08.01.2015 Views

A Step by Step Guide for SPSS and Exercise Studies

A Step by Step Guide for SPSS and Exercise Studies

A Step by Step Guide for SPSS and Exercise Studies

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Statistical tests 83<br />

Figure 24<br />

residuals against the predicted values (to save these values in the data file,<br />

use the Save option below). If this assumption is met, there should be no<br />

specific pattern in the spread of the residuals. However, in Figure 25, there<br />

seems to be a pattern as the variability of the residuals decreases with<br />

increasing predicted values (i.e., heteroscedasticity). Usually, the literature<br />

reports cases where the variability of the residuals increases at larger<br />

predicted values. An alternative way to test this assumption is to re-run the<br />

Univariate analysis using the residuals as the dependent variable. A nonsignificant<br />

Levene test will signify that the assumption of homogeneous<br />

variances cannot be rejected. Heteroscedasticity reduces the power of<br />

ANOVA to discover significant results (i.e., Type II error).<br />

3. The relationship between the residuals <strong>and</strong> the predicted values of the<br />

dependent variable should be linear. However, inspection of Figure 25<br />

reveals that this relationship is non-linear. Non-linearity also reduces the<br />

power of ANOVA.<br />

4. The residuals are independent. This assumption requires that the score of<br />

one participant is not related in any way to the score of another participant.<br />

This assumption is violated when the order in which participants are<br />

assessed may influence their per<strong>for</strong>mance. For example, imagine an<br />

inexperienced researcher who is able to provide better instructions to the<br />

participants who join his/her experiment later on rather than earlier. Better

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!