The danger of statistics out of context: how numbers don’t tell the whole story



Statistics unaccompanied by the appropriate context can inadvertently cause harmful misconceptions. Conversations surrounding these statistics can leave out the students being directly affected and result in them being overshawdowed by the data.

People can lie, but numbers cannot. Statistics are everywhere, and they’re meant to convey impartial information in an easily digestible manner. Visuals, charts and labels allow the viewer to process a story much faster than written data. But what happens when a quick glance at a statistic becomes a false perception of the truth?

The Cypress has published various articles with statistics of all kinds throughout the years. However, statistics do not always tell the whole story. A recent example is an article published in late 2022. The article covers a school subcommittee meeting that took place on Oct. 18, showcasing multiple datasets with 2021 Massachusetts Comprehensive Assessment System (MCAS) scores and attendance rates at the top of the page, which were all used by the subcommittee in their discussion. The subcommittee meeting and subsequent article were largely focused on racial and socioeconomic disparities in the provided MCAS scores.

The statistics used by the school subcommittee were official and provided by MCAS. However, because of the lack of context given, the factual information isn’t enough, according to math teacher Lisa Rodriguez.

“Without being able to go into the background of what’s occurring at each school and defining all the variables, you’re limited in your storytelling and now people are going to read that and think, ‘oh my goodness, these parents need to care more about their children’s education,’ which is not the case,” Rodriguez said.

The Town of Brookline Updated Disparity Report March 2022, published by the Office of Diversity, Inclusion, and Community Relations, offers contextualization as to why socioeconomic status plays a role in test scores by researching the disparate access to educational resources.

“Research indicates that school conditions contribute more to socioeconomic status differences in learning rates than family characteristics. In 2014, the high school dropout rate among persons 16–24 years old was highest in low-income families (11.6 percent) as compared to high-income families (2.8 percent). Those from higher social-socioeconomic backgrounds tend to be more successful in developing career aspirations, and are generally better prepared for the world of work because they have access to resources such as career offices, guidance counselors, better schools and familial experience with higher education. Given its centrality to future success, it is imperative that Brookline focus on increasing the educational resources that lead to better outcomes for low-socioeconomic households,” the report said.

Showing the disparate data without the nuanced explanation of how socioeconomic status can impact test scores, may convey the impression that lower income students are less intelligent, which is simply not true.

The race breakdown in Brookline MCAS scores is also flawed, which could skew the data, according to Rodriguez.

“Only talking about people who are solely African American or Black manipulates the data. I remember [Senior] Jacquovia [Higgs]’s [DoRRS] speech, where she talks about being one of a hundred Black students, but there are more. She was only talking about kids who identify only as [Black]. So let’s say that MCAS is saying only 10 sophomores are fully African-American or Black, and those 10 kids don’t do as well. That data is totally skewed compared to like, if you get like 300 white kids to take the test and 10 white kids don’t do as well. Their score is gonna be way higher. So that already kind of messes it up a little bit,” Rodriguez said.

Additionally, in the MCAS data presented in the article, students with disabilities were presented in a table with the racial categories with no further explanation. There isn’t clarity on whether students with disabilities were taken out of the data groups for their subsequent races in order to include them in a separate category, or on why racial groups and students with disabilities were represented as separate categories in a single table.

The article also mentions that the school subcommittee said the high school provides M7 bus passes for “socioeconomically disadvantaged students.” While the school district does give students M7s, the context in the article is incomplete and can lead to misunderstanding who uses M7s.

“The Brookline Public Schools provide free busing for students K-8 who live more than 1.5 miles from school, as well as for students with special needs and students who use the federal school lunch program,” according to the Town of Brookline Updated Disparity Report March 2022.

Furthermore, the school subcommittee members talked about “reinforcing the importance of school” to improve disparities in attendance. Without discussing the varied and complex reasons why students of color may have lower attendance rates, offering “reinforcing the importance of school” as a solution makes it seem like students with lower attendance rates simply do not care. Systemic racism cannot be solved by “reinforcing the importance of school.”

Marika Alibhai, a statistics teacher at the high school, said that any presentation of data requires context. She teaches lessons on interpreting and skewing data.

“The very first unit we do is learning sample design and design of experiments because there’s no point in analyzing bad data, right?” Alibhai said. “How do we collect good data? There’s the five W’s and the H, if you will. When was it collected? Where was it collected? Why was it collected? Who’s included in the data? And then how was the data collected? Knowing all those things are really important. Is this number worth repeating or investigating?”

There are a couple of things that can skew data, according to Alibhai.

“We teach in stats all the time that when you take a list of numbers and make a graph, if you change things that are referred to as the bin width or the bin alignment in a histogram, you can make it look a little bit different,” Alibhai said. “If you want it to look a little more symmetric, you have to play with how the graph is set up to make it look a little bit more symmetric. If you want to make it look a little bit more skewed, you can play with it to make it look a little bit more skewed. It’s all valid graph-making. You can make it look a little bit different, and neither one is wrong, but it does tell a slightly different story.”

The Cypress covered a School Curriculum Subcommittee meeting where the story told, without context, was that Black students are doing worse, which can lead to the conclusion that Black students are less smart, according to junior Amara Ukomadu.

“When you’re just putting out information saying that these students are getting worse grades and they aren’t performing in classes, you also have to talk about the impact of why that’s the case. You have to mention both sides of the story,” Ukomadu said.

Statistics are incredibly powerful tools, but they can be manipulated to tell many different stories. It is important to realize that not all statistics are just the facts. A statistic, just like a single point of view, might not always be the most authentic way to tell a story, and these numbers are not without their victims.

These statistics can not only spread misinformation, but also have real impacts on students, according to junior Bruno Amaral.

“I think if you’re telling people of color, ‘Hey, you guys aren’t doing as well [as white students],’ then they think it’s a norm to not do well. It’s not what we want. We want to encourage students to try their best no matter what,” Amaral said.