deals with the statistical testing of data from a preliminary usability study. No calculations, but you should re-familiarize yourself with how to interpret the results of a t-test.

A usability study was conducted of two voting machines manufactured by different companies - DRE and OptiScan. Users were asked to vote a simulated ballot. A statistician performed the appropriate statistical test on the efficiency measured in time on task. draw a conclusion as to which voting machine has better efficiency and why.

Note that you are only concerned with efficiency, but not effectiveness nor satisfaction. While usability is multidimensional dealing with effectiveness, efficiency, and satisfaction, you currently only have efficiency data and that is the only data to be considered.

Here is the memo that the statistician left for you before leaving for an out of town family emergency.

I did a two-tailed unpaired t test. The P value is less than 0.0001, and this difference is considered to be extremely statistically significant.

The mean time to vote using the DRE machine as 429 seconds, and the mean time using the OptiScan machine was 633 seconds.

The number of users for the DRE machine was 42, and the number of users for the OptiScan machine was 51. but the number of users is still sufficient for the t-test, and so this does not impact the conclusion.

Lastly, the completion rate (number of users who completed the sample ballot) was 42 out of 42 for the DRE machine (100%), and was 50 out of 51 for the OptiScan machine (98%). Both are really good completion rates.