Hi again - First off many thanks for attempting to address my problem so fully, and secondly, sorry for not providing any feedback as yet. I posted this just before going away for the weekend, fully expecting no one to bother responding, and have only come back into email contact now.
Let me clarify: I am doing a Masters in Human Computer Interaction. For my dissertation I have designed a new method for achieving look-free dialling from a mobile phone, that i want to compare against voice activated dialling. Before performing the main body of the research, however, (the comparison against voice activation), i wanted to make sure that i had configured my system in the optimal way. I created 8 slight variations, and want to know which one is the best (fastest), which i will then go on to test against voice activation. Before i go on, let me just put this into context: I am very short of time and intended these to be 'pilot' tests before the main body of research, and therefore they were done very quickly and not in line with a thorough experimental design or control. I have a hunch which configuration is best and it is not essential that i do this anaylsis before continuing, however i just wanted to add a bit more 'meat' to my justification for selecting config x over config a, b, or, c etc.. I therefore just want a 'rough and ready' analysis of the data to see whether one condition does actually stand out as faster than the rest. I do not want practice effects to be taken into account. Ok, here are the details: I assembled a phonebook of 10 names. In each condition, every name was dialled in a random order of presentation. Time for a successful dial was recorded, being the time from the presentation of the name, to the system ackonwledging selection of the correct name by the user. I was the only participant in all conditions. I am now left with 10 times in 8 conditions and a mean dialling time for each condition. As far as the little questionnaire posted by D. Burrell (i think?) i would say that the discussion between you two was for me somewhere between confusing and incomprehensible - i really have never understood stats, but once upon a time i did at least know which buttons to press on the stats package i was using (then it was Minitab). Anyhow, like i said my main aim here is to just get this done as opposed to educate myself, although i do appreciate the time you guys put into the explanations. I hope this clears up my problem, and describes the type of answer i am looking for adequately. Any help you could give with actually using SPSS to do the test you recommend would also be greatly appreciated as i am not very familiar with this package. Thanks again, DN . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
