Calculations are organised by table type, which may be Flat or Cross. This is determined by whether you are creating a single or a two-dimensional table:
Flat table presents a question or a sequence of questions in the rows or the columns only (not both).Cross table presents one or more questions in both the rows and the columns which then intersect to form the body of the table.While most calculation types are available for both flat and cross tables, some are restricted to one context or the other, if they are only valid in that context. Where a calculation is not available, it is shown as not applicable in the descriptions below.
There are several groups of calculations which apply to the specific type of data being referred to. You can jump to the calculations for each group by following the links below:
These calculations apply to single and multiple closed questions.
| Calculation | FLAT | CROSS |
|---|---|---|
| Counts | The number of respondents who cited the response. | |
| Percentage | The number of people who cited the response, divided by the base number of people who were asked the question, multiplied by 100. | Not applicable |
| Cumulative percentage | The total of percentages. | Not applicable |
| Percentage across | Not applicable | The percentage compared to the total for the row (i.e. the horizontal percentage). |
| Percentage down | Not applicable | The percentage compared to the total for the column (i.e. the vertical percentage). |
| Overall percentage | Not applicable | The percentage calculated based on the total of all persons interviewed. |
| Answered frequency | The number of people who cited the response, divided by the base number of people who were asked this question (excluding DK responses), multiplied by 100. | Not applicable |
| Mention rate |
The number of people who cited the response, divided by the total answers (sum of citations), multiplied by 100. |
Not applicable |
| Asked frequency |
The number of people who cited the response, divided by the base number to whom the question was presented (whether they gave an answer or not), multiplied by 100. |
Not applicable |
| Significance | Not applicable | The significance deviations. |
| Counts when independent | Not applicable | The count value, if the variables are independent. |
| Test value | Not applicable | The test value. |
| Affinity index | Not applicable | The indices in base 100. For an example, see "affinity index" example, below. |
| Column Significance testing | Not applicable | The significance differences between columns. |
| Z-score (closed) | Not applicable | Z-score (also known as "standard score" allows you to compare observations from different normal distributions, allowing you to reduce the effects of different sample sizes between columns. Z-score (closed), as opposed to Z-score (numeric), is calculated using the percentages on a cell within the results table (and therefore appears within that cell). |
| Row significance | Not applicable | The significance differences between the means of two rows. Applies to one question per row. |
| Interval (-) | Lower percentage of the confidence interval calculated at 95%. To modify the error percentage, open the advanced properties and modify the value 95. | Not applicable |
| Interval (+) | Upper percentage of the confidence interval calculated at 95%. To modify the error percentage, open the advanced properties and modify the value 95. | Not applicable |
| Weighting coefficient | The mean coefficient of the weights for the individuals present in the response. | Not applicable |
|
Effective base |
The sum of weight factors squared, divided by the sum of the squared weight factors. | The sum of weight factors squared, divided by the sum of the squared weight factors. |
| Mention rate across | Not applicable |
The number of selected responses, divided by the total answers (sum of citations), multiplied by 100. As the percentage is based on the number of responses rather than the number of respondents, the percentage may be greater than 100% for a multi-coded question. |
| Mention rate down | Not applicable |
The number of selected responses, divided by the total answers (sum of citations), multiplied by 100. As the percentage is based on the number of responses rather than the number of respondents, the percentage could be greater than 100% for a multi-coded question. |
| Column on column significance | Not applicable |
A Chi2 test with Cochran Adjustment. |
These calculations apply to single and multiple closed questions.
| Calculation | FLAT | CROSS |
|---|---|---|
| Response mode |
The response text of the code which has the modal count (i.e. the most frequently-selected response). When two or more responses tie for the modal count, the first encountered code is displayed. |
The response text of the code which has the modal count (i.e. the most frequently-selected response). When two or more responses tie for the modal count, the first encountered code is displayed. |
| Average number of responses | Advanced options for average number of responses. | Advanced options for average number of responses. |
These calculations apply to numeric questions.
| Calculation | FLAT | CROSS |
|---|---|---|
| Mean | The sum of values divided by the base of persons interviewed on this response (excluding DK responses). | The sum of values divided by the base of persons interviewed on this response (excluding DK responses). |
| Standard deviation | The dispersion around the mean. Where n= Sample base and Xi= value of each observation, Standard deviation= Sum (xi - column mean)² divided by n. | The dispersion around the mean. Where n= Sample base and Xi= value of each observation, Standard deviation= Sum (xi - column mean)² divided by n. |
| Number of interviews | The number of persons in the column who gave a response to the numerical question. | The number of persons in the column who gave a response to the numerical question. |
| Number of 'don't knows' | The number of persons in the column who did not give a response to the numeric question (in other words, Don't Know, or DK). | The number of persons in the column who did not give a response to the numeric question (in other words, Don't Know, or DK). |
| Sum | The sum of values cited for all the respondents. | The sum of values cited for all the respondents in the columns. |
| Minimum | The minimum value cited for all the respondents. | The minimum value cited for all the respondents in the columns. |
| Maximum | The maximum value cited for all the respondents. | The maximum value cited for all the respondents in the columns. |
| Standard error |
The error margin when you want to assign a value to a calculated mean in a sample. |
The error margin when you want to assign a value to a calculated mean in a column. |
| CI on mean (-) | The lower mean of the confidence interval calculated, by default, at 95%. To modify the error percentage, open the advanced properties and modify the value 95. | The lower mean of the confidence interval calculated, by default, at 95%. To modify the error percentage, open the advanced properties and modify the value 95. |
| CI on mean (+) | The upper mean of the confidence interval calculated, by default, at 95%. To modify the error percentage, open the advanced properties and modify the value 95. | The upper mean of the confidence interval calculated, by default, at 95%. To modify the error percentage, open the advanced properties and modify the value 95. |
| Variance | The mean of squares of deviations from mean for the column. | The mean of squares of deviations from mean for the column. |
| Median | The value around which there are two equivalent groups of counts. | The value around which there are two equivalent groups of counts. |
| Smoothing coefficient | Designates the value of the Fisher smoothing coefficient. The Fisher smoothing coefficient (kurtosis) is a measure allowing evaluation of the shape of a distribution of values for a variable compared to a theoretical distribution, as follows:
|
Not applicable |
| Asymetry coefficient | Returns the asymmetry of a distribution. This function characterizes the degree of asymmetry of a distribution with respect to its mean. A positive asymmetry (skewness) indicates a unilateral distribution offset towards the most positive values. A negative asymmetry (skewness) indicates a unilateral distribution offset towards the most negative values. | Not applicable |
| Gini coefficient | Returns the concentration coefficient. The concentration coefficient allows definition of the concentration, which varies between 0 (equidistribution) and 1 (maximum concentration). | Not applicable |
| Significance (Mean) | Not applicable | The significance differences between means. For details of the formula used, see the Knowledge Base article Significance and Column Significativity test formulas. |
| Column significance testing | Not applicable |
The significance differences between the means of two columns. For details of the formula used, see the Knowledge Base article Significance and Column Significativity test formulas. Note: Your column significance testing for a particular cell (on closed calculations, this does not apply to numeric ones) can change when the total base of their table grows, even though the count and percentage for the cells in the table remains the same. For details, see the article Significance and Count Threshold.
|
| Z-score (numeric) | Not applicable | Z-score (also known as "standard score" allows you to compare observations from different normal distributions, allowing you to reduce the effects of different sample sizes between columns. Z-score (numeric), as opposed to Z-score (closed) is calculated using the stats of a results table (and therefore appears as an additional row in the stats sections of the table). |
| Standard deviation estimator | Provides the estimated standard deviation. This option is available on standard error and confidence interval calculations only. | Provides the estimated standard deviation. This option is available on standard error and confidence interval calculations only. |
| Mode | Returns the mode (i.e. the value that occurs the most frequently). | Returns the mode (i.e. the value that occurs the most frequently). |
| Percentile | Returns the number of records that fall below the specified percentage. Specify the percentage in the advanced properties for this calculation. | Returns the number of records that fall below the specified percentage. Specify the percentage in the advanced properties for this calculation. |
Script calculations allow you to define advanced analyses that you cannot produce with the built-in calculations, or define tables in a more efficient manner (e.g. fitting more information into a single break). They can also be used to carry out arithmetic operations on the results of other arithmetic operations, without displaying the intermediate operations. Scripts can return a number (e.g. weighted counts or unweighted counts) or a string (such as displaying a custom text or message to indicate that something is significant). For detailed information on script calculations, please refer to this Knowledge Base article: Script calculations.
Script calculations have a script property that determines which variable is used as the basis of the calculation. To set this, click the script button in the calculation properties. You can then enter the reference to the variable (e.g. ??Q1??).
For a worked example of script calculations, please see the following Knowledge Base article: Script Calculations.
| Calculation | FLAT | CROSS |
|---|---|---|
| Script (sum) |
Returns the sum split by the dimensions of your table. An example can be found in the following article: Mean Summary Tables. In the script property, enter the shortcut of the variable you want to use as the basis of the calculation (e.g. ??Q1??). |
Returns the sum split by the dimensions of your table. An example can be found in the following article: Mean Summary Tables. In the script property, enter the shortcut of the variable you want to use as the basis of the calculation (e.g. ??Q1??). |
| Script (percentage) |
Returns the percentage split by the dimensions of your table. In the script property, enter the shortcut of the variable you want to use as the basis of the calculation (e.g. ??Q1??). |
Not applicable |
| Script (horizontal percentage) | Not applicable |
Returns the percentage compared to the total row. In the script property, enter the shortcut of the variable you want to use as the basis of the calculation (e.g. ??Q1??). |
| Script (vertical percentage) | Not applicable |
Returns the percentage compared to the total column. In the script property, enter the shortcut of the variable you want to use as the basis of the calculation (e.g. ??Q1??). |
| Script (mean) |
Returns a mean value for a variable which will have numeric analysis values assigned. You can define these analysis values in the script. For example, you can refer to ??rating?? for a numeric or for a closed question On(??rating??,1,2,3,4,5,nr) but you can change the analysis values for a closed question if you require: On(??rating??,100,75,50,25,0, The following article explains this in depth: https://support.askia.com/hc/ |
Returns a mean value for a variable which will have numeric analysis values assigned. You can define these analysis values in the script. For example, you can refer to ??rating?? for a numeric or for a closed question On(??rating??,1,2,3,4,5,nr) but you can change the analysis values for a closed question if you require: On(??rating??,100,75,50,25,0,
The following article explains this in depth: https://support.askia.com/hc/ |
| Script (standard deviation) |
Returns the standard deviation for a variable which will have numeric analysis values assigned. You can define these analysis values in the script. For example, you can refer to ??rating?? for a numeric or for a closed question On(??rating??,1,2,3,4,5,nr) but you can change the analysis values for a closed question if you require: On(??rating??,100,75,50,25,0, nr). |
Returns the standard deviation for a variable which will have numeric analysis values assigned. You can define these analysis values in the script. For example, you can refer to ??rating?? for a numeric or for a closed question On(??rating??,1,2,3,4,5,nr) but you can change the analysis values for a closed question if you require: On(??rating??,100,75,50,25,0,nr). |
| Script (standard error) |
Returns the standard error (for a variable you reference in the script) which will have numeric analysis values assigned. You can define these analysis values in the script. For example, you can refer to ??rating?? for a numeric or for a closed question On(??rating??,1,2,3,4,5,nr) but you can change the analysis values for a closed question if you require: On(??rating??,100,75,50,25,0,nr). |
Returns the standard error for a variable which will have numeric analysis values assigned. You can define these analysis values in the script. For example, you can refer to ??rating?? for a numeric or for a closed question On(??rating??,1,2,3,4,5,nr) but you can change the analysis values for a closed question if you require: On(??rating??,100,75,50,25,0,nr). |
| Script (count) |
Returns the count for the specified variable, split by the dimensions of your table. The option e.g.
|
Returns the count for the specified variable, split by the dimensions of your table. The option e.g.
|
| Script (frequency) |
Returns the count (for a variable you reference in the script), split by the dimensions of your table. e.g.
|
Returns the count (for a variable you reference in the script), split by the dimensions of your table. e.g.
|
| Script (base) |
Returns total rows or total columns based on the variable (or condition) in the script. |
Not applicable |
| Calculation | FLAT | CROSS |
|---|---|---|
| Chi2 test | Not applicable | Advanced options for Chi2 tests. |
| Paired preference | Not applicable | Advanced options for paired preference tests. |
| Correlation | Not applicable | The R correlation coefficient. The formula used can be found here. |
| Calculation | FLAT | CROSS |
|---|---|---|
| Arithmetic calculation (number) |
Allows you to create bespoke calculations using askiaanalyse’s new scripting language. The calculations can be programmed to return numbers or text found in any table in your portfolio. |
Allows you to create bespoke calculations using askiaanalyse’s new scripting language. The calculations can be programmed to return numbers or text found in any table in your portfolio. |
| Arithmetic calculation (string) |
Allows you to create bespoke calculations using askiaanalyse’s new scripting language. The calculations can be programmed to return numbers or text found in any table in your portfolio. |
Allows you to create bespoke calculations using askiaanalyse’s new scripting language. The calculations can be programmed to return numbers or text found in any table in your portfolio. |
|
Preferred color |
Men: 76 |
Women:74 |
Sample: 150 |
|
Red: |
15 |
25 |
40 people |
|
Profile column |
19.7% |
33.8% |
26.7% |
|
Affinity index |
(19.7% / 26.7% ) * 100 = 73.8 |
(33.8% / 26.7%)* 100 = 126.6 |
100 |