organization To make a user-centric categories, the information architect has technical assistance in decision making, as is the so-called 'card sorting' cards or management.
The technique of 'card sorting' is based on observation of how users together and associate with each other a predetermined number of cards labeled with the different categories Theme of the website.
Thus, based on the behavior of the users, you can organize and classify information a web site according to their mental model.
The technique of 'card sorting' is based on observation of how users together and associate with each other a predetermined number of cards labeled with the different categories Theme of the website.
Thus, based on the behavior of the users, you can organize and classify information a web site according to their mental model.
Types 'card sorting'
We can distinguish between two types of 'card sorting': open and closed.
In the 'card sorting' open the user can freely include the categories in the number of sets you think necessary, while in the closed, groups or sets are predefined and labeled and the user only need to put each category in the group creating appropriate.
This second type of 'card sorting' is recommended to check if a classification information is familiar and understandable to the user, while the "Open" aims to discover what kind of classification of categories would be more correct.
Test Driving The first step is to recruit participants for the test. The selection must take into account these features and profiles should be consistent with the target audience of the website. No point having young participants and recent graduates in computer engineering, when the website is intended to be used by housewives, uneducated and elderly.
The number of participants in the test must be at least 5. The time allotted for the test depend on the number of categories to sort, although indicative, for 40 categories should allocate at least 20 minutes.
At the beginning of the test should explain to participants the objective of the test, and that this is not the assessment of the participants. In addition, participants indicated what should be the criterion for grouping of the categories by similarity.
then gives users the cards with the different categories, ensuring that each set delivered to each user is disordered (shuffled). Delivered to all users cards with a predefined can influence the form of management by the participants.
For 'card sorting' open type, once they have completed the test participants may be asked to give names to each of the groups created.
There are, likewise, software applications with which to carry out the virtual test, as EZSort or WebSort.
qualitative analysis of the results
This type of analysis can be performed when the participants as well as the number of categories, are not very numerous.
consists, first, to observe individually to each user during the test, and record all aspects of how each user organizes the cards, what problems have to do the job, which brings immediate and intuitive categories and questions about what else, what questions do during the test, etc.
then discusses "eyeball" the groups created by the participants, noting which categories most closely related to what categories according to the user's mental model.
Quantitative analysis of the results
This type of analysis is the statistical processing of data, and the subsequent "summary" of the results through graphical representations that facilitate their interpretation by the information architect. It is, therefore, a proper analysis to test a large number of participants and categories to sort.
first create a table of co-occurrences in a spreadsheet. In this one, with many rows and columns as the number of different categories, indicate the number of times each pair of categories have been placed in one group, resulting in a symmetric matrix as follows.
Then, on the co-occurrence table apply dimensional reduction algorithms such as clustering algorithms and multidimensional scaling (MDS), whose function is to simplify the relationships between categories to a number of dimensions easily interpreted by visual inspection (2D or 3D).
For the application of these algorithms have to have a statistical software package that allows it, such as SPSS or Statistica.
As can be seen in the figures, the result of applying clustering algorithm is a dendrogram, and the MDS geometric representation of the categories, separated according to their own di-similarity between them. Dendrogram (clustering), MDS (Multidimensional Scaling). Conclusions
technique 'card sorting' technique is a useful categorization of user-centric content, providing a double tool: to help decision making in the conceptual design phase (open card sorting) and to evaluate a particular organization of categories in usability evaluation stages (closed card sorting).
Source: http://www.nosolousabilidad.com
We can distinguish between two types of 'card sorting': open and closed.
In the 'card sorting' open the user can freely include the categories in the number of sets you think necessary, while in the closed, groups or sets are predefined and labeled and the user only need to put each category in the group creating appropriate.
This second type of 'card sorting' is recommended to check if a classification information is familiar and understandable to the user, while the "Open" aims to discover what kind of classification of categories would be more correct.
Test Driving The first step is to recruit participants for the test. The selection must take into account these features and profiles should be consistent with the target audience of the website. No point having young participants and recent graduates in computer engineering, when the website is intended to be used by housewives, uneducated and elderly.
The number of participants in the test must be at least 5. The time allotted for the test depend on the number of categories to sort, although indicative, for 40 categories should allocate at least 20 minutes.
At the beginning of the test should explain to participants the objective of the test, and that this is not the assessment of the participants. In addition, participants indicated what should be the criterion for grouping of the categories by similarity.
then gives users the cards with the different categories, ensuring that each set delivered to each user is disordered (shuffled). Delivered to all users cards with a predefined can influence the form of management by the participants.
For 'card sorting' open type, once they have completed the test participants may be asked to give names to each of the groups created.
There are, likewise, software applications with which to carry out the virtual test, as EZSort or WebSort.
qualitative analysis of the results
This type of analysis can be performed when the participants as well as the number of categories, are not very numerous.
consists, first, to observe individually to each user during the test, and record all aspects of how each user organizes the cards, what problems have to do the job, which brings immediate and intuitive categories and questions about what else, what questions do during the test, etc.
then discusses "eyeball" the groups created by the participants, noting which categories most closely related to what categories according to the user's mental model.
Quantitative analysis of the results
This type of analysis is the statistical processing of data, and the subsequent "summary" of the results through graphical representations that facilitate their interpretation by the information architect. It is, therefore, a proper analysis to test a large number of participants and categories to sort.
first create a table of co-occurrences in a spreadsheet. In this one, with many rows and columns as the number of different categories, indicate the number of times each pair of categories have been placed in one group, resulting in a symmetric matrix as follows.
Then, on the co-occurrence table apply dimensional reduction algorithms such as clustering algorithms and multidimensional scaling (MDS), whose function is to simplify the relationships between categories to a number of dimensions easily interpreted by visual inspection (2D or 3D).
For the application of these algorithms have to have a statistical software package that allows it, such as SPSS or Statistica.
As can be seen in the figures, the result of applying clustering algorithm is a dendrogram, and the MDS geometric representation of the categories, separated according to their own di-similarity between them. Dendrogram (clustering), MDS (Multidimensional Scaling). Conclusions
technique 'card sorting' technique is a useful categorization of user-centric content, providing a double tool: to help decision making in the conceptual design phase (open card sorting) and to evaluate a particular organization of categories in usability evaluation stages (closed card sorting).
Source: http://www.nosolousabilidad.com
0 comments:
Post a Comment