Validation Checks#
Tools, then Check for Errors, runs every check against the open survey and lists the findings in the validation rail. Every export runs the same checks first. This page explains the severities, states which findings stop an export, and lists the checks under the source named beneath each finding. The rail itself is described in Questions, Search, Validation, and Comments.
Severities#
Severity |
Meaning |
|---|---|
Error |
The survey is inconsistent: an input that leads nowhere, a missing answer type, or a duplicate choice value or output name. |
Warning |
A quality problem to review: empty text, an unbounded roster, or a name an exporter will rewrite. |
Info |
An advisory: a computation that has only a description, or the state of a translation. |
In the rail, a red dot marks an error, an amber dot a warning, and a blue dot information. The Validation badge in the navigation bar takes the colour of the worst finding, and turns green when there is none.
Validation and Export#
The Specification, Paper Form, and DDI 3.3 exports proceed whatever the checks find. The Blaise 5 export, in Blaise Colectica Questionnaires, refuses to run while any finding is an error, or while a question produces no output. Instead of its options dialog it shows a message counting the issues to fix, and the validation rail lists them. No warning or information row stops an export.
Checks in Every Edition#
Each finding names its source: Structure, Questions, Outputs, Logic, or Translations. In the messages below, italics stand for the name, value, or number the finding reports.
Structure#
Check |
Severity |
Message |
|---|---|---|
|
Warning |
The survey has no questions. |
|
Warning |
The section is empty. |
|
Warning |
The display logic contains no items. |
|
Warning |
The roster is empty. |
|
Error |
The input alias is not connected to an answer. |
|
Error |
The repeat count must be at least 1. |
|
Error |
The maximum number of repeats must be at least 1. |
|
Error |
The roster repeats by an answer, but no answer is selected. |
|
Warning |
The roster repeats as many times as the answer to question, which has no minimum and maximum, so its size is unbounded. |
|
Warning |
Question allows up to n, but the roster allows at most m repeats. |
Questions#
Check |
Severity |
Message |
|---|---|---|
|
Error |
The question has no answer type. For a grid: The grid has no answer type. |
|
Warning |
The question produces no output. |
|
Warning |
The question has no text. |
|
Warning |
The statement has no text. For a note: The note has no text. |
|
Error |
The choice list has no choices. |
|
Error |
Choice n has no value. |
|
Error |
The choice value value is used more than once. |
|
Warning |
Choice value has no label. |
|
Error |
The other-please-specify text is attached to a choice that no longer exists. |
|
Warning |
The other-please-specify text is not attached to a choice. |
|
Error |
The minimum low is greater than the maximum high. |
|
Error |
The minimum length low is greater than the maximum length high. |
|
Warning |
The edit check has no message to show the respondent. |
|
Info |
The computation has only a description. |
Outputs#
Check |
Severity |
Message |
|---|---|---|
|
Error |
The output name name is also produced by item. Within one item: The output name name is produced more than once by this item. |
|
Warning |
The output name name contains characters that statistical packages may not allow. Without a name: The output has no name because the question has no name. |
Logic#
Check |
Severity |
Message |
|---|---|---|
|
Error |
The display logic “synopsis” refers to question, which is asked after this point; or refers to an answer inside the roster roster, which is not supported; or is incomplete or refers to an answer that cannot be found. |
|
Error |
The edit check “synopsis”, followed by the same explanations. |
|
Warning |
The condition on question contradicts item, so no respondent can ever see this item. |
Translations#
Check |
Severity |
Message |
|---|---|---|
|
Warning |
The language code code is not recognised. |
|
Info |
n of m strings have no language translation. |
Checks for Blaise 5#
Note
This feature is available in Blaise Colectica Questionnaires.
Blaise Colectica Questionnaires adds checks whose source is Blaise 5. They describe what the Blaise 5 generator cannot express or will change. Errors in this group, like every other error, stop the Blaise 5 export; warnings do not.
Check |
Severity |
Message |
|---|---|---|
|
Error |
The choice value value is not a whole number. Blaise requires numeric choice values. |
|
Error |
The display logic has only a description. Add formal display logic or Blaise code before exporting to Blaise. |
|
Error |
The edit check has only a description. Add a formal condition or Blaise code before exporting to Blaise. |
|
Warning |
The computation has no formula and will not be exported to Blaise. |
|
Error |
Blaise export does not support while and until loops. Use a roster instead. |
|
Error |
Blaise supports one answer type per question, or a choice list followed by an attached other-please-specify text. |
|
Error |
The answer of a choice grid must be a choice list for Blaise export. |
|
Error |
The rows of a grid must be a choice list for Blaise export. |
|
Warning |
The name name is not a valid Blaise identifier and will be exported as identifier. |
|
Warning |
The date type type is exported to Blaise as free text; comparisons on it in display logic will not work. |
|
Warning |
Blaise limits text answers to n characters; the maximum length m will be reduced. |
|
Error |
The computation uses function, which cannot be exported to Blaise. |
|
Warning |
The text contains a caret (^), which Blaise reserves for fills; it will be removed on export. |
|
Warning |
The Blaise code refers to |
See also
Test a Survey works through the findings; Generate a Blaise 5 Survey covers the Blaise 5 export.