Background
Overview
The goal of this project was to answer a hypothetical business question with a card sort activity. I used Mural, Google Sheets, Figma, and Tableau to collect, clean, analyze, and visualize the data. The card sort was done in class on a Friday morning and I finished the analysis that afternoon, so this was a very short project.
Premise
“A caregiver service wants to make their website easier to navigate. Information about an individual caregiver is listed on their profile page, but there’s no way to filter caregivers by these attributes. The business is adding a filter function so users can view only caregivers who meet certain criteria.”
Business Question
What caregiver attributes belong in the search filter?
Constraints
1) Space is limited, put as few attributes in the filter as necessary.
2) Time is limited, be quick but thorough.
Process
Card Sort
Working in Mural, 19 participants independently sorted 49 caregiver attributes into either List Filter, Caregiver Profile, or an alternative category or subcategory that they came up with.
Data Cleaning
I copied the data from Mural into a spreadsheet and combined participant-generated main and subcategories to eliminate conceptual overlap. This reduced the total number of categories from 30 to 5.
I then converted these columns of text to attribute count by category. This revealed that 11 attributes were sorted more often than not into list filter.
Analysis
Since I was interested in identifying attributes to be included in the list filter, I conducted the analysis on only these 11 attributes that were sorted into list filter more often than caregiver profile.
I used binomial tests to decide which of these attributes should be included in the list filter. These p-values indicate the probability that an attribute was sorted into list filter as many times as it was by chance alone.
Small p-values provide evidence that most participants feel a given attribute belongs in the list filter. Using an alpha level of .05 as a tentative cut-off, only 5 of the 11 attributes should be included in the list filter.
* indicates p < .05, ** indicates p < .01, *** indicates p < .001, ~ indicates marginal significance.
Location, Infants, Children, Adults, and In-home Care clearly belong in the list filter. Seniors and Care Facility were not significant but included in the filter because of their conceptual similarity to at least one other attribute in the filter. This created more useful and intuitive categorization.
Key Findings
Based on the findings of this card sort activity and analysis, I found that users want to filter caregivers by the age groups they work with, the type of care they offer, and their location.
Recommendations
I created a low-fidelity mockup to demonstrate how a business might choose to display these attributes in the list filter. This design separates attributes into two subcategories, Age and Type of Care, and includes a location search function with a map.
