Sample...sample....sample.....

What to do, what to do....that was my thoughts on this weeks assignment. Taking a look at the different types of sampling in the text gave me some ideas about my research project. I'm not sure about you guys, but most of my research is being conducted at my place of employment and not at the school. Taking this into account, I tried to look at what would give me the best data that could be used for a research firm not only to make a decision, but to make a more informed decision than previously available. If that were the case, I would choose quota sampling. I liked how they described in the book that an interviewer would be assigned 100 interviews and he would disperse those interviews among the intended respondents. Although my research is focusing on hourly associates that work night shift in a plant environment, reading the breakdown described in the book actually got me thinking that maybe I need to look a little further into my research. This would include mostly hourly associates, just like the book, but could also include (in limited numbers and on a much smaller scale) Managers at the plant, team leaders, trainers, visitors to the plant that will be working night shift, and so on. Of course the main focus would still be on the hourly associates, because that is where the benefit of my research will be most useful, but it will also include a broader scope of individuals that could be captured underneath the same umbrella.

-DS

Bad questionaire....

Three things I found wrong with the questionaire are face validity, criterion validity and construct validity. When I took the questionaire, I was thinking "what the hell is this thing talking about?". The questions were all over the place. If we look at face validity, we can see that if you try and look at the data, you don't really know what the questionaire is supposed to measure. Also, the criterion validity can't be reached because with the questionaire being so broad and covering so many different topics, it is impossible to correlate with anything. When we look ath the construct validity, it is kind of the same thing as we see with the criterion validity in that it is so broad that you can't relate to a hypothesis. This would be a data miner's nightmare.