Bad Surveying in The Fourth Grade

At the Friends School, where I attended pre, elementary, and middle school, there were annual science fairs. When I was in the 4^th grade I did an experiment to find out if organic flowers smelt better then non-organic flowers. I bought some flowers that were not organic and then took some organic flowers that had been sitting on the windowsill. I blindfolded the people who I surveyed, and let them smell each. My hypothesis was that the organic ones would smell better, and it was confirmed.

I remember surveying my dad first. After I gave him the non-organic ones and then gave him the organic flowers I introduced the organic flowers with a great amount of lavish: “now smell this one!” After that my mom told me that if I introduced either flower differently I might influence the answers of the people being surveyed. In other words, my asking the question in such a way that implicitly favored one of the answers introduced bias into the survey.

There were also several other problems with the survey, some of which I knew at the time, others I didn’t. One problem, which my mom had also mentioned to me at the time (she said that if I included this in my final project my teachers would be impressed) was that the sample size was relatively small and—she didn’t word it like this—that my results were therefore not statistically significant.

Another problem with my survey is that there may have been confounding variables as the organic flower had been grown at home and I didn’t know much about the store-bought non-organic flowers.