Art is subjective. Math is a science. For most of your life, you’ve probably been pretty clear on this concept. Things like beauty and colour are held in the eye of the beholder — we all perceive shades differently, and respond in our own way to a piece of art or a stanza of music. But math is math, right? Two is always two.

*Except when it’s not.*

Many people are surprised when I say that objective data analysis isn’t really a thing. Statistics are one of those things we like to feel are concrete — especially when we’re basing life-changing decisions on them. But there’s something very important you need to understand about statistics. Your worldview gets baked right into them. The way you see the world around you can affect even the most basic math.

Let’s look at an example.

# Where You Stand Affects Your View

Saying that objective data analysis is basically impossible sounds very wrong. But watch how even the most basic math is open to interpretation in this example.

Let’s say we’re doing some research in the education sector and we want to talk about how average class size affects outcomes. *Averages, those are pretty scientific, right?*

We study three classes. Take a look at the image below and calculate the average class size.

While that probably seems like a pretty simple task, there’s a real possibility that not all of you have arrived at the same answer. (And I’m not just saying that because some of you are reading this before you’ve had your morning coffee!)

**The average class size at this school depends entirely on who you ask.**

Even though there is nothing challenging or complex about the math involved in this question, we still can’t count on objective data analysis. Why? Because the “correct” answer depends on your worldview. Let’s look more closely.

If we ask the students how many students are in a class, we get the following answers:

Now let’s ask the professors how many students are in a class.

# The Locus of Power in Your Analysis

The average class size depends entirely on whose point of view you’re taking. That is, where you put the *locus of power* (or centre of power) in your analysis — on the professors or on the students.

How often do we automatically put the centre of power in a specific place and simply assume that it’s correct? (Not that it’s necessarily incorrect — but it’s not the only option.)

Let’s look at the math.

Both answers are technically correct. The math is sound. But how does that work? The answer to the question “what’s the average class size?” depends on whether you’re a teacher or a student. And that’s why objective data analysis isn’t really a thing — because there will always be assumptions you need to make, and making assumptions removes objectivity.

# Achieving Objective Data Analysis

*So how do I make sure my analysis is free from assumptions then?*

Short answer: you can’t. Because statistics are never just about numbers, but what those numbers represent, there will always be the possibility that you are interpreting those numbers differently than someone else.

We’ve touched on this before with our story about numbers going in the wrong direction. The fact that the number of transgender murder victims is at an all-time high seems like bad news. Until we recognize that we had assumed that *all transgender murder victims were always being reported as such* — which is just not true. Once we recognize that an assumption was made there, other theories become possible:

- More transgender murder victims are being correctly identified
- There are more people identifying as transgender

*So what can I do?*

Just because truly objective data analysis isn’t really possible doesn’t mean you can’t take steps to make your results as transparent and accurate as possible. Here are some ways to ensure that you understand any biases or assumptions that may exist in your data and communicate them to your audience.

- Create a data biography to help uncover hidden bias in your data
- Carefully document any edits or assumptions you make when cleaning data
- Examine your data carefully and be aware of common statistical mistakes
- Test how other perspectives might change your results
- Remember that, even without bias, averages are not always trustworthy
- Make sure you avoid incorrect comparisons
- Take care not to communicate your results in a way that might mislead your audience
- Understand how to communicate uncertainty

Still need help? Our team of experts is happy to help you collect and analyze data or communicate your results in an honest, engaging way. Get in touch with us today.