Readers are advised that this appendix is intended primarily for those who can interpret the results of a basic statistical analysis. If you are unfamiliar with statistics, or uncomfortable with it, we gently suggest that you might not find these appendices to be relevant to you.

Note: SAS^® is a registered trademark of SAS Institute Inc., Cary, North Carolina [U.S.].

This appendix provides results for contingency tables done for the thesis. The output below is labeled, however some readers may find parts of it to be hard to understand. The notes below provide a brief overview of the most important elements in reading the tables (overview only: not intended to be comprehensive).

These have a rectangular cell structure, with cells arranged in rows and columns. The rows define one categorical variable and its values, the columns the other categorical variable. The term categorical variable refers to discrete variables, e.g., a survey question that has 2 replies: "yes" or "no." This is to distinguish it from a non-categorical ("continuous") variable with a range of values in interval(s), e.g. a temperature reading that usually falls between 0 and 100 degrees Centigrade.

In the contingency tables, rows and columns are labeled with the identity of the associated variables, which come from the survey done for this thesis. For example, you will see below columns labeled "Pre-test" and Post-test," the two study periods for the survey. Rows in each table are labeled with the value for the requisite variable, say for a specific example "Yes" and "No." Thus the cell that falls in the intersection of column number c, and row number r, gives the statistics for those responses where we have (row, column) = (r,c). For a more demonstrative example, in the Pre-Test column, if we consider a table where the first row is labeled "No," then the statistics in that cell reflect the people in the pre-test survey who answered "No" to the question that defines the rows.

Each of the main cells in the contingency tables list 3 numbers:

• Frequency = raw number of responses for cell (can be zero if no responses in that cell).

• Percent = cell count as percentage of table total = (cell frequency)/(table total count).

• Col Pct = column percentage = (cell frequency)/(total count for column cell is in).

The SAS system can compute additional cell statistics, but we have limited it to the most relevant ones for the tables here. If you look at the last row/last column in each table, you will see row and column subtotals with their percentages, as well as the overall table totals in the lower right corner of each table.

Two of the chi-square statistics are of interest: the one simply labeled "Chi-Square," a better designation for which is the goodness-of-fit chi-square, and the "Likelihood Ratio Chi-Square." These statistics are usually close in value to each other. (They are asymptotically equivalent, which means they converge to the same value as the sample size increases without bound.)

The SAS tables include a number of other versions of chi-square statistics, Fisher's exact test (only for 2X2 tables), and a number of other test statistics. The thesis did not use these other statistics; if you understand them you might find them of interest. Otherwise, you can safely ignore them here if you wish (in the context of interpreting the thesis).

If you still have difficulty understanding these tables, or have questions regarding statistics in the table that are not covered above, please consult an intermediate statistics text, and /or college-level texts on contingency tables [categorical data analysis] for clarification. Please note that Beyond Veg authors/editors cannot provide statistical tutoring via email.