Concept Lab is a tool that allows you to compare concepts or entities on a variety of attributes using either pairwise comparison significance testing or indexing directly in your Results page.
Learn more about Concept Lab in the Lighthouse Academy!
In This Article
- Automated Concept Lab Report
- Setting Up Concept Lab
- Custom Setups
- Attribute Options and Methods
- Significance Testing
- Adjusting the Concept Lab
- Concept Lab Exports
Automated Concept Lab Report
Once a survey using Smart Loops or Group Logic begins fielding, the platform automatically generates a Concept Lab report. This report is created based on any supported group configurations included in your survey, so you'll have results ready without any extra setup.
To access an automated Concept Lab report, once a survey has begun fielding, navigate to the Stats page.
- Click Reports in the left sidebar.
- Select Automated Concept Lab Report beneath the Automated Reports section. The report will automatically populate with the Concept Lab and all questions included within the selected group. The questions are collapsed by default.
⚠️ Note: If the survey contains multiple groups, each subsequent group will appear directly after the final question of the previous group. Subgroups do not receive their own automated report. The automated report also requires that Concept Lab be enabled for the survey owner's account.
Click here to learn more about automated Concept Lab reports!
Setting Up Concept Lab
- On the Results page, click the Data Lab tab, and drag the Concept Lab component into your results.
- Click the note icon on the right side of the header, and give the Concept Lab a more specific name.
- Click the drop-down in the header to select another pre-defined group (identified by any group logic or Smart Loops used in the survey), or click Custom Set up to start with no pre-populated information in the table.
- Edit the labels for the entities or attributes by clicking within the appropriate text box.
- In each attribute row, enter the question or subquestion numbers you want to evaluate, then use the method picker to choose how those answers are calculated. The attribute row headers read Label and Hold Shift to apply to all.
- Click the Run comparison button. While the comparison is processing, the button reads Running comparison.
💡 Tip: Data source is a column heading in the Entities block and is read-only—it displays the question reference behind each entity rather than offering a selection.
⚠️ Note: Rows bound to a preset can be unlocked with the switch to advanced mode toggle, and returned with switch to basic mode.
Custom Setups
When you choose Custom setup from the drop-down, you must define every entity, and list all relevant Question/Subquestion numbers, separated by commas, in each attribute row.
⚠️ Note: If the questions referenced do not match in structure, the Run comparison button is disabled rather than hidden. The button only disappears when the editor itself is read-only—for example, when the feature is turned off or you are viewing a share link.
⚠️ Note: Run comparison is also refused when an attribute lists fewer references than there are entities. You will see the error: "For each attribute, please make sure there is a question number or subquestion number that corresponds to each entity."
Attribute Options and Methods
In Basic mode, attributes are generated from Radio, Checkbox, Slider, Star, Smiley, and Progressive Matrix question types in your survey. The method drop-down lists the individual answer options as programmed, followed by the available box scores and means.
For Progressive Matrix questions, any sub-type is supported. The sub-type determines the method control: single-choice (radio) sub-types offer a single pick with box scores, while any other sub-type offers a multi-select of answer options.
| Method | Description |
|---|---|
| Individual answer option | Compares entities on a single scale point. To obtain top box, select the option that corresponds with your most positive scale point (on a 5-pt scale, this will be A1 or A5, depending on how you programmed). |
| Top 2 Box | Combines the two most positive answer options. Offered from each end of the scale. Requires at least 3 positive answer options. |
| Top 3 Box | Combines the three most positive answer options. Offered from each end of the scale. Requires at least 4 positive answer options. |
| T-Weighted | A weighted box score counting the most positive option at 100% and the next at 50%—commonly described as a 1.5 top box. Offered from each end of the scale. Requires at least 3 positive answer options. |
| Mean | The arithmetic average score for each question/sub-question; coded as A1=1, A2=2, etc. |
| Mean (reverse coded) | The arithmetic average score for each question/sub-question, coded as A1=5, A2=4, etc., on a 5-pt scale. |
💡 Tip: Hold Shift while selecting a method to apply it to every attribute row at once.
⚠️ Note: If the questions referenced in an attribute have differing numbers of answer options, that attribute will not offer any method options. The field is flagged when you click out of it, and Run comparison stays disabled until it is corrected.
Significance Testing
The resulting Concept Lab displays the selected concepts side-by-side, with the attributes listed in rows. Click the Analysis drop-down to select Pairwise SigTest (default) or Indexing.
Pairwise SigTest compares the percent of respondents who fell into the selected measurement (for example, Top 2 Box) displayed in each cell. An uppercase letter in a specific cell indicates that attribute for that concept is statistically significantly different from the specified concept, at the 95% confidence level. A lowercase letter indicates that attribute is statistically significantly different at the 90% confidence level.
Indexing determines if the individual concept value is statistically significantly higher or lower than the average attribute value across all concepts. Higher is indicated by an up arrow, and lower is indicated by a down arrow.
⚠️ Note: To change the confidence levels from the default of 95% and 90%, open the View options sidebar panel and adjust the Primary and Secondary drop-downs under Confidence Levels for Sigtesting. Keep in mind that this changes the confidence levels for all significance testing on this report. The Primary and Secondary drop-downs found under the Banner tables/Crosstabs export option apply only to that export, not to the Concept Lab on the page.
Adjusting the Concept Lab
- Click the Analysis drop-down to toggle the analysis between Pairwise SigTest and Indexing.
-
Click the Split By drop-down menu to split the concepts by traits, variables, and Single Choice Virtual Questions. Only variables with between 2 and 20 values are offered.
⚠️ Note: Only Single-Choice Virtual Questions are supported for Split By. Virtual Questions with overlapping answer options or those configured as Multiple Choice will not appear as available Split By options in the Concept Lab. Learn more in Using Split By in Concept Lab.
- To remove a concept, hover over it and click the remove concept control. To add one, click the Choose another concept cell at the right edge of the table. There is no limit on how many concepts can be in the table at once—it scrolls horizontally.
- Click the Show drop-down to change what concepts you view, selecting from the default, All, or Top 3, Top 25%, or None. Top 3 is hidden when there are fewer than 4 concepts, and Top 25% is hidden when there are fewer than 3. Adding or removing concepts will automatically change this drop-down menu to custom.
- Click the Rank by: drop-down to rearrange the order of concepts from left to right. Original order arranges concepts as you entered them in the table. Select an attribute from the list to rearrange the order of concepts from highest percentage of that attribute on the left, to lowest on the right. Sample size is also available as a ranking option.
- Click the Expand icon to view and manipulate the Concept Lab in Full Screen mode.
Concept Lab Exports
Once a Concept Lab has been added to your Results page, it will appear in Excel Summary and PowerPoint exports. Keep in mind that whatever selections are in view (e.g., Indexing for Top 25%) will dictate what you see in the export.
There is also a dedicated Concept Lab .xlsx export available in the Export panel, and a per-block Export data button on the Concept Lab itself.
View a demo report. (Note that not all Concept Lab features will be accessible.)