Uncertainty in Data and Trends
|Above copied from the NY Times|
Our students are currently in distance learning not because of a well designed controlled study. Students are in front of their laptops because of a pandemic. These are not normal times so how well distance learning performs maybe due in part to the pandemic. There are no live soccer matches to watch, for example. There is no recreational soccer league. Most parents are staying at home. These are certainly not the conditions from which we should draw conclusions on which one is better: distance- or in-classroom- learning. On top of these, we need to work on an acceptable sample that can represent all students in the entire population. We need statistics.
Correctly applying statistics, however, is not a guarantee for universality. In fact, when statistics is used, uncertainty is highlighted. It is quite ironic that statistics and probabilities are like conjoined twins. And when we present statistics, we often favor what helps us in proving a point. Take, for instance, graphs of reported COVID-19 cases in the United States as presented by Rachel Maddow at MSNBC:
|Above copied from Rachel Maddow (MSNBC)|
It is true that displaying only the graph for New York can lead us to think that there is a real down trend in cases, but it is equally misleading to show a graph for the rest of the United States to represent better the current status of the pandemic in the country. This is likewise not an accurate representation of where the country stands. The following are graphs for each state from the New York Times.
Without doubt, drawing graphs for each county within a state likewise yields different pictures. These are so different from each other that statistics actually tells us that people are indeed facing different realities.
Furthermore, showing the number of reported cases for each day does not give us a quick glance at how fast the infection is really spreading. This rate is actually very important in gauging how fast the virus is spreading, an essential piece of information for making public health decisions. In this case, however, there is actually a clear picture from each and every state as reported by the NY Times:
Not all states are included above but the entire table from NY Times shows that the infection rate in all states has slowed down. Of course, if one expands a state and look at the county-level, there are places where infection rates are still at a peak. For example, here are some localities in the state of Virginia:
The growth rate in cases in Galax is clearly not on a down trend yet. Thus, there is an expected range of opinion on whether communities should keep stay-at-home orders or not. This is statistics. There is uncertainty. And on top of all of these, there is really a lot of things that we do not know about this virus. The governor of New York finds, for example, that most of those admitted in hospitals in the state are already staying in their homes.
|Above copied from CNBC|
But we are not completely in the dark. Statistics tells us that COVID-19 affects the elderly more significantly. There are indeed cases involving younger individuals, but deaths and serious cases are really overwhelmingly coming from those who are 50+ years old. For instance, here are the case fatality rates from Spain, China, Italy and South Korea:
|Above copied from OurWorldinData|
The above tells us that we really need to focus on these age groups that are clearly more vulnerable to COVID-19. There is something useful in statistics. We could only hope that we do not allow the uncertainty and differences in opinions in preventing us from taking the proper action. Staying at home and keeping things closed also have dire consequences. There is indeed uncertainty but there are things that are certain.
Going back to distance-learning and in-classroom learning, there is likewise certainty. Distance-learning has not appeared only during this pandemic. It has been here for more than a decade now and yet, most of us still choose in-person face-to-face setting.