Summary:
Labels in a card sorting study must be neutral to prevent keyword matching and encourage careful, conceptual groupings from users.
Card sorting is a specialty UX research method used to understand users’ mental models of the information architecture (IA) within your digital product. Users organize individually labeled notecards into groups according to criteria that make sense to them.
The phrasing on each card label can significantly influence how users group cards. Card labels should be neutral to prevent participant bias and encourage users to consider each card carefully.
Keyword Matching in Card Sorting
In card-sorting studies, participants sometimes tend to group cards with similar keywords rather than conceptually organize them carefully. This happens because it requires less cognitive effort to identify patterns through keywords than to think about the meaning of each card and how it’s conceptually related to other cards.
Keyword matching is a problem because it focuses on superficial similarities rather than conceptual relationships between items. This can lead to inaccurate groupings that don’t accurately represent users’ mental models regarding the relationships between items.
Example: Intranet Card Sort
Let’s imagine you are running a card-sorting study for your company’s intranet with the following cards. Each card corresponds to pages users can navigate to on the intranet. Participants are asked to organize cards into groups that make the most sense to them and provide category names for each group.
Cards |
---|
Employee Directory |
Professional Development Workshops |
Health and Safety Regulations |
Employee Benefits |
Certification Programs |
HR Contact Information |
Employee Online Training |
Participants will likely create groups with similar keywords with this set of cards, placing cards with Employee together and those with HR together.
❌ Keyword Matching: Example Groups | |||
---|---|---|---|
Group 1: Employee Resources* | Group 2: HR* | Group 3: Continuous Learning | Group 4: Regulations |
Employee Directory | HR Contact Information | Professional Development Workshops | Health and Safety Regulations
|
Employee Benefits | HR Policies and Procedures | Certification Programs |
|
Employee Online Training |
|
|
|
Does grouping according to keywords say much about how conceptually they think about each card? No.
Let’s focus on group 1, which includes Employee Directory, Employee Benefits, and Employee Online Training. Grouping all cards with Employee in them may result in a vague category label like Employee Resources to encompass all items.
Vague categories don’t accurately represent how users think about each group of cards and how they relate to one another. This can ultimately lead to content organization decisions that don’t reflect users’ true mental models.
To avoid this problem, researchers need to encourage users to think beyond keyword matches. Designing cards that require participants to carefully consider each card’s purpose ensures that the resulting groupings reflect users’ conceptual understanding and the relationships between items. This approach more accurately represents how users interact with your site’s navigation, where they expect similar items with similar conceptual meanings to be placed near each other, not just those with the same keyword.
Card labels should compel participants to recognize that Employee Directory, Employee Benefits, and Employee Online Training serve distinct purposes, despite all containing the word Employee.
If participants were encouraged to consider each card carefully, they’d likely see that Employee Online Training fits better with Group 3, which includes other continuous learning-related cards.
✅ Content Matching: Example Groupings | |||
---|---|---|---|
Group 1: Employee Resources | Group 2: HR | Group 3: Continuous Learning | Group 4: Regulations |
Employee Directory | HR Contact Information | Professional Development Workshops | Health and Safety Regulations |
Employee Benefits | HR Policies and Procedures | Certification Programs |
|
|
| Employee Online Training* |
|
Card Labeling Strategies
We must label cards strategically to encourage deeper consideration of the content and meaning of each card (instead of superficial card labels). Use synonyms, non-parallel structures, and in-depth descriptions to prevent keyword matching.
Synonyms
One way to avoid having participants simply match up keywords is to use different words for the same concept.
For example, if we already have a card labeled Employee Benefits, another card labeled Employee Directory will be likely to produce keyword matching. We can use the synonym Staff instead and label the second card Staff Directory. Using the synonym Staff shifts the focus from specific words to the broader conceptual meaning and associations with other cards.
Original Cards | Revised Cards |
---|---|
Employee Benefits | Employee Benefits |
Employee Directory | Staff Directory |
Understanding which synonyms users see as interchangeable can also be valuable when building out a taxonomy and crafting search-friendly content.
Non-Parallel Structures
Using non-parallel structures for card labels can effectively disrupt users from quickly scanning for similarly structured labels and making groupings without much thought.
For example, we might rephrase the Employee Benefits card to be Benefits for Employees. This approach structurally deviates from similarly written cards like Employee Directory and Employee Online Training.
Original Cards | Revised Cards |
---|---|
Employee Benefits | Benefits for Employees |
Employee Directory | Staff Directory |
Employee Online Training | Employee Online Training |
In-Depth Descriptions
In-depth descriptions in card labels can help mitigate keyword matching by providing additional context about the content.
For example, instead of Employee Online Training, a more detailed label might be Interactive digital courses for skill enhancements. This conveys more details about the medium and purpose of the training. Short descriptions encourage participants to reflect on the specific content represented by the cards they’re organizing.
Facilitation Strategies to Encourage Conceptual Grouping
In addition to modifying card labels, there are a few facilitation strategies researchers can employ to encourage conceptual groupings in card sorting.
Explain the Objective
Explain the study’s objectives at the beginning of the session. Help users understand the importance of focusing on concepts rather than terminology before they begin sorting cards. Explain that the aim is to understand how they conceptualize and categorize the information on the cards. This briefing will help them set the right mindset from the start.
Encourage Thinking Out Loud
Encourage users to think out loud as they sort cards. Asking users to think out loud will help you understand whether users grouped cards based on similar keywords or considered each card conceptually.
Ask for Subcategories
If your participants engage in keyword matching, try asking them to subcategorize cards or combine cards into a larger group. This may encourage them to think more deeply about their organization structures, and work to differentiate them further.
Allow Iterative Changes
Allow participants to make iterative changes to their groupings. After their initial sort, ask users follow-up questions such as:
- Were any items especially easy or difficult to place?
- Did any items seem to belong in two or more groups?
- What are your thoughts on the items left unsorted (if any)?
These questions usually require deep consideration of the participant’s rationale. Often, participants discuss how they would organize it differently. Allow them to make those refinements after careful reflection.
Conclusion
Effective card sorting goes beyond keyword matches. Card labels should encourage users to deeply consider each card and its association with others, which is crucial for the usefulness of your study. Through card labeling strategies and facilitating techniques, researchers can prevent keyword matching and encourage participants to make groupings that reflect their mental models.
Post a Comment