Chapter 15

Epilogue:
SPSS, Data Analysis and the EZ Research Project

15.1 Computer-Based Analyses of Research Data

It seems appropriate to conclude with a review of what we have learned in this book about the use of the computer – and the SPSS statistical package in particular – in the analysis of data obtained in a research project such as the hypothetical study at EZ Manufacturing. This text has introduced you to an extremely powerful tool and has helped you develop the skills needed to use it in analyzing data.

In the first several chapters, you learned learned the basics for using the Syntax Code method and the Point-and-Click method in using SPSS for Windows. You learned how to create a data file consisting of scores from on several distinct variables related to a research question. You learned how to edit that file and to obtain a printout of the file. You also learned how to generate simple frequency distributions of the scores on each variable in the file, as well as how to transform existing variables into newly-created variables. We saw that this procedure could be helpful in getting an overall feel for the data, but that drawing meaningful conclusions from this cursory description of the data was next to impossible.

We then proceeded to the statistical analysis of the data so that we could obtain a deeper understanding of the results. We introduced you to the wide range of descriptive and inferential statistics that could be computed using SPSS on our data file. In Chapter 6, you learned how to generate a variety of descriptive statistics, such as the mean, median, mode, range, and standard deviation. This number crunching SPSS routine saves the researcher hours of tedious labor that would be required to compute these statistics by hand. It's also much more accurate. These statistics can be very useful in helping the researcher get to know the data by obtaining single numbers that summarize the entire set of scores for each variable. They are also essential in making comparisons between important groups in the study (e.g., comparing mean leadership performance scores of men versus women).

Chapter 7 introduced you to a routine for creating frequency matrices and contingency tables for scores in various combinations of levels of two or more nominal variables (e.g., the number of task oriented male versus female employees). In Chapter 8 you learned various correlational subroutines that are useful in describing direction and degree of relationships between two quantitative variables (e.g., the correlation between Achievement Needs and Task Skills), and in predicting scores on one variable from another variable.

The above statistical analyses took us a long way toward understanding our data and the relationships between variables relevant to our basic research questions. However, in order to determine whether there were real differences among the various groups being studied, we had to rely on several inferential statistics that permit such conclusions. In Chapters 9 and 10 we learned how to use the t-test procedure to analyze differences between mean scores of groups on the same dependent variable (e.g., Social Skills scores at two levels of a single independent variable (e.g., Gender).

In Chapters 11 and 12, a more sophisticated procedure was introduced (One-way ANOVA) which allowed the researcher to test for significant differences between three or more groups (e.g., mean Leadership Performance scores of masculine sex typed, feminine sex typed, and androgynous employees). We concluded in Chapters 13 and 14 with an extremely powerful procedure (Factorial ANOVA) which permits simultaneous testing for differences between levels of two independent variables (e.g., Gender and Leadership Style) in addition to assessing their interactive, or combined, effects on the dependent variable.

We assume this introduction has given you an appreciation for the value of the computer in helping the investigator answer both simple and complex research questions. The computer not only saves time and effort, but it can enable the researcher to pose and answer questions that would be difficult, if not impossible, to address by eyeballing the data or by hand computations. We believe that the basic skills you have learned in conducting the above analyses on our hypothetical research project will be easily transferable to future projects that you may undertake. Indeed, you should have acquired a better understanding of statistics and research methods from the examples and exercises, and we hope that you will be motivated to apply this knowledge and skill to whatever career you pursue.

This, of course, brings us back to one of the motivating factors for the EZ project in the first place – the mega consulting fee that upper management at EZ Manufacturing is paying you for conducting this study! What are you going to be able to report to them that might be helpful in the development of their affirmative action program? Let's consider some of the conclusions and recommendations that you might make on the basis of your research and data analysis.