outcome using data that contains more than two categories.

The programmability extension comes with SPSS 15. This extension allows the user to extend the syntax language used for SPSS commands. You can write program code in other programs, such as Python, and use this to control SPSS syntax. Create user defined procedures and use syntax for opening and saving new files. You can even save the syntax to use within the automated production facility for commands that are used frequently.

SPSS allows researchers to transform variables by computing new variables. For example, if you had a series of scores on subscales of a measurement, SPSS would let you add them together to get a total score. You simply name the new variable (for example, "Total") and then direct the program to add the variables you select together. You do this by clicking on "Transform" and then clicking "Compute" from the drop-down menu and entering the variables with a plus sign between them.

SPSS allows researchers to transform variables by multiplying them together. For example, if you had a series of weekly paychecks for a group of people and the number of weeks each person worked, SPSS would let you multiply them together to get a total payment for each person. You simply name the new variable (for example, "Totalpay") and then direct the program to multiply the variables you select together. You do this by clicking on "Transform" and then clicking "Compute" from the drop-down menu and entering the variables with a multiplication sign (x) between them.

Social science researchers use recoding when they are scoring measurements. For example, if a psychologist collects data from a measure of internal-external locus of control, some of the questions will have "yes" responses to indicate internal control and some will have "no" responses to indicate internal control. Locus of control measures whether people feel they control what happens to them or outside forces, such as luck or chance determine what happens to them. Once you enter data into the SPSS program, you need to recode the variables in the same direction so they can be totaled. Recoding involves changing the "yes" responses (coded as 1) to "no" responses (coded as 2). You can make this change by clicking on "Transform" and when the menu drops down clicking "Recode." You move the items to be recoded into the variable list and then indicate how they are to be recoded (for example, put "1" under "Old Value" and "2" under "New Value").

Once you have recoded the variables in a measure, you obtain the total score. You do this by adding the individual items together by using the "compute" key as if you were computing variables by adding.

T-Tests compare the means of scores from the same group of people at two points in time --- for example, students' scores on their mid-term exam and their scores on their final exam. T-Tests can also compare the scores from two different groups --- for example, the final exam scores from male students and those of female students. You can use SPSS to conduct a T-Test. SPSS provides the output which indicates the T value (the test statistic) and the significance (the likelihood the results are due to chance rather than the independent variable) for the comparison between the means. A significance level of less than .05 indicates that the likelihood the results are due to chance are less than 5 out of 100.

A One-Way ANOVA (Analysis of Variance) is used to compare means when there is one independent variable and one dependent variable, similar to the T-Test. SPSS analyzes data for a One-Way ANOVA. The results are very similar to the results of a T-test, with the same significance level.

A Two-Way ANOVA (Analysis of Variance) is used to compare means when there are two independent variables and one dependent variable --- for example, if the researcher wants to compare the scores of males and females in private and public school on their SATs. SPSS analyzes data for a Two-Way ANOVA or a Three-Way ANOVA. A Three-Way AVOVA is used to compare means when there are three independent variables and one dependent variable --- for example if the researcher wants to compare the scores of males and females in private and public school from two different states. As you increase the number of independent variables, it becomes more difficult to interpret the results.

A MANOVA (Multiple Analysis of Variance) compares the means of one or more classes of independent variables --- for example, gender and type of school --- on two or more dependent variables --- for example, SAT scores and grade point average. As the number of independent and dependent variables increases, interpretation becomes more difficult.

Cluster analysis with the Statistical Package for the Social Sciences start with a set of cases or participants that you want to categorize into groups of similar cases or clusters. Start by selecting which variables are most important to use to identify the groups. The Statistical Package for the Social Sciences provides three distinct methods for cluster analysis: k-means, hierarchical, and two-step. The procedure you use depends on the size of your data. Use the SPSS hierarchical clustering procedure for small data sets and to find solutions with varying numbers of clusters. There are two main types of hierarchical clustering procedures: agglomerative and divisive.

Agglomerative hierarchical clustering starts by putting each case or participant into a separate cluster. The procedure then goes through steps in which similar cases are combined into successively larger clusters. Once a case has been put in a cluster, it cannot be separated from the other members of that cluster. Clusters are allowed to merge with other clusters, but cannot be split. The final step puts all the cases into one cluster. Neither the first step -- each case is in a separate cluster -- nor the last step -- all cases in the same cluster -- is an appropriate solution.

Divisive hierarchical clustering proceeds in the opposite manner of agglomerative hierarchical clustering by placing all cases in one large cluster. The method proceeds in steps to form smaller and smaller clusters until each case is in an individual cluster. The SPSS procedure displays the clusters and the cases included in them at each step. Neither one large cluster nor each case as a separate cluster is an appropriate solution.

Selecting the best solution to clustering procedures is an art as much as a science. There is no correct or incorrect appropriate number of clusters. You determine the appropriate number of clusters based on how you intend to use the clusters. Review the clusters solutions at each successive step and choose the solution in which the clusters can be interpreted in the most logical way. You do this by reviewing the successive solutions to determine which solution contains an appropriate number of clusters that appear to contain logically similar cases.

One problem commonly associated with searching through catalogues is the inability of the user to find all instantiations of a particular work. Hierarchical catalogues start with a category at the top that encompasses all information associated with that topic. Users can then navigate their way through lower categories in order to zero in on the particular piece of information they want.

Consider how a person would search through a hierarchical catalogue at a library. If he is looking for a 2000 edition paperback copy of "The Iliad" that has been translated by Ian Johnston, he can start with the abstract work titled "The Iliad." He can then access all translations by Ian Johnston. He can then access the 2000 paperback addition and finally, go find it on the shelf.

While searching and browsing through catalogues is made much easier with the implementation of a hierarchical structure, there are some drawbacks. Its main drawback is its complexity. In order to deal with their complexity, catalogue builders must include complex matching algorithms that may slow down the system considerably. Ironically, if these drawbacks cannot be rectified, the searching and browsing benefits conferred on users by the hierarchical structure may be counteracted by the system's decreased performance.

Due to the complexity of the set-up process, many catalogue builders may be put off by hierarchical catalogues. However, once the system is properly set up, users can expect a much easier and much more comprehensive catalogue searching experience. Maintainers of catalogues will also find their job has become much easier. Finally, research projects can be expected to improve in general as researchers are much less likely to miss important information searching through hierarchical catalogues.