Analyzing Quantitative Data

Short-cut to Examples

Questionnaire Example.
Compiling Data With Excel Example.
Compiling Data With SPSS Example.
Cross-tabulation Example.
Comparing Means Example
Correlation analysis Example.
Using SPSS Example
Introduction

In terms of argumentation, the purposes of research are:

• To find out what kind of argument you want to make (find an answer to your research question); and
• To garner evicence to support the argument you will make (test the value of your answer by trying to convince others that it is true.).
Remember, the essence of research is comparison - especially quantitative research.  What difference does a treatment make in outcomes?  Do two variables increase or decrease together?  Are members of one group more likely than members of another group to experience a certain outcome?  For example, are men more willing to speak in class than women?  Does willingness to speak in class correlate with age or GPA?  Are women more likely than men to major in Speech Communication?

Steps

1.  State your research questions in a clear, precise way.  If your main question(s) imply other questions, state these as sub-questions.

2.  Determine what kinds of comparisons you need to make in order to answer your questions.

• What variables do you need to measure?
• Are you interested in differences between groups?  Correlations between variables?

3.  Decide how to measure each variable, and determine what kind of variable it is (e.g., nominal, ordinal, interval, or ratio).
This will help you decide what kind of comparisons you will want to make.

4.  Set up blank tables.

5.  Devise a research method and gather your data.

6.  Summarize your data in such a way that you can fill in the blanks in your tables. (Analyze your data.)

7.  Interpret your data:  What is the answer to each of your questions?  What kind of argument will you make?

Example

I assign group projects in many of my courses.  I have noticed that most students have very good experiences with their group projects, but a few do not.  I have noticed the same mixture of results with group projects in workplace settings.  These experiences led me to wonder about the differences between people and between groups that might help account for the different experiences people have with group projects.  With that as my overall research question (and mindful of the need to provide an example of each kind of data analysis I want to illustrate here) I posed the following specific research questions:

RQ1:  Do men and women differ in the way they tend to organize their groups?  Here, I decided I am most interested in whether people divide the overall project into discrete tasks and work independently on each task, work together on all phases of a project, or use a mixture of these, working together on some tasks and working separately on others.
RQ2:  Do different ways of organizing groups lead to different assessments of the outcome?  Here I decided to measure how positively people feel about the social aspects of the experience, how positively they feel about the learning process, and how highly they rate the final product.
RQ3:  Independently of how groups are organized, is age related to how people evaluate the results of group projects?  E.g., do older, as people grow older and gain life experiences, do they get better at working in groups?

For this example, I used a short survey questionnaire to gather data. To see the questions I used, with a brief discussion of the questionnaire, click here.

Questionnaire Example.
I compiled the data on an Excel spreadsheet; I also compiled the data on SPSS. To see a detailed discussion of how to compile data, click on one of the following:
Compiling Data With Excel Example.
Compiling Data With SPSS Example.
RQ1:  Gender is a nominal variable.  It would be possible to measure how groups are organized in such a way that it would be ordinal or even ratio-level, but the simplest measure I can think of simply asks people to classify themselves into one of three groups - producing another nominal variable.  For this question, I decided to produce a simple cross-tabulation, Gender by Group Organization.  To see how I set the table up, filled it with data, and interpreted the data click here:
Cross-tabulation Example.
RQ2:  I decided to create three indexes to measure how people feel about the process and outcome of group projects.  I created a series of statements about outcomes and asked subjects to respond on a five-point Likert-type scale, with responses ranging from "Disagree Strongly" to "Agree Strongly," then I averaged the results of thematically-related scales to produce te indexes.  Although the results of a Likert-type scale are actually ordinal, it is conventional to treat them as interval-level data, and it can be shown that analysing the data as interval-level produces results similar to analyses at the ordinal level.  Thus, for this question I decided to compare means, Group Outcomes Evaluations by Group Organization.  To see how I set the table up, filled it with data, and interpreted the data click here:
Comparing Means Example.
RQ3: Since age is a ratio-level variable, and I have decided to treat the indexes of outcome evaluations as interval-level variables, the obvious analysis for RQ3 is Pearson correlation.   (If I treated the evaluations as ordinal data, I would use Spearman Rank-Order Correlations; the outcomes tend to be very similar in magnitude with data of this type.) To answer this question, I decided to correlate Age with Group Outcomes Evaluations. To see how I set the table up, filled it with data, and interpreted the data click here:
Correlation Analysis Example.
For simplicity, I have set these examples up using Excel - most other spreadsheet softwares will perform similarly.  However, students who expect to do much research will find it useful to know how to use a more versatile and more powerful data analysis software.  Accordingly, I have also prepared a detailed set of examples to guide you through the use of SPSS for Windows.  (This is entirely optional - most of you will find Excel entirely sufficient for your needs in this class.)
Using SPSS Example.
A final note:  I have not mentioned an additional advantage of setting up your tables before you gather data.  Two advantages, actually:  If you have your tables set up in advance, you are less likely to gather more data than you need.  The second advantage -  setting your tables up in advance helps ensure that you will gather as much data as you need!