Table Statistics

A table statistic automatically generates a row for the chosen statistic, using the table's rows as the basis for the computation. Table statistics are calculated based on one value for each row. If row values are ranges, you must use row statistics (see Row Options).

Use Setup|Profile Settings|Statistics to specify default table statistics settings to be applied to every table for a new job.
Use Setup|Job Settings|Statistics to specify default table statistics settings to be applied to every new table for an existing job.
Use Setup|Tables|Statistics or Setup|Globally Modify Tables|Table Statistics to modify table statistics settings for pre-existing tables in an existing job.
Setup|Profile Settings|Wording for Rows or Setup|Job Settings|Wording for Rows let you modify the default wording for statistic rows (text is automatically applied to these rows). Setup|Profile Settings| Statistics Rows or Setup|Job Settings|Statistics Rows let you modify the order in which the rows should print.

In Setup|Tables, select Statistics:

Sample size for statistic base

Displays the base for statistics, which is the total sample minus any ‘no answers’ and any rows specifically excluded from the statistic base.

Mean

Specifies the values to use in calculating any descriptive statistics: the mean, standard deviation and standard error. The mean is the sum of the values divided by the number of values, so the values in all rows are used in calculating these statistics for tables.
When row values are defined with ranges, you must use a row mean. You specify a row mean using Row Options.

Mean confidence interval - lower

Displays the lower mean confidence interval based on the selected confidence interval level specified using the Select Level option. The confidence interval level specified is used for calculating both the lower and upper mean confidence interval.

Mean confidence interval - upper

Displays the upper mean confidence interval based on the selected confidence interval level specified using the Select Level option. The confidence interval level specified is used for calculating both the lower and upper mean confidence interval.

Standard deviation

Measures variance from the mean for a set of values.

Standard error

Measures the standard error of means. It is calculated by dividing the standard deviation by the square root of the sample size.

Median

Reports the middle value of all values; half the values of the variable fall below the median and half fall above. For tables with ranges of values, use a grouped median.

Note: The Median does interpolate if the value falls at exactly 50%. The Grouped median does interpolate if you are using ranges.

Changing values for statistics

Grouped median

When row values are defined with ranges, you can specify a grouped median. If the values are not in a range and you want to do a group median, you can edit the row logic and use the at symbol (@) to change a single value to a range.

The @ symbol is used only for grouped medians. The grouped median is used on tables that have a range of values represented by a single value.

The regular median is used on tables where the median can be calculated on the actual values. The regular median might be used, for example, on a table that represents a question in which a respondent was asked to state his or her age in years. The @ is not used for the regular median. When a regular median is calculated, it uses the value in parentheses (or the # sign, described above, can be used to specify a different value for the calculation).

The grouped median is used on tables, for example, where a respondent is asked the range in which his or her age falls. Another example of a question type that would call for a grouped median in a table is income, where respondents are asked to indicate an income range, rather than a precise dollar value.

To calculate the grouped median, you need to specify mutually exclusive ranges with the @ sign. For example, for the following five age categories, we could use ranges in place of the original category values (1-5):

UNDER 25	1/79 (1)	@15-24.9
25-34	1/79 (2)	@25-34.9
35-44	1/79 (3)	@35-44.9
45-54	1/79 (4)	@45-54.9
55+	1/79 (5)	@55-64.9

For example, UNDER 25 would use the range 15-24.9 rather than the original value of 1; 25-34 would use the range 25-34.9 rather than the original value of 2 and so on.

To calculate both the grouped median and use a different value for the mean (described above), we could specify:

UNDER 25	1/79 (1)	#17.5 @15-24.9
25-34	1/79 (2)	#29.5 @25-34.9
35-44	1/79 (3)	#39.5 @35-44.9
45-54	1/79 (4)	#49.5 @45-54.9
55+	1/79 (5)	#59.5 @55-64.9

1st Quartile

Represents the value where 25% of the values fall below that value and 75% of the values fall above it.

3rd Quartile

Represents the value where 25% of the values fall above that value and 75% of the values fall below it.

Mode

Reports the value that occurs the most often.

Minimum

Shows the minimum value.

Maximum

Shows the maximum value.

Note: The Minimum and Maximum options are most meaningful when specified on a table defined using a Frequency Row(s).

Effective sample size for statistic base

Creates a row on your table displaying the effective sample size used when calculating T-scores for significance testing on means.

Mean number of mentions

Creates a row on your table displaying the total number of responses divided by the total number of respondents.

Percentiles

Select the Percentiles (from 0 to 99) that you would like to show on your table.
Percentiles show the value where the selected percentage of the values fall below and the remainder fall above.
For example, if you select the 10th Percentile, this will display the value where 10% of the values fall below and 90% fall above.

Show dollar sign ($)

For tables with dollar variables, displays dollar signs.

Show comma (,)

For tables, displays commas.

Scaling

When values are too large to fit in a cell, you can choose to scale all statistics.

Choose one of the scaling options (Divide by 10 (move the decimal point one place), Divide by 100, Divide by 1,000 or Divide by 10,000).

Statistical testing

You select statistical tests for banner columns when you use Setup|Banners.
You must choose at least one test if you want statistical testing on the table: Means, Percents and Chi-Square.
Your selection applies only to the current table.
For banner columns, you must select the test you want to conduct from the Statistical Testing dialog box in Setup|Banners.

Statistical testing must be selected for a banner AND selected for a table.

To select statistical testing for a banner, use Setup|Banners. To select statistical testing for a table, use any of the following:

Setup|Job Settings|Statistics
Setup|Tables and choose Statistics
Setup|Globally Modify Tables|Table Statistics

Decimal places

You can define the decimal precision for all statistics when selecting Plain text report or Enhanced text report as the Report viewing format option in Run|Tables.
Central tendency applies to Mean, Median and Mode and other table statistics (Minimum, Maximum, 1st Quartile, etc.). You can select 0, 1, 2 or 3 decimal places from the dropdown. The default for Central tendency decimal precision is 2 decimal places.
Variability applies to Standard deviation and Standard error. You can select 0, 1, 2 or 3 decimal places from the dropdown. The default for Variability decimal precision is 2 decimal places.

Select rows, if any to exclude from statistics

Every row in a table is included in the statistic base (except rows indicated as net rows by NET or as subtotal rows by SUB). Select rows you want to exclude (they are then highlighted).

Note: The Select rows, if any to exclude from statistics option is not available when using Setup|Globally Modify Tables|Table Statistics.