January 6, 2015

Developing User Study to Evaluate Force Interactions

HCI research, user study design


A user study was developed to further study Force interactions. In the study, we explored how various types of feedback mechanisms (visual, auditory, haptic, and teletactic) allow remote interaction with distant objects to be more immersive.

We also studied preferences between direct vs. traditional remote control and examined the importance of matching the shape feedback with the shape of the distant object being  controlled.

Experimental Design

The key hypothesis that was tested was: Wrist-worn and hand-held devices that change shape and colors make mid-air gestural interaction more immersive. The independent variables (stay constant throughout the experiment) were the device configurations, the wrist-worn controller and handheld controllers. The dependent variable (measuring this to draw conclusions) is participant reaction to the devices.  This includes things like comfort, anxiety, willingness to use in the future, etc. The control condition (will not change during the experiment) is that both devices will have  sound and vibromotor feedback enabled. The experimental condition (will change during the experiment) is that shape change and visual feedback is enabled and disabled. The design of the experiment is between-groups, which means that all participants will try all conditions of the experiment. In order to account for any bias we will randomize the order of the conditions using Latin Square ordering (within-subject design).


A total of 8 participants were recruited from the local community. Female and male participants were equal in proportion with the age range between 23-27. The participants ranged from being undergrads, graduate students and working professionals.


The user study began by providing each participant training on the 4 device conditions: wrist-worn and handheld devices with shape and visual feedback enabled and disabled. Following this the participants used each of the device conditions with 4 of the user scenarios that were created. Ordering of the device conditions with the user scenarios was done using Latin Square; half the participants had feedback enabled first while the other half with it disabled first. Following this we also explored other aspects like miss-matching shape feedback with the distant object. For example, when the participants reached for a large yellow textbook they felt a sphere instead. The participants also had a chance to drive the Sphero with the manufacturers mobile app. Upon completion of the study participants were given a questionnaire with open ended and Likert questions. Each participant was compensated $30 for their participation in the study.


Five point Likert scale questions (strongly disagree, disagree, neutral, agree, strongly agree) were placed in the questionnaire. Below are the results from key Likert questions and responses to open ended questions.

Participants were asked about which form of feedback  was important to providing an immersive experience. Auditory feedback indicating selection and de-selection of objects, inflation of the balloon, and feeling the balloon when squeezing to select distant objects were highlighted as important. Visual feedback of the local copy and the remote object/minion was considered unimportant. Some participants did not even notice the color change in the handheld/wrist-worn controllers when they selected a new minion to interact with.

Participants were also asked about their preference with regard to the 4 device conditions and the mobile app which provided indirect joystick driving control of minions. Direct control provided by the prototype devices was generally preferred over the indirect joystick in the mobile app.  Participants commented that with direct control they could move minions to specific locations with greater ease as they can point towards the final destination they want the minion to go.

We found that the participants found matching the shape feedback with the object being controlled important. When the shape did not match with the object that they were controlling the user perceived the shape change as an arbitrary signal of selection. However when the shape feedback did match they perceived it as touching the distant object.

return to top