Eyes Free Messaging

Eyes Free Messaging was a user study conducted to evaluate two features: receiving (Feature 1) and sending (Feature 2) emoticons. Each feature included two distinct methods for completing the task. The primary objective of the study was to determine which method was most accurately recognized (received) and input (sent) by users.

Link to published video: https://youtu.be/ZgILDekW4UM?si=9c8x2up5Ws7gOUmo

Olufunmilola Obielodan work

The Study

Eyes Free Messaging evaluated two features of messaging, receiving emoticons (Feature 1) and sending emoticons (Feature 2). In this user study, participants were blindfolded and unable to rely on audio cues. All participants experienced both features and both sets associated with them. To reduce bias and fatigue effects, participants were randomly assigned to begin with either Set 1 or Set 2 for each feature.

For Feature 1, participants began with a training phase for Set 1. They were introduced to each vibration pattern in a fixed order, and the entire set was repeated twice. After training, the testing phase began. Participants were presented with each of the six emoticons five times in a random sequence. A pause of approximately five seconds was given between each vibration. As they felt each vibration, they verbally identified the corresponding emoticon. This randomized set was presented once. The same process was then repeated for Feature 1 Set 2.

Feature 2 followed a similar structure. Participants began with a training phase for a set of gesture-based inputs and their corresponding emoticons. As the study was conducted with a blind fold, we verbally described each gesture. Participants then attempted to draw each gesture on the phone screen as it was described. After training, we verbally instructed participants on which emoticon to send using the gesture input. These instructions were randomized as well and each of the six emoticons was sent five times. This process was repeated for Feature 2 Set 2.

At the end of the study, we will ask participants which of the two sets they preferred for each feature and why. (This is different from the accuracy test that is being conducted by the study)

Set Characteristics

We chose to group our sets for each feature by similar defining characteristics. For feature 1 our sets are comprised as follows:

Set 1: Frequency

Set 2: Number of Buzzes

Similarly for Feature 2 our sets are comprised as follows:

Set 1: Drawing

Set 2: Number of Taps + Swiping

Hypothesis

We hypothesized that for Feature 1 (receiving vibrations), Set 2 would result in lower user accuracy. There is an increasing number of buzzes used for each emoticon in Set 2, which we believed could cause users to lose count beyond the third vibration. As a result, we expected higher error rates for the final three emoticons in that set, while the first three were likely to be identified more accurately. In contrast, Set 1 was designed to make each emoticon vibration more distinct from the others, which we anticipated would help users more easily recognize the emoticons within a short time frame. Overall, we expected Set 1 to yield a higher average recognition accuracy.
For Feature 2 (sending via touch gestures), we predicted that Set 2 would lead to higher accuracy. The gestures in Set 2 were simpler and less prone to being misinterpreted by the phone's input system, while Set 1 relied on users drawing more specific shapes. Though this may potentially lead to recognition issues, we believed it would feel more intuitive initially, since it was based on familiar shapes and concepts. As a result, we anticipated that Set 1 might perform better early on, but Set 2 would ultimately surpass it in accuracy due to its lower learning curve. In summary, we expected Set 1 to have lower system accuracy and Set 2 to show higher user accuracy over time.

Results

coming soon