Theme 3: Reflections

I have learned to clarify certain notions during this week, which include mixed research, validity and reliability.  These are rather often used within studies of human-computer interaction, but not that often equally well explained like during this theme.

Obviously the theme of the week was quantitative methods but in the end I felt that the paper about emotions didn´t really fit under that category. For me quantitative methods should be briefly, clearly explained and not involve abstract, complicated, subjective terms like this paper had. In addition this paper used a qualitative method as well so a mixed research was inevitable.  Would the study have been better and more comprehensible if it only contained one method of the two, in this case just a qualitative method? This question is kind of contradictory to what I said about mixed research earlier but again it´s just another evidence of how difficult it is to choose the most appropriate method or methods.

The lecture about quantitative research and the picture of the four dartboards helped me a lot to separate the terms validity and reliability from each other. Usually I mix them up quite often. Apparently as long as you hit the target with a particular acceptable range of standard deviation, the results are said to be reliable. While validity on the other hand requires you to exactly match the optimal hit. Hence, with that kind of definition, every study lacks of validity more or less according to me, because you can´t hit the bull´s eye frequently, can you? Certainly it depends on how to define what´s valid in a particular case. Concretely speaking one dartboard´s total area might be another dartboard´s or another study´s bull´s eye. So to speak, the main question, the method and other important aspects create the area of the bull’s eye. A study with a transparent question and method will most likely have a greater area of validity.

The lab exercise and the SPSS gave me useful software for how to match and compare multiple variables with each other, which could be quite messily only with excel. It seems to be a very handy tool, if you for instance want to know exactly how many participants in percentage chose the second alternative for all the questions within a survey. However I don´t think this tool makes the survey more reliable, It just explicates and emphasizes particular combinations of answers.