Edit Option of table_tests=[$r region]

TABLE_TESTS=[$R col_list BY row_list] is an edit option. (***NO equal sign between $R and the regions.***)

If you want to use the region (or portion of a table) that is made up of the banner points two through five and rows one through 10 and you want to testor statistical significance using an ANOVA, you can add this to the LOCAL_EDIT statement in the TABLE_SET that describes the row:

LOCAL_EDIT={: ANOVA=[$R  2 TO 5 BY 1 TO 10] }

If you want to use the same banner points on a different table and use rows one through five and you want to test for statistical significance using aCHI-SQUARE, you can add this to the LOCAL_EDIT statement in the TABLE_SET that describes the row:

LOCAL_EDIT={: CHI_SQUARE=[$R  2 TO 5 BY 1 TO 4] }

If there are several regions in the banner that you want tested for statistical significance using a CHI-SQUARE, your LOCAL_EDIT statement mightlook like this:

LOCAL_EDIT={: CHI_SQUARE=[$R  2 TO 5  BY 1 TO 4]CHI_SQUARE=[$R  6 TO 8  BY 1 TO 4]CHI_SQUARE=[$R  9 TO 11 BY 1 TO 4]CHI_SQUARE=[$R 12 TO 15 BY 1 TO 4] }

If you want to do either an ANOVA or a CHI-SQUARE on every table where one of these two tests is appropriate, you can add this to the TABLE_SET thatdescribes the banner:

EDIT={: TABLE_TESTS=[$R  2 TO 5  BY 1 TO LAST]TABLE_TESTS=[$R  6 TO 8  BY 1 TO LAST]TABLE_TESTS=[$R  9 TO 11 BY 1 TO LAST]TABLE_TESTS=[$R 12 TO 15 BY 1 TO LAST] }

Use “1 TO LAST” to describe the rows you want tested. It is not necessary to know the number of rows on each table. If specific rows need to be excluded from the statistics (i.e., DON’T KNOW), then use the stub front option

[-COLUMN_STATISTICS_VALUES] to exclude specific rows on a question by question basis. If you have a job with more than one banner, the appropriate regions should be attached to the TABLE_SET that describes each banner using this EDIT option.

Once the regions have been attached to the banner, then for each TABLE_SET that describes the row you can add one of the following LOCAL_EDIT statements:

LOCAL_EDIT={: ANOVA }LOCAL_EDIT={: CHI_SQUARE }LOCAL_EDIT={: -TABLE_TESTS }

The first example says to do an ANOVA on this table for the regions specified in the banner. The second example says do a CHI-SQUARE on this table for the regions described in the banner. The third example says to do neither an ANOVA nor CHI-SQUARE on this table even though regions are specified with the banner.

The default format, EDIT=CHI_SQUARE_ANOVA_FORMAT, for printing the ANOVA and CHI-SQUARE statistics is at the bottom of the table with the banner categories used in the test designated by arrows.

ANOVA; F:     <--     2.94     -->NUM,DEN:             3,146PROB:                0.0345

It is possible with this format that there may not be enough room between the arrows to print the calculated chi-square value for the CHI-SQUARE test or the F-value for the ANOVA. Statistical tests can be summarized at the bottom of a table using the option -CHI_SQUARE_ANOVA_FORMAT on the EDIT statement and labeled using the syntax TABLE_TESTS=[$T=”text” $R col_list BY row_list]. The advantage of using this method is that there is more room to print the results.

EDIT={: -CHI_SQUARE_ANOVA_FORMATTABLE_TESTS=[$T="STATS FOR RELATIONSHIP " $R  2 TO 5  BY 1 TO LAST]TABLE_TESTS=[$T="STATS FOR FREQ OF USE  " $R  6 TO 8  BY 1 TO LAST]TABLE_TESTS=[$T="STATS FOR POTENTIAL    " $R  9 TO 11 BY 1 TO LAST]TABLE_TESTS=[$T="STATS FOR SALES        " $R 12 TO 15 BY 1 TO LAST] }

The results for an ANOVA would be printed as:

STATS FOR RELATIONSHIPanova = 2.94, df1,df2 = (3,146) prob = 0.0345STATS FOR FREQ OF USEanova = 0.18, df1,df2 = (2,124) prob = 0.8385STATS FOR POTENTIALanova = 0.52, df1,df2 = (2,124) prob = 0.6014STATS FOR SALESanova = 0.38, df1,df2 = (3,125) prob = 0.7704

The -CHI_SQUARE_ANOVA_FORMAT also allows you to test non-contiguous columns and to use the same column in more than one test, which is not possible when using the default format.

EDIT={: -CHI_SQUARE_ANOVA_FORMATANOVA=[$T="STATS FOR LOW VS MEDIUM  " $R   6,7 BY 1 TO LAST]ANOVA=[$T="STATS FOR LOW VS HIGH    " $R   6,8 BY 1 TO LAST]ANOVA=[$T="STATS FOR MEDIUM VS HIGH " $R   7,8 BY 1 TO LAST] }

You cannot do both a CHI-SQUARE and an ANOVA on the same table. When ANOVA=[region] is specified, it cannot be overridden by CHI_SQUARE.

When CHI_SQUARE=[region] is specified, it cannot be overridden by ANOVA. When TABLE_TESTS=[region] is specified and both ANOVA and CHI_SQUARE are specified, you will get an error message saying both tests cannot be done on the same table. (***It was possible to do both tests on the same table with the SPL software but this will not be implemented in C-Mentor unless a user comes up with a practical example.***)

Example spec:

~COMMENT TABTESTS.SPX&TABTESTS.DOC~DEFTABSET={BAN1:EDIT={:PDEC=0SDEC=1-ROW_NA-COL_TNAPUTCHARS=---CALL_TABLE=""RUNNING_LINES=2CWIDTH=5SWIDTH=25}BANNER={: