Description of Sample at Baseline

1. Describe the sample characteristics and baseline values, comparing the two groups'characteristics.

With any analysis, the first step is to examine your data, assessing the degree of missing data, the potential miscodes, and conducting a descriptive analysis. This typically results in a Table describing the sample. Since this is an intervention study with two groups, you want to describe the characteristics of the 2 groups, rather than the entire sample. The example table shell below gives you an illustration of how this might look. This dataset includes only two demographic characteristics - gender and age. You would also want to consider how the group means differed at baseline on fear of statistics, confidence in coping with statistics and depression. Run the t-tests to determine if there are significant differences in mean age, fear of statistics, confidence in coping, and depression between the two groups at baseline.  You can run the t-test procedure and get the descriptive statistics at the same time.  Make sure that you assess the test assumptions.  Add the statistical test results to the last column of the table.

Write a paragraph summarizing findings; comment on meeting or not meeting test assumptions.   See "Presenting the Results" on p. 243 in Pallant for an example of how to report testing results.

Table 1  ....title table so that it can "stand-alone" if separated from the paper (e.g., include description of sample, sample size, and what the table is describing).

Part II:   Hypothesis testing of mean differences

2.   Single group pretest-posttest

For the entire sample (regardless of the intervention), test the hypothesis below, making sure that you test the assumptions for the statistical test. Write a paragraph describing these results, reporting on the assumption testing, the descriptive statistic (means, SD), and the statistical test. See "Presenting the Results".

Hypothesis:  Having either of the classes will improve the mean confidence in coping with statistics (i.e., compare Time 1 to Time 2).

3.   Two group pretest-posttest

Assume that the researcher had the following two hypotheses to test.

Hypothesis:  The "confidence building" groupwill have a higher increase in confidence in coping with statistics than the "math skills" group.

Hypothesis: The "math skills" group will have higher scores on the statistics exam than the "confidence building group.

A simple approach to testing these hypotheses is to calculate change scores and compare the mean change scores for the two groups.  To compute this score in SPSS:  Transform/Compute - enter a name for the new variable (e.g., confide_chg), then indicate which scores should be subtracted (confid2 - confid1), hit OK.  The variable should be created - check the frequency.  Conduct your statistical test of the difference in the means, making sure that you test assumptions.  Write a paragraph describing the results.

4.  More than 2 groups

The t-test cannot be used to test differences across more than two groups.  Assume that the researcher wants to examine baseline fear of statistics in relation to exam performance:

Hypothesis: Performance on the statistics exam varies by baseline fear of statistics.

One approach is to create an ordinal variable from the interval-level fear of statistics variable (e.g, Low-Medium-High) and then compare the mean exam scores by the three groups.  Do the appropriate statistical tests and summarize in a paragraph.  Again, make sure that you test the assumptions.See "Presenting the Results".

How to create a three-level variable from the baseline fear of statistics:

-Transform/recode into different variable - click over baseline fear variable, add a name for the output variable e.g., Base_fear_re3
-Click the "Old and New Values" button and enter the data based on these values: Low fear= 30-38, Medium fear=39-42, High fear=43+
-Don't forget that you must click the "Change" button before the OK can be clicked.
-Do a frequency to check that you created the new variable correctly.