How much do we talk about climate change?
For this story, I am re-publishing an old post I had on my blog, which originally came out in 2020, with some adaptations. I didn’t manage to update the data and I’ll tell you why.
I was wondering how much the media historically reported on topics of climate change as it is definitely true that in recent years this topic has acquired a much higher media presence than it ever had, but what are the stats?
Climate change as a phenomenon is a thing solidly known in science at least since the early 1970s, and I could rant about how long it took the world to actually take notice and start acting and how behind where we should be we are, but I will not do so. Here’s a good timeline of the discoveries.
The #FridaysForFuture movement, started in August 2018 by Greta Thunberg, has expanded into a global youth movement; public awareness of the climate catastrophe and sensibility towards the environment have increased; vegetarianism/veganism are now social phenomena not relegated to cultural niches anymore; people are embracing recycling and up-cycling more willingly than ever; brands are arising that propose sustainable shopping; there’s more appreciation for fair and low-impact production methods. How much is the media discussing these topics? Yes, of course it’s a chicken-and-egg situation: the more the media discusses them, the more people will embrace them, and vice-versa.
The data card
The viz displays the count of stories (articles) published in the New York Times containing the keyphrases “climate change“ or “global warming” and, for comparison, the total of stories. There is a clear and unsurprising upward trend in more recent years, especially for “climate change”.
From all stories and climate change ones, it appears there is a peak between the years 2006 and 2009. Specifically, while all stories averaged at about 102k before 2006, the average jumps to about 160k for the period 2006-2009, to then decline again to about 88k afterwards. The years of the 2008-2009 financial crisis might have at least in part contributed to this peak, however there might be a ton of other reasons why the NYT published more in that period and it would be very interesting to see what other papers did at the same time.
Regarding climate change, a peak around the same years is visible, obviously following the overall counts, and then a subsequent upward trend, which is quite pronounced for “climate change” (in 2019 alone about 4300 stories have been published that contained the phrase “climate change”). I certainly do have the impression that “climate change” is nowadays the most common catchall phrase, while “global warning” seems to have faded out of fashion. Obviously the two terms are different things, but in ordinary speak they might get used interchangeably, which is another reason why I chose to track both.
2009, the peak year in the plot, was the year of the Copenhagen UN Climate Conference, known for not having reached an agreement on the ambitious plans it was aiming for. The NYT specifically titled on the 14th of November 2009 “Leaders will delay deal on climate change”, a big shame.
It is also interesting to notice that “climate change” seems to have been more common in stories than “global warming” even in the years up to 2000 - I would have thought that “global warming” was the preferred common phrase some decades ago, but apparently not.
Note that a section called “Climate” started appearing on the paper in 2017.
The data - finding a proxy
Choosing a source
In need of a credible proxy press source which also had to have a good interface to retrieve data, I’ve chosen the New York Times. It is an American paper, but it has an international presence and readership (I’ve been a subscriber for a few years, for instance). It is also a reputable outlet with large circulation.
Sure, there’s several biases in using just one source to proxy a general news presence of a topic, but with the due caveats this is a decent source. To do better, I would have had to go and look at several papers, possibly choosing amongst those with the highest circulation country-wise, which would have been a bit much for the sake of this little illustrative work. I have tried looking at The Guardian as well (British paper, authoritative) but its search feature didn’t allow for retrieving the needed data.
Note that the fact that I’ve chosen a reputable paper to do this also means that I’m getting data about how much the topic has been discussed by non-deniers (apart from the stories about that phenomenon). I’m not interested in the opinions of papers presenting fakes, pseudo-science and generally low-quality stuff. Maybe one day I can use those to quantify the effect of falsities, but that’s a story for another day.
The NYT has a search feature that allows you to retrieve articles dating back since its early days in 1851 (another reason why is suited my requirements): the search is purely phrase-based, namely you type a query and get back all the articles containing it, with a counter. The search feature is not case-sensitive and it yields documents which contain the phrase in its entirety. You can filter results down by date range and by section.
Note: in republishing this piece here I tried adding data for the most recent years (from 2020 onwards), but apparently the search function is not working well when I filter for a date range, which is why I wasn’t able to do so.
Choosing what to look for
To measure the NYT’s coverage of climate change, I decided to search for “climate change” and “global warming” year by year starting from 1980 to present and take note of the count of articles retrieved. I’ve figured these two would be the most relevant phrases present in articles related to the topic so they would make for a sensible assessment of its prevalence; no doubt there’d be plenty of other suitable ones (“environment”, “fridays for future”, “pollution”, “ocean acidification”, “carbon emissions”, “carbon footprint”, …) but they’d either be too vague/too specific or risk denoting something else unrelated to the area of interest (for instance, an article containing the word “environment” might not talk about climate change at all). Anyway, I’m surely underestimating the count of relevant articles (by excluding those on climate change which don’t contain any of these specific phrases, there might be and there might be many), but I think I can safely assume that the vast majority of articles reporting on climate change will contain at least one of these two phrases.
Note that articles containing both phrases will figure in both counts because with the NYT search feature it is not possible to separate them: it is not possible in general to run a semantic query, as it is for instance on Google Trends where you can query for a topic, which encompasses all phrases in its semantic sphere. Note that I’ve thought of using Google Trends to have at least some indicative information, but of course it doesn’t go far back in time enough. This is why counts have been kept separated in the plot.
I’ve also taken the counts of all stories written in total on the NYT (in red in the graph) and those in some specific sections of the paper (“Arts”, “Science”, not shown), to contextualise numbers. Because of course the structuring into sections might have most likely evolved in time (and it has), using sections only has an anecdotal value. Furthermore, I’ve noticed that the counts of stories in all sections do not sum up to the total counts in a year - this is obviously because not all stories are attached to a section (I’d imagine general ones, breaking news, reviews and summaries may not).
The data is taken up until 2019 (again, I didn’t manage to update the data).