Sunday, September 9, 2012
The Right Experiment
Me, always wanting to get an experiment right and prove some weird theory that pops up in my mind, I didn't realize that there was so much to getting an experiment right, as well as getting accurate results. Not only did I learn that I can't find the answer to why something happens, but there is so much to reconsider when doing an experiment. We have to think of everything that has the potential to ruin an experiment, like for example, when I am going and doing some sort of environmental experiment, and it's supposed to be a nice clear day, and then suddenly, clouds come in and it's thunder and lightening. There goes my experiment! So I kind of no longer have an interest in any environmental experiment from that, and so why not do some experiment that involves statistics! Where can I go wrong from there? Why don't I mail out a personal survey to a hundred people asking them questions about how much energy they consume? Well, it's not that simple, because I have to give everyone an equal chance of getting the letter. So why not do random assignment? Everyone get's an equaled chance of being chosen! One thing that I have to realize when sending out a survey is my word choice and experimenter bias. Two great things that have a great effect on whether or not if I am to get accurate results. When it comes to communication, your word choice is very important. When asking a certain question, you have to consider if it was a biased question. When trying to get results, and you ask someone a question, it's about phrasing the question, "How many people do you think consume just as much energy as you do?" When I ask that question, it's not too accurate, but if I phrased the question in a different way, "How many people do you know of that consume just as much energy as you do?" Just by the switch of words, I just avoided getting statistics that could have been very off! Although, I don't think I can be too biased with a statistics experiment, I have to take into acknowledgement Experimenter bias. When doing an experiment, everything I do for that experiment matters! When reading Isaac's post about Pampers, there is so much I could do, even though I don't realize it, it changed the results! With doing things the correct way, without experimenter bias, random assignments, enthusiasm, and many other things taken into consideration, I could do the RIGHT experiment!
Subscribe to:
Post Comments (Atom)
Great post Eli. Just to further everyone's understanding of experimental bias, I would like to share another example. My example is going to me in regards to religion. If I were to do an experiment to find out the world's views on the 10 plagues, as a Jew I would word my questions differently to get the person to answer them in the way that I believe. For example, instead of asking, "Do you think that science played a factor in the 10 plagues?" I would ask, " Do you think that the 10 plagues happened because of just science?" Obviously, nobody believes that the 10 plagues happened just because of science. Rather, some people believe that science had something to do with them. Most sensical people believe that God did the plagues. By asking people questions in a certain way, I could get the results that I want: that it was God who made and carried out the 10 plagues.
ReplyDelete