You may have received a report from your vendor with a cross-tab that you need to dig into. You’ve replicated the cross-tab in KnowledgeHound, but things aren’t lining up the way you expected. You might wonder why the data may look different in KnowledgeHound- here are a few things to check first.

On slide decks and in data tables, your supplier will often include footnotes with details about the source of stats and other information. Use this as a guide to determine how to build similar visualizations on KnowledgeHound.

Choose the correct study

Sometimes similar projects are available but are potentially split up by market, supplier, or time period. Ensure your study is the one you’re really looking for by looking at the name, supplier, date of field, attached documents, and other context.

Choose the right question/variables

Variables may be labeled differently than how they were asked on the survey, so it will be helpful to search the question list thoroughly. You can also review the IDs of questions on the “questions tab”. (Not seeing a variable you need? Some, such as banners, programming variables, or other derived variables, may be available but hidden. If you suspect that this is the case, reach out to KnowledgeHound support for help.)

Utilize the correct filters/crosstabs

While reviewing the footer of your report, keep an eye out for additional filters that may be applied. Does this visual represent a certain time period, audience, or other segment of the whole sample?

Toggle the necessary weighting scheme

The data may need to be weighted or unweighted. KnowledgeHound lets you toggle between available weighting schemes in the data file. Learn more about weighting here. (Not seeing a weighting variable you need? It is possible that some weighting schemes have been hidden or are unavailable. Reach out to KnowledgeHound support for help.)

Arrange your axes

KnowledgeHound is a flexible tool, and how you structure your visual may determine how it should be interpreted. Consider transposing the X and Y axes for additional views. Learn more about how to interpret a visual here.

Investigate base sizes

On any given variable, you have the option to show the ‘total’ available responses for that question. Each question on KnowledgeHound includes all frequencies, presenting data based to those who were exposed to the question. Variables in reports, such as brand funnel questions, may be based to the total sample instead. Try viewing your information as a spreadsheet and switching from % to numeric view. Should base sizes differ than you expect, it may be worth reviewing each individual variable to see if one has a significantly different base size than expected.

Revisit the data file that was shared with KnowledgeHound

If the above steps don’t provide clarity on potential differences, it’s possible the mismatch has to do with the structure of the data file itself. Learn more about our Data Quality Guide here to see how data should be sent to KnowledgeHound for a polished output. Reach out to our support for additional help!