47-512  Applied Research Methods

SPSS Introduction:  Descriptive Statistics

due in class 03/03/03

Note: due to illness I have not gotten this material onto the web until much later than anticipated.

Please accept my apologies for the inconvenience and

 just do the best you can to get the assignment done.

 

Purpose. 

• Increase facility with data exploration using SPSS and data sets provided in class. 
• Use both graphs and statistics to understand trends.
• Understand basics of statistical comparisons

 

How to start.  Copy the data from the disk passed our in class to the hard drive of the computer on which you are working.  Open SPSS.  From the menu, select file, open, data and open the World95.sav data file you just copied onto your hard drive.  Remember if you click down on the tabs on the lower left of the screen, you can switch between screens that permit you to store information about your variables  (variable view) and to make adjustments (i.e., change the number of decimal places showing from 2 to 0 when you don’t need them as in this exercise).  Variable names need to have 8 characters or less; longer and more descriptive titles can be used as labels.  These are assigned by typing into the appropriate spaces on the variable view screen.

 

Remember:  Each participant gets one row, and one row only.  Each variable gets one column.   What are your sampling units here? This is an example of a “participant” that is not an individual.

 

In this exercise, you are being asked to compare information about literacy rates (a) across different types of countries, and (b) within countries across gender.

 

What to do?  Always look at your data first.  Look at the distribution of literacy, lit_male, and lit_fema by going to GRAPHS and choosing histogram.  For another view, go to GRAPHS, choose Boxplot and create a simple boxplot for separate variables. 

 

Now begin to compare.  Choose a simple boxplot for groups of cases and compare the variable literacy using the category axis region.

 

Forgotten what these plots tell you?  Go to HELP and open the topic. Opening boxplot, for example, allows you to be reminded that, “Boxplots show the median, interquartile range, outliers, and extreme cases of individual variables.” 

 

QUESTION 1  Is the literacy rate in the Middle East different from the literacy rate in Latin America?  Now you want to limit your data to examine only those two regions. 

 

• How are these two regions coded?  Go to the variable view (bottom tab) for your data and you will see that Region is coded by numbers and you can select the … next to Values to see what they are.  You find Middle East is 5 and Latin America is 6.
• Go to the data view for your data and to the menu selecting DataàSelect CasesàIf condition is satisfied. 
• Click on Region in the list of variables and then the arrow to the right of that  list.  The variable name should appear in the formula box.  You want to include countries where the region is 5 or 6 (i.e., greater than 4), so now type in > 4 next to the variable name in the formula box.  Click continue and okay.
• Check what has been temporarily included and excluded.
• Now you want to create confidence intervals around each average literacy rate. 
• Go to Analyze à Explore
• You want to get statistics with a 95% confidence interval for the dependent variable percent of people who read (literacy) factored by the region.  Click on the statistics box at the bottom to be sure you will get your CI.

 

ANSWER 1  Are the literacy rates in these two regions different?  Say yes or no and give the statistical evidence you just obtained.

 

QUESTION 2  Are the male and female literacy rates the same or different within countries. SO, what you really want to know is whether the answer will be 0 (or whether the 95%CI for the answer will include 0) when you subtract the female rate from the male rate.  Let’s do it.

 

• Back to the data and the menu to get all cases included again Data à Select cases à All.
• Now you want to get the differences for each country in male and female literacy rates.  Go to Transform à Compute
• Your new variable can be diff_mf, which will remind you that you computed the difference by taking male and subtracting female.  So type diff_mf into the box labeled target variable.  Now scroll down the variable names below and select males who read (%) and the arrow next to that box to move lit_male into the Numeric Expression box.  Then click in the Equation Box and  type in the minus sign (-).  Then scroll and select lit_fema and move it into the equation so that it now reade lit_male – lit_fema.  Select Okay.
• Now go back to the data and notice that you have added a variable column.
• What does it look like?  Create a simple boxplot of this separate variable. Is it sort of normal?

• Create a confidence interval around this difference score by going to Analyze à Explore and proceeding as you did above.

 

ANSWER 2  Are the literacy rates for males and females within country significantly different from each other?  Say yes or no and give the statistical evidence you just obtained.