2013-14 Term 1 G2 TeamYOLO A10 Conclusions
Before we proceed to our conclusions, there are some things that we would like to note:
- The experiment was conducted among 30 participants only. A bigger group would have been more ideal, but a sample size of 30 is generally regarded as adequate to simulate a normal distribution.
- The experiment was conducted among SMU, NUS and NTU students only. This limited our randomness of our sample from our targeted audience.
- We have ensured that within each block of SMU, NUS and NTU students picked for the experiment, the students were randomly selected.
- We have ensured that each participant engages in the same tasks as the other participants.
- We have designed the experiment such that each participant uses both proposed layouts of the app.
- The experiment also splits the ordering of the layout into 1-2 and 2-1 i.e. (Design X then Design Y, or Design Y then Design X), and that half of the participants will take the first ordering of the layout and half will take the second ordering of the layout.
Design X was the original design for BillSplitter prior to A10. It allows renaming of faces, but each of the faces are visually identical.
This resulted in:
- Better accuracy in submitted answers
- Slower task completion time
Design Y was a new design for BillSplitter that we wanted to test. The use of multicoloured faces was designed to help people differentiate between the people to split the bills for. However, it does not allow users to rename faces. Users will only see numbered faces.
This resulted in:
- Worse accuracy in submitted answers
- Faster task completion time
Why not combine the best of both worlds? Combined with the responses from the Qualtrics survey, it gives us reason to let users have the option to skip the renaming process, and yet view the multicoloured faces!