Phase 4 | Evaluative Research | 04.15.2018
In this week, we started to sum up the findings and insights from our evaluative research phase, and began to finalize our wireframes for the project. Then, we put these into slides for Monday presentation.
We are focusing on building a networking platform for incoming students that leverages AI to facilitate two-way learning through person-to-person conversation.
We systhesis a diagram from our early research that shows how international students’s perceived competence varied over time before and after they come to US, and it is clear that right after they come to US within 3–6 months, most students would experience a severe gap. Thus, our goal is to smooth that gaps, and ease the transition process for international students.
So we believe the intervention should take place as soon as possible, especially when international student just receive their offers, particularly during their Pre-Arrival Excitement. By providing information while students are at a pre-arrival emotional high, we can ease the transition upon arrival.
After conducting research observing incoming students’ conversations, we found that the the break down we assumed doesn’t happen so frequently, which is why we pivot to the design direction that takes more proactive intervention.
The diagram Need vs Want helps us frame our problems better, and facilitate us to decide which parts we should be more proactive and give more prompts, and which parts we should let users to chat with each others. We want to focus on the things that are highly needed while doesn't seem to be wanted. These parts of things or we say knowledge are exactly the stuffs that students don’t know they don’t know. So we have to actively prompt these type of knowledge to have them be aware of it.
Prototyping and Speed Dating
After making several prototypes, we began to do another speed dating with users to help us which is the best way for conversation intervention, interest matching, as well as marking the confusion.
Here are the feedbacks we got from 5 speed datings.
For the matching function, users like to not see the other person’s photos or names, because they either think it is irrelevant or possible to make them bias. But they do want to know what their interests are.
So we replaced photo with illustrations, but still leaves some basic bio and interests on the page. We will provide 5 options each week, and according to their research we also found that a limited number of choices makes uses easy to choose, and feel having the control, while unlimited options makes it overwhelming.
For the conversation intervention, most people like it not too intrusive. They enjoy it to be subtle, and able check it and also ignore it. They sometimes feel it annoying if it keeps popping up notifications.
We decided to use suggestion chips for intervention, the system would automatically generate several related knowledge topics if it’s needed. The suggestion chips are mainly a couple words, so it would cause a heavy cognitive burden. If a user clicks one of the chips, a prompt will automatically generate to their chat box, but they can still continue editing it before sending it.
Ultimately, we allow users to mark the confusion or save the dialog for later review. The user can easily select the sentence or words and hold it for 1 second, a small bar would popup. If the user choose to mark it confusing, a question such as “What do you mean by….” would be added to chat box. In this way, the user save the time for typing, and the system can also identify what part is confusing, so that it can utilise that data to provide better service later.
