Thursday, July 24, 2008

SPSS for Dummies by Arthur Griffith John Wiley & Sons © 2007

Here is another book on SPSS which is available in the Books24x7 collection.

SPSS for Dummies
by Arthur Griffith
John Wiley & Sons © 2007 (360 pages)
ISBN:9780470113448

Covering all of the key analysis topics step-by-step, this friendly, plain-English guide gets you up and running with the latest version of the software so that you can start performing calculations right away.

I have picked an excerpt from Chapter 14 on some examples of analysis which I believe might be useful for new researchers:

Independent-samples T test

The independent-samples T test compares the means of two sets of values from one variable.

Click picture to enlarge

The table displays the two means and the standard deviation and standard error for the two means. The Independent Samples Test table provides further information about the mean in two rows of numbers — one for equal variances and one for unequal variances:

* If the significance of the Levene test, the number in the second column, is high (greater than 0.05 or so), the values in the first row are applicable.

* If the significance of the Levene test is low, the numbers in the second row are more applicable.

* If the significance of the T test, the 2-tailed significance, is low, this indicates a significant difference in the two means.

* If none of the numbers of the 95% confidence interval are 0, it indicates the difference is significant.

One-way ANOVA

Click picture to enlarge

ANOVA is an analysis of variance. A one-way ANOVA is the analysis of the variance of the values (of a dependent variable) by comparing them against another set of values (the independent variable). It is a test of the hypothesis that the mean of the tested variable is equal to that of the factor.

Correlation

The group of tests in this section determines the similarity or difference in the way two variables change in value from one case (row) to another through the data.

Click picture to enlarge

Correlation figures vary from −1 to +1, and the larger the value, the stronger the correlation. In Figure above, you can see that the variables have a correlation of 1 with themselves and .880 with one another, which is a significant correlation.

Regression

Regression analysis is about predicting the future (the unknown) based on data collected from the past (the known). A regression analysis determines the mathematical equation to be used to figure out what will happen, within a certain range of probability. It analyzes one variable, the dependent variable, taking into consideration the effect on it by one or more factors, the independent variables. The analysis determines that some independent variables have more effect than others, so their weights must be taken into account when they are the basis of a prediction. Regression analysis, therefore, is the process of looking for predictors and determining how well they predict.

Click picture to enlarge

When only one independent variable is taken into account, it's called a simple regression. If you use more than one independent variable, it's called multiple regression.

No comments:

Post a Comment