Shine's Spotlight: ‘Spoon-Fed:Why Almost Everything We've Been Told About Food is Wrong’ by Tim Spector
Tim Spector is both a professor of genetics at King’s College London and an author of popular science books, such as this one we’re reviewing and “The Diet Myth” amongst others. His research specialises in a range of topics including being the lead researcher on the COVID symptom study app created by a company he co-founded called Zoe. They are a nutritional science company who investigate how our biological responses to food are based on things such as bacteria in your gut (an expanding area of research, known as the microbiome).
This book was bought for me as a present by a friend studying Public Health. It has become apparent over the last few years that what we eat is a largely understudied and badly communicated area of all our health. It is claimed that part of what makes this so difficult is that food companies lobby governments and fund research, which will tend to result in subsidies and scientific research which support the sort of food that these companies produce, which may not actually be best for our health, e.g. soft drinks, breakfast cereals etc.
Therefore, there is a lack of genuinely independent research into what foods are best for us, and even when there is good research, it is often badly communicated to us. Tim Spector aims to help correct this with this book exploring a range of food myths many of us believe, supported by research and data.
The book is split into 23 chapters, each addressing a food myth, i.e. things many of us believe about food that aren’t actually true. These range from “Breakfast is the most important meal of the day” to “Calories accurately measure how fattening food is ” to “Drinking alcohol is always bad for you”. This means it is definitely a book you can dip in and out of, rather than needing to read it as one extended narrative. I found Spector’s writing clear, engaging and thought provoking, around a topic that we are bombarded with a constant (and often incorrect or misleading) dialogue.
Data analysis is subject agnostic, and hence I will share some general points about data analysis and data communication which I learnt from this book.
Are your controlled variables actually controlled?
When comparing the effect of the independent variable on a dependent variable in an experiment, it’s essential that the control variable is actually controlled. This might seem obvious, but perhaps it can be easy to wrongly assume that they are.
This is covered in the food myth “Nutritional guidelines and diet plans apply to everyone”. We all respond very differently to the same types of food. Therefore, what may be good nutritional advice for one individual may be bad for another. An example of this is that normal people can vary up to 10x in their blood sugar’s response to the same food. Therefore in nutrition, they have to be very careful when comparing the effects of varying diets between individuals.
It may be the same in any data analysis you are doing, if you are comparing the effect of a change in an independent variable on a dependent variable on two separate samples (or events, scenarios etc), are the factors you believe are stable between the two samples actually controlled?
2. Sometimes the simple answer isn’t correct
Occasionally some facts might appear obvious, but if you do a dive into the data you actually get what feels like a counter-intuitive answer. This is covered in the food myth “Local food is always best”. One example given as to why this is a food myth, is the idea that if you live in the UK, then Welsh lamb would surely be far better for the environment than frozen lamb from New Zealand which has to be transported halfway around the world? Well it turns out because the actual conditions for rearing sheep are so much more beneficial in New Zealand (e.g. rarely having to supplement pasture-feeding with animal feed), and shipping is fairly low-carbon per product at large scales, that the carbon footprint of lamb from New Zealand may actually be better than local lamb.
Sometimes there may be assumptions in your business area, that when you consider the data behind it, that may be wrong.
3. Remember that things aren’t always linear
When exploring a relationship between two variables, it can be possible to make a mistake in assuming they are linear when they are actually non-linear, but you haven't investigated both ends of the spectrum. An example of this is covered in the food myth “We all need to reduce our salt intake”. This section illustrates that very high salt intake is bad for your health, but it turns out that very low salt intake is bad for your health too. So just saying “reduce your salt intake” is bad health advice (also turns out, it’s only actually very high salt intake that is bad and most of us are probably fine with how much salt we have).
So sometimes we make the mistake of assuming that because something is very bad in high levels, that the more it is reduced the better, but it may be that there is a perfect middle ground.
4. It may be worth checking the funding behind studies…
When reading research studies or published data, it is always worth looking into the source behind this data, as the author or the funding behind the author may give some bias. An example of this is covered in the food myth “ Sugar-free foods and drinks are a safe way to lose weight”. Between 2010 and 2017 the US HQ of Coca-Cola spent $140 million on research grants to academic scientists, many of which went on to publish work saying that sugar or artificially sweetened products are safe to consume. This isn’t to say that this work is necessarily wrong, but it’s worth bearing in mind where possible the potential bias behind research or data analysis.
5. Is showing people data actually improving anything?
Sometimes those of us who work in data can assume that if we just show stakeholders/decision makers/customers more data, they will go on to make better decisions. This is especially true in the current climate where live data dashboards are particularly in vogue. An example of extra data not really helping anyone is demonstrated is the food myth “Food labeling helps us make healthier choices”. Apparently only a quarter of the British public pay attention to food labeling, and furthermore there have been some studies that show people can end up eating more when there are calorie labels, because for example, if they have a low calorie drink, they may feel they can justify eating more food in a restaurant.
So, before spending hours finding ways to show more data to your customer or client, consider if this will actually have the positive effect you are hoping for.
Please get in touch with any good data-related books you think I should read or check out what I’ve been reading on my GoodReads.