Girls have more left-overs than boys on average in the school canteens.
We choose this topic, because the results can infer how much food the dining halls leave every day, and reflect how frugal the students are and the quality of food offered by different canteens. At first, we want to find out whether boys like sports better than girls in UNNC. This topic was rejected because there are more girls than boys on this campus, and girls have menstrual periods, during which they cannot do much sports. The target populations in our final project are boys and girls in the first-floor and the second-floor canteens. We used quato sampling because it is a practical, cheap and quick method of obtaining samples based on the fact that a sampling frame is not available. To obtain the samples that can reflect the population as well as possible, we chose samples as large as possible and extent the survey time. 30 boys and 30 girls in the first-floor canteen, 30 boys and 30 girls in the second-floor canteen are included in final samples. We cannot use a larger sample because time is limited and we can just borrow one electronic loadcell scale so that we could only just do survey in one of the four exits. Our way to deal with absent or ambiguous is to ignore them. For example, some left-overs containing drinks are not involved in our statistics.
We need to train our observers to ensure that the left-overs are poured out as much as possible and that we use the electronic loadcell right. We have three main problems before we started. Firstly, we wanted to use scales with a hook, but the number it shows can not be read easily, then we chose an electronic loadcell scale instead. Secondly, we wanted to use plastic bag, but there were still some left-overs stuck to it, so we weighed the plates directly. Thirdly, our time is limited, so not all the population in each group had the same possibility to be chosen. Then we spent as much time as we...