Significances

This feature allows you to create a table or graph to show results of significance testing in online report.

To add a Significances to your report, select the Significances option from the Insert dropdown menu on the Report Items page.

Use the Name table/chart text box to enter the name of a table or chart.
Provide the Description of a table or chart.
In the Available Responses text box select the responses the significances calculation will be based on.
Use the Select "brand of interest" drop-down box to choose one of items significances will be calculated for. This list corresponds to the items selected for the row responses.
Select groups of respondents that will be used as a control group and as a test group for calculating table data. Click the Apply button to create data set based on the selected respondent groups. Click the delete button next to the set you need to delete.
Click the Edit Groups button to create new or edit existing group.
Input logical expression to create a sample for table or chart, then click Apply. (For example: Q1.A2 AND (Q3.A5 OR Q4.A1)).
Select one of the following options to calculate significances: Number, Percent, Response rate, Difference, Lift, Zscore, Probability,Significance.
Select the Table option to display calculated data as a table.
Select the Chart option to display data as a graph.

Calculation Formulas

Values calculated for either control or test group separately

Name	Formula	Description
Number (n)	Any number	Equals to the number of respondents belonging to this group who selected this answer choice.
Percent (%)	Number / general number of respondents in group * 100%	Equals to Number divided by the general number of respondents in this group and multiplied by 100%.
Response rate	Number / general number of respondents in group	Equals to Number divided by the general number of respondents in this group.

Values calculated for both test and control groups

Name	Formula	Description
Difference	Response rate (T1) - Response rate (C1)	Subtract Response rate of control group (C1) from Response rate of test group (T1).
Lift	(Percent of T1 - Percent of C1)/ Percents of C1	Subtract Percent of control group (?1) from Percent of test group (T1). Divide the value obtained in step 1 by Percent of control group.
Zscore	ABS(difference/SQRT((Response rate of T1(1- Response rate of T1) /n of T1)+ (Response rate of C1(1- Response rate of C1)/n of C1)))	Subtract Response rate of test group (T1) from 1 and divide it by the number of respondents of test group and multiply it by Response rate of test group. Subtract Response rate of control group (C1) from 1 and divide it by the number of respondents of control group and multiply it by Response rate of control group. Evaluate square root of the sum of values obtained in steps 1 and 2. Divide Difference of test and control groups by the value obtained in step 3. Evaluate absolute value for the result obtained in step 4.
Probability	ABS(2*NORMSDIST(Zscore)-1)	Evaluate the standard normal cumulative distribution function for Zscore of test and control groups and multiply it by 2. Subtract 1 from the value obtained in step 1. Evaluate absolute value for the result obtained in step 2.
Significance	IF(Zscore >1.96,"S1",IF(Zscore >1.645,"S2",IF(Zscore >1.283,"S3","")))	Status indicators: S1 Highly significant indication of difference in response rates. S2 Significant indication of difference in response rates although not conclusive. S3 Slight indication of difference in response rates. Blank No significant indication of difference in response rates.

Click Save to save your changes.
Click Preview to see how your report looks.
Click Back to return to the New Report page.
Click Next to go to the Report Layout page.