VIDUR MADHAV
EXPERIMENTS ON UI INTERFACES THROUGH
Eye Tracking & Statistics
Interface I
Interface II
Primary Hypothesis
Click through Rate | Time to Click | Dwell Time | Return Rate
Click through Rate
Null Hypothesis The click through rate I get on my Version I will be equal to the click through rate on my Version II.
Alternative Hypothesis The click through rate I get on my Version I is going to be greater than that of Version II, because the former has larger pictures which will attract people more.
Time to Click
Null Hypothesis The time to click I get on my Version I will be equal to the time to click on my Version II.
Alternative Hypothesis The time to click I get on my version I is going to be greater than that of Version II, because version II has even column structures which makes it easier to compare taxi services
Dwell Time
Null Hypothesis The dwell time I get on my Version I will be equal to the dwell time on my Version II.
Alternative Hypothesis The dwell time I get on my Version I will be greater than the dwell time in version II, because the user’s eyes would need to travel vertically multiple times on the screen since it has two sections on the top and two sections on the bottom.
Return Rate
Null Hypothesis The return rate I get on my Version I will be equal to the return rate on my Version II.
Alternative Hypothesis The return rate I get on my version I will be greater than version II, because version II has all the content stored on the top half of the screen, and users are not required to move their eyes multiple times. Users will return on version II more than the former because it takes significantly less effort.
A / B Testing
Data collected over 120 unique visits to the interfaces: https://still-spire-33534.herokuapp.com
Version I Calculated Data
Click through rate 82.85%
Time to click 5945.55ms
Dwell time 22465.5 ms
Return rate 20.68%
Version II Calculated Data
Click through rate 78.57%
Time to click 7263.1 ms
Dwell time 7530 ms
Return rate 27.27%
Eye Tracking on the Interfaces
Understanding the user's behavior through the visual aesthetic on the screen. Looking at the time spent and verifying the statistical analysis conducted in the A/C Test.
User I
User II
Statistical Test & Analysis
1. Chi Square Test: Click through Rate
2. T- Test: Time to Click
X = 0.1816
Degree of Freedom = 1
Probability (p) = 0.05
Since 0.1816 is not greater than 3.84, it is not statistically significant
The p-value corresponding to 0.1816 approximately 0.99 giving us roughly 99% probability of the statistical significance.
T = 0.619
Degree of Freedom= 45
Probability (p) = 0.05
Since 0.619 is not greater than 1.68 then it’s not statistically significant.
The p-value corresponding to 0.619 is more than 0.90 giving us > 90% probability of the statistical significance.