Given that the data consist of two groups, viz. Users and non users, t test would be more appropriate. Since the two categories don't arise from the same sample - paired t test can be straight away ruled out. Now, independent t test would be most rewarding. Since two samples are used, it would be worthy to go with independent group t test (as known in Stats) or independent samplessamples (please note the plural) t test, if you stick on with SPSS. Setting the hypothesis is crucial. I believe the researcher wants to prove that the difference between means of two groups us equal to zero. Well, then go ahead. Do not worry, even if your null hypothesis gets rejected for some or many variables.