The US appears focused on math education, especially when it comes to curing American innumeracy. Yet we are in an age when enormous new data sources are now available in academia, business, and government. Elliot Schrage, vice president of communications at Facebook, argued that our kids should study “statistics, because the ability to understand data [will] be the most powerful skill in the twenty-first century.” Hal Varian, chief economist at Google, noted “I keep saying that the sexy job in the next 10 years will be statisticians. And I’m not kidding.” The National Science Foundation reported that the employment rate was highest (at 99%) for PhD mathematicians and statisticians compared to other scientific fields. Therefore, in addition to math education, shouldn’t we be focused on statistics education?
The lack of American quantitative literacy has been fairly well documented. To give an example, in 2003, only one-third of US adults were able to interpret a graph to solve a basic percentage comparison problem on the National Assessment of Adult Literacy. In many international comparisons, including the 2013 OECD numeracy ranking, US adults were near the middle or bottom. Despite these unfortunate trends, professor and author of Higher Education, Andrew Hacker, who is currently writing a new book on numeracy, told me: “People can be very numerate when they want to be. Look at the way they handle baseball statistics. Or build fantasy teams. Or how supermarket shoppers work out the highest-quality-lowest-cost cart. Plus playing those grid numbers games.” So perhaps numeracy depends, in part, on context.
It is important to clarify that quantitative literacy, as it has been assessed is not the same thing as statistical literacy, although there are some overlapping aspects such as the ability to interpret numerical summaries and graphical displays of data. In fact, now is a time when data journalism is on the rise through new sites such as The Upshot, FiveThirtyEight, and of course Quartz. Nate Silver has pointed out: “More than 80% of American adults spend at least some time with the news each day.” Yet what fraction of that 80% can actually understand data and appropriately interpret the data journalism increasingly appearing in their newsfeeds? The answer is, we don’t really know.
According to a report outlining a Pre-K through 12 statistics education framework endorsed by the American Statistical Association (ASA), “A statistically literate high-school graduate will know how to interpret data in the morning newspaper and will ask the right questions about statistical claims.” Rebecca Nichols, ASA’s Director of Education, told me:
“Sound statistical thinking and reasoning takes time to develop. As mathematical thinking is developed over years of study starting in the early grades, so should statistical thinking. The current Common Core standards being implemented in the majority of the [United] states do not have much statistical content in the early grades and then a large amount of statistics content starting in grade 6 and the middle grades. Because statistical thinking takes time to develop, I am concerned that it will be difficult for middle school teachers who have not been prepared to teach statistics to tackle such a large task of introducing students to so many statistical concepts in those middle years.”
Just like mathematical thinking, perhaps we should be laying down the foundations of statistical thinking starting in kindergarten, as all education builds over time. Harvard social scientist and statistician Gary King told me: “Everyone should learn about ‘inference’ (using facts you have to learn about facts you don’t have) as it is a crucial part of life and an important way that collections of people learn more than anyone could. An education in statistics is merely a formal way to do that, so it certainly seems like a good idea.” He adds that “whether a particular high school or elementary school has a teacher prepared to give that formal training is another question.”
Here are some starting benchmarks from the ASA (pdf) on what it means to be statistically educated (see the full list here). Students should believe and understand why:
- Data beat anecdotes;
- variability is natural, predictable, and quantifiable;
- random sampling allows results of surveys and experiments to be extended to the population from which the sample was taken;
- random assignment in comparative experiments allows cause-and-effect conclusions to be drawn;
- association is not causation;
- statistical significance does not necessarily imply practical importance, especially for studies with large sample sizes;
- finding no statistically significant difference or relationship does not necessarily mean there is no difference or no relationship in the population, especially for studies with small sample sizes.
A kindergarten statistics curriculum, as directed by the ASA framework, might look like the following: Given that children are surrounded by data, such as student preferences (e.g. favorite types of music) and measurements (e.g. number of books, heights), there are many opportunities for young students to be introduced to “data sense—an understanding that data are more than just numbers. Statistics changes numbers into information.” Teachers could have students collect data on other students in the classroom and then think about questions to ask about the variability in that data and begin to sow the seeds of statistical thinking.
The challenges to promote widespread statistical literacy in the population are many. There are educators at all levels—including those teaching future teachers—who are teaching statistics without any formal statistics training. Also, teacher preparation programs do not often prepare teachers appropriately to teach statistical literacy, and the quality of statistics assessments need to be more consistent and improved in that they often do “not test conceptual understanding and statistical reasoning,” says Nichols. And given that in international comparisons in science and math, the US is in the middle of the pack or lower, which is troublesome for statistical literacy. Along with quantitative and science literacy, which is necessary for all STEM fields, these skills are likely linked. Other countries may already be preparing their future workforces with better basic competencies than the US. Unfortunately, we don’t know what fraction of the US adult population is statistically literate according to the ASA benchmarks. But we do know that developing statistical literacy takes time and that it is more important than ever today and in the future. This is why we must start teaching it as early as possible.