Statistics projects of interest

Each term there are a few student statistics projects of interest. Data samples are usually convenience samples, hence extrapolation is problematic. At best these provide anecdotal glimpses into the system being studied. The students are in their first and probably their only statistics course - an introduction to statistics. Four of the seventy reports contained data that I found interesting.

One student studied alcohol sales at a small family owned and operated local neighborhood store. The student found that 53% of their sales was of the highest proof alcohol available in the store. The top seller was not the cheapest of the high proof alcohols, just the strongest. Coupled with their own knowledge of their neighborhood, the student concluded that purchasers were "most likely drinking to get drunk, not drinking for flavor or an addition to a meal."

In a related vein, another student looking at cigarette sales in a local family store found that the top seller garnered 55% of sales among eight brands offered. When customers were asked why they preferred this brand, the student quoted respondents as noting that the brand offered the strongest "high." Based on a couple of charts, the top selling brand does appear to be the one with the highest nicotine content. Clearly the cigarette is only a vehicle for imbibing nicotine.

While the above two studies suggest groups that seek the strongest drugs available to them, another study indicated a steady drop in nightly evening church service attendance despite the season of Lent usually seeing an uptick in church attendance. This study was particularly interesting as a decision was carried out earlier in the year by an order of the church to forcibly retire a number of popular church leaders. The denomination, not known to be gaining ground globally, may have unintentionally acted against its future best interests in this action. The church did not explain its actions to its parishioners, and now the church is apparently losing a key constituency: the regulars who come to nightly services.

The fourth study that caught my attention was a study that hints at the possibility that the farther a student's home is from their elementary school, the lower the probability that they will eventually attend college. The study was conceptually interesting and was carried out to the best of the ability of the student. There are, however, so many confounding factors that the results can only be treated as "deserving of further study." The most fundamental problem has to do with the distribution of students - fewer may be living farther from the school. The geographic units used in the study are probably not population equal.

Despite the potential problems, the differentials appear to be stronger than the population differentials. On a very rainy island where children walk to school, there may be a distance effect on students in elementary schools. Obviously there are dozens of intervening variables. One would want attendance versus distance data broken down by individual student and distance covered.

If this effect is real, however, then there is an interesting result. For the school studied, children come from multi-generation ancestral home sites. Thus their parents often grew up in the same location. These are indigenous peoples who do not move from location to location in the manner seen in western nations. One could postulate that this effect also plays out over the generations: those who live near schools are more likely to go to college, value an education, and pass it along to their children who also grow up near the school.

Hence the school has local, state, and church leaders all living very close to the school, while areas further from the school produce fewer local, state, and church leaders.

If, and this remains a big if, if there is a distance disadvantage, then actions that reduce this effect should be beneficial to the larger society. Transport, such as buses, is too costly for this community, let alone expanded to a state wide initiative. A larger number of smaller schools closer to where children are raised is also probably not affordable nor practical. Solutions to enhancing the education of the most distant students are not obvious to this author.

For the particular school studied, of interest is that the lowest rates of college admission were seen for the students who are indeed farthest from the elementary school, but whom are closest to the national campus of the college. Some of these students are actually closer to a school in the neighboring municipality, but the municipal boundary is also a cultural boundary of a sort and families choose to send children to the further elementary school. The college is, however, probably not in a position to be all that helpful to these children. The college is small, budgets are way beyond tight, and socio-cultural complications would lead to bedeviling details.

Of the other studies, most are of routine and pedestrian matters, hours of sleep, number of betel nut chewed, number of sodas consumed per day (maxing out at seven!).

Popular posts from this blog

Box and whisker plots in Google Sheets

Creating histograms with Google Sheets

Traditional food dishes of Micronesia