The purpose of this master thesis was to evaluate the effect of different instruction types in a remote usability test with predefined task assignments where testusers were set up as single-user and pairs using think a-loud technique and constructive interaction.
The test setup was conducted as a remote asynchronous evaluation on an online mail and communications system. 54 teachers in the Faroese public school system were placed in two major groups as 18 single users and 18 pairs and then again divided into three minor groups getting different instructions types. The pairs were furthermore grouped into two groups of males, 2 x females and 2 x combination male-female.
The results showed that single users identified more usability problems than pairs in all instruction types. According to instruction types the results showed that both single-users and pairs that recieved a conceptual definition of usability problems identified more usability problems than those who were instructed by means of examples. Furthermore, the results showed that the gender combination of pairs did also have an impact on the results recieved.