by Caryl Teh
Data is everywhere – a statistic or graph in a news article, the nutritional information in the food we eat, a sign-up form for a new membership, the customer reviews of a product we’re looking to buy, and the cookies on the websites we surf. Data is useful – it informs human decisions about medicine, government policy and entertainment. But data can also be misused, and the way data is collected, analysed and presented can change your final conclusions about it.
All data tells you a story. But the important question is, should you believe it? Here are 4 tips that you should apply when trying to suss out whether what you’re seeing is an accurate interpretation of the numbers, or an incomplete or misleading one.
Check the Source!
You’ve heard of the term “fake news”. Here’s a simple step you should take to prevent yourself from worsening the problem: before you share an article, check. the. source.
- Does the article come from an objective, reliable, trustworthy person or website?
- Is this article funded by a certain person or organisation it is claiming to support?
- Might the article have a hidden agenda?
- Has the author shown a tendency to take one particular side, cherry-pick information, or ignore counter-considerations?
If you’re reading a research paper:
- Has it been peer reviewed?
- Does it mention any other clear quality control process?
- Has this paper been retracted or flagged as a predatory journal?
Find the Factors
You’ve heard of correlation or causation claims – they sound something like “Water is bad for you because 100% of people who drink water end up dead!” Ok, that’s an extreme example, but when presented with two things that are allegedly linked, ask yourself two questions:
- Are there other factors at play that might have been discounted or ignored? (Like disease or lifestyle habits for example…)
- What is the sample size? If the claim is based on a very small amount of data, it might not be wise to put money on it or let it be a swaying influence on your decision making.
What Are Data Visuals Telling You?
Different types of visuals can change the angle of a story because they highlight different aspects of information. Here are some of the most common types of visuals:
- Line chart: trend of one variable over time
- Pie chart: proportions, ie. parts compared to the whole
- Bar chart & histogram: distribution of a few variables
- Scatter plot: the relationship between two variables
- Maps: information across a space
Check out this video to see some examples at play.
Things to look out for:
- Does the visual used support the story the author is trying to tell?
- Is the scale of the frequency axis consistent, or has a portion of it been cut out and if so why?
Data in Social Media – Understand How The “Algorithm” Works
Ever experienced this: you searched for “the best chocolate cake” last week and since then, it’s everywhere you look! “Sinful Chocolate Indulgence”, “7 Layer Chocolate Cake Heaven”, you get the idea.
This might not come as a surprise to you, but when you’re on the internet, you’re exchanging your data for entertainment. The aim of any website “algorithm” (sound familiar?) is to keep you on their site for as long as possible. It takes what you seem to like (based on what you click or how long you watched a video, what IG post you liked, what news article link you shared) and recommends more similar content.
So rest assured, the algorithm is not evil nor can it read your mind. It just gets good at reading your behaviour. Check out this video for a visual illustration of how algorithms work.
We hope this has given you the basics of how to be a more careful interpreter of data. For a deeper dive into data literacy, including how it’s used in news, sports and politics, we highly recommend this Study Hall: Data Literacy YouTube playlist.
Source:
Study Hall: Data Literacy (Crash Course x Arizona State University)
