Tracking Global Energy Narratives

Team Members

Max Kortlander, Roberto Pizzato, Miles Zornig.

Introduction

For the past two decades, there has been an ongoing debate about the reality of climate change. While some actors remain skeptical of the notion, many scientists have observed severe impacts of climate change on physical and ecological systems that observed throughout the last century (Adger et al. 77). As a response, individuals, organizations and whole societies have adjusted (or tried) their behaviour towards climate change, by notably changing their energy policies (Ibid.).

NGO’s such as the Global Call for Climate Action (GCCA) have been trying to spread awareness on this issue, by also improving the public discourse on climate change.

Being a guest at the 2015 Digital Methods Winter School, GCCA provided essential concepts of their narrative in order for the researchers to track them down within the websphere. Five concepts that the GCCA initiated played a crucial role within the scope of this research: “coal”, “fossil fuels”, “dirty energy”, “100% renewable energy” and “phase-out”. The task the researchers received was to track down these concepts within the websphere of eight countries, from which the four English-speaking countries were chosen and focused on: Australia, India, the United States (US) and Canada.

By tracking down the concepts given by the GCCA, it was especially relevant to investigate where the concepts originated from and, most importantly, where the four concepts were used in relation to other concepts that the GCCA has been trying to push: “phase-out” and “phase-in”. If someone in any of those countries was to consult the Internet to find information about the controversial topic of climate change, he or she is likely to use a search engine. Thus, this research will focus on the first search engine result pages (SERPs) of the four countries in question. By doing so, it will be possible to determine whether or not the narrative of the GCCA has moved from inside the climate-’bubble’ to the outside, the mainstream. The local first SERPs of each country will be here be considered to be the ‘mainstream’, as Google, being a algorithmic search engine, organises information on the web and has become the main source for finding information online. As such, Google does not only act as a facilitator for accessing knowledge, but also as a gatekeeper of information (Hinman 67; Rogers 95).

After querying the four concepts on each local Google domain, the first step will be to organised the actors present on each and weigh their importance within the discussion regarding “coal”, “fossil fuels”, “dirty energy” and “100% renewable energy”.

After, the concepts of “phase-out” will be traced on each local SERP in order to establish which actors make use of this concept with regards to each of the queries.

Last but not least, this work will also take an analysis of Wikipedia, utilising front-end- and back-end- analysis.

Research Questions

Are key concepts from GCCA’s narrative gaining traction, and if so, are these concepts discussed within the environment-bubble only, or also within the mainstream? Which public is addressed?

Methodology

Conceptualization

Before proceeding with our analysis, it was first necessary to define an inside and outside to the “bubble” addressed in the introduction. The GCCA and climate professionals can be said to be inside of the bubble, while non-climate professionals comprise the outside of the bubble. With this distinction, we then defined three spheres through which we expected to see GCCA-narratives travel. The first sphere is the GCCA, where climate-related concepts originate. The second sphere comprises climate professionals, while the third sphere consists of non-climate professionals.

Sample selection

A sample of countries (India, Brazil, Australia, Germany, US, Canada, France and Indonesia) was provided by GCCA in order to evaluate the effectiveness of their work in those countries. Due to time and linguistic limitations, the researchers decided to only focus on the four English-speaking countries, namely the United States, Canada, Australia and India.

Query design and choice of keywords

The next crucial step was the ‘query design’, defined by Rogers as “[…] the practice of formulating a query so that the result can be interpreted as indications and findings […]” (Rogers 110). From the list of queries provided to us by the GCCA, we were able to come up with a list of the four most relevant queries of this study: [coal], [“fossil fuels”], [“dirty energy”] and [“100% renewable energy”]. Quotation marks were added to the queries consisting of more than one word, since without quotation marks, as Rogers argued, these “[…] queried keywords may return results for synonyms or other inexactitudes, providing equivalents as opposed to matches […]” (Ibid 111). In this case, GCCA provided the researchers with a list of words used by the NGO. The GCCA asked to track down the queries in combination with the concept of “phase-out”. Given that these keywords could be written in two different ways, they were typed in as “phase-out” and “phase out”, and “phase-in” and “phase in” to be scraped by the Lippmannian Device.

Last but not least, the query [“100% renewable energy”] suggest the idea of ‘phasing-into” more sustainable energy sources, therefore, this query was also search in combination with the keywords “phase-in” and “phase in”.

To sum up, the queries and the keywords chosen provided the substance of GCCA narrative that has been used sensitize the phase-out from fossil fuels, in other words to make the people aware of the topic.

Browser configuration

To ensure that personal settings and the browser history did not interfere with this research, a depersonalized browser was installed to conduct this research. To ensure the depersonalization of the browser, the following steps have to be considered: Download and open Firefox, open settings, click the option ‘Tell site that I do not want to be tracked’, choose the option ‘Never remember the history’, and ensure that the location bar suggests ‘nothing’. This was important in order to avoid any bias that could potentially interfere with the results throughout the process of this research.

Search engine selection

Google was used as the main search engine to conduct this research, as the Webcertain Global Search and Social Report 2013 indicates that the United States, India, Australia and Canada have a Google market-share of 67%, 97%, 93% and 67%, respectively (Return On Now). Google is therefore the most used search engine in these countries and will be serve as the basis of this research.

Search engine customization

It was necessary to take some precautionary steps before diving into the research with Google’s search engine. First, it was important to customize Google’s search settings, by disabling Google’s instant predictions, in order to enable the customization of ‘results per page’, which was set to 20 results per first SERP in order to have some margin. Then, the following address had to be inserted into the browser’s address bar: www.google.com/history/optout. This gives the user a possibility of disabling Google’s customizations based on search activity to prevent the search engine’s algorithms from interfering with the research.

Data collection

Google News and Google features were included in the data collection in order to simulate a Google Search hic et nunc, by users from each country. The only results that were omitted were Google Ads and Sponsored links.

Analysis of local Google first SERPs

Throughout the forthcoming analysis, the following definitions will be used:

Ranking

The order in which the search results appeared on their respective SERPs.

Actor

The source of the page, in other words information on the person, group or organization behind the respective website.

URL

A direct link to the website.

Weight

In order to establish the relevance of each actor, a specific weight was applied to the position occupied by each actor in terms of ranking. So that, the higher is the ranking of an actor in a local SERP and higher is the weight of that actor.

Triangulation tool

The Triangulation Tool was used in order to establish which actors were the most displayed in the selected sample.

Data visualization

Once the data was collected in a spreadsheet, the next step was to clean it for further analysis and evaluation. Microsoft Excel’s automated counting function was then used to calculate the occurrences of the respective positions on nuclear energy per country, as well as the type of actors represented and their rankings. The results were then plotted on clustered column charts for further interpretation. Also, the triangulation tool provided by the Digital Methods Initiative (DMI) was used to analyze the commonality of search results between different local SERPs and to visualize them as well.

Bubble lines

The visualization of the “weighted” actors was enabled by another tool of the Digital Methods website, Bubble line. The weight assigned to the three most relevant actors per country allowed a visualization of the actors’ relevance, in other words the dimension of each bubble was proportional to the weight of its ranking.

Harvesting URL’s of each local Google first SERP

[coal], [“fossil fuels”], [“dirty energy”] and [“100% renewable energy”] were individually queried on the local Google domains of India, Australia, the US and Canada. In order to extract all URL’s of the first search result page, the selection source of the first SERP was copied into the Harvester-tool.

Scraping of harvested URL’s

Afterwards, it was relevant to determine on which of the harvested links contained a combination of our query and the keywords “phase-out” and “phase out”. For this, the list of URLs retrieved by the Harvester for each individual query was copy-pasted into the top box of the Lippmannian device. Then, the respective keywords had to be entered into the box below. The retrieved URLs for each query inserted into the local Google domain’s search bar were then ‘scraped’ for the keywords “phase-out”, “phase out” and the query itself, as the device was supposed to only give us an indication of the links that contained both our query and the keyword. For example, the links harvested from Canada’s local Google first SERP on [coal] were inserted into the Lippmannian device and scraped for the keywords “phase-out”, “phase out” and “coal”. After all links were scraped for these keywords, the CSV file was downloaded from the ‘Output’-menu of the tool, imported and opened in an Excel sheet. This process was repeated 12 times, for all four queries within all four countries’ local Google first SERP. For the query [“100% renewable energy”], the process was repeated by using the keywords “phase-in”, “phase in” and “100% renewable energy” to see which links could contain information on countries ‘phasing-into’ more sustainable forms of energy production.

Findings

[“Fossil fuels”]

The query “fossil fuel” provided results four SERPs local pages completely different from the others: in each local Google Search page one ‘Google News’ and three ‘In-depth articles’ are to be found.

For the query “Fossil fuels” four URLs were displayed among all the four countries selected: the Wikipedia page titled “Fossil fuel”, the article titled “Energy Story - Where Fossil Fuels Come From”, the article “Harvard defies divestment campaigners and invests tens of millions of dollars in fossil fuels” which appeared on the Guardian online on Wednesday 14 January 2015 and an interactive quiz about energy by Eco Kids, an initiative of Earth Day Canada.

While the above-quoted Wikipedia page is ranked as the first one among all the four countries analysed, the article by the English online newspaper is the only Google News results in common among the selected sample. Energy Quest, the education website of the California Energy Commission, emerged as a relevant actor among the whole sample.

It is interesting to notice that in none of the above-quoted webpages was possible to find the keywords. Even more remarkable is that within the Wikipedia page titled to “Fossil fuels” there is no internal link for the “Fossil-fuel phase-out” Wikipedia page. That is that, the best ranked and reasonably one of the major source of information about fossil fuels does not display the word phase-out at all, while in the Wikipedia page “Nuclear power” there is a section about “Nuclear phase out”.

For the query “fossil fuels”, 13 out of 42 results are Google News results (31%), of which 9 are ‘In-depth articles’ and 4 news published the day results were retriggered. In other words, Google News represent about one third of the total amount of webpages collected and emerge as a fundamental feature, not only when a user browse their websites, but also when she looks for information through a search engine.

In India the first result was an advertised link by the Chantam House - Royal Institute of International Affairs, a think tank sponsored by important oil company such as Shell and Exxon aming others. This was the only case of a Google Ad as first result in the analysed sample.

[“coal”]

As it was for the query “Fossil fuels”, the percentage of results which belong to the feature Google News is about one third, 13 out of 42 (31%).

When typing “coal” as query in the Google Search page, the only result displayed in all the four different countries is the Wikipedia page named “Coal”. Nonetheless, the Google feature displaying a short abstract of the query is different in each of the four countries: in the US the text is taken from the US Environmental Protection Agency, in Canada it is an abstract from the Wikipedia page titled “Coal”, while in Australia the short text is taken from the website of the integrated energy company “Origin Energy” and in India from the “World Coal Association”. Two are the actors which are displayed three times each: the fashion brand Coal Header – not relevant for this research – and The Economic Times.

[“dirty energy”]

Even though “dirty energy” is a not neutral query, this particular narrative of GCCA appears to have weak traction. Five out 10 results (50%) are common to all the countries among the local SERPs of our sample. Two of them are advocacy group related to the topic of dirty energy: the NGO International River and the website Dirty Energy Money, an initiative of the advocacy group Oil Change, The others are are related to a movie called “Dirty Energy” and “The dirty energy dilemma”. Advocacy groups represents the 33% (13 out of 40) of the total results.

News are represented only by the Press Agency ‘Inter Press Service’, displayed in three out of four countries. That is, that the narrative of GCCA did not get to the mainstream through News & Media websites – which probably do not use this definition because of its strong connotation.

[“100% renewable energy”]

Two are the actors in common between all the four countries: the Wikipedia page “100% renewable energy” and “Go 100% renewable energy”. On one hand it can be said that this query went mainstream because it can be found on Wikipedia, on the other it can be argued that is likely to be found only searching for that. In total 31 out 41 of the actors (76%) displayed are advocacy groups fighting for the switch to the use of 100% renewables sources of energy. The scenario appears completely dominated by NGOs and advocacy groups, while News & Media sources are only four out of 41 (10%).

On one hand, the query went mainstream because it can be found on Wikipedia, on the other it can be argued that is likely to be found only searching for that. In total 31 out 41 of the actors (76%) displayed are advocacy groups fighting for the switch to the use of 100% renewables sources of energy. The scenario appears completely dominated by NGOs and advocacy groups, while News & Media sources are only four out of 41 (10%).

It can be argued that the communication of GCCA has only marginally gone out of the bubble.

News Websites

Considering all the queries used in the four countries, at the first glance it can be noticed that the highest percentage of News & Media websites is in the US 12 out 44 (27%), while in Australia is 9 out of 43 (21%), in Canada 8 out 43 (19%) and in India 7 out 39 (18%).

The query that showed the smallest number of News & Media websites is “dirty energy”, only three, followed by “100% renewable energy”, with four news URLs. Thus, it can be said that these two queries have not been used consistently out of the so called bubble.

However, the large amount of Google News links has not to be confused with the number of News & Media websites displayed, but it represents only the number of articles shown by Google News.

Environmental advocacy groups

The advocacy groups and NGOs displayed in the sample are 37 out of 169 (22%), but only four of them appeared in the SERPs retriggered using the queries “fossil fuels”, none in relation to “coal”. That is when using neutral queries, the traction of the GCCA narrative seem to be not really effective out of the so-called bubble.

Findings of scraped URL-lists

The Lippmanian device showed that, in all four countries, the Wikipedia Article on “Coal” contains a link on “Phase out coal”. Australia has an additional Wikipedia article, “Coal in Australia”, also linking its users to “Phase out coal”. The Scraper also revealed five articles by the Mining Australian Magazine, where “coal” and “phase-out” are used together in the same article. Beacon News and the Natural Resources Department, actors found on Canada’s local Google domain, also use both the query “coal” and the keyword “phase-out” together in one article each. In India, there are – in addition to its Wikipedia page on “Coal” – two websites containing articles where “phase-out” and “coal” are used together – indiabeyondcoal.org and the website of the Indian Ministry of Coal.

When scraping [fossil fuels] for the keywords “phase out”, “phase-out” and “fossil fuels”, the Lippmannian Device showed that all countries have a Wikipedia article on “Fossil Fuels”. However, no direct link can be found on the page of the article on the issue of “phase-out”, but it is the link to “Coal” on the page that links the user to a page on “Coal” containing the link “Phase out coal”. On the local Google domain of the US, Canada and India, the scraper revealed that among the harvested URL’s, the Climate Movement 350.org and the magazine Scientific American had one and three publications respectively, where the query “fossil fuels” and “phase-out” were used together in the same article. Also, considering this exact query, India and the US had the most publications talking about “fossil fuels” with regards to “phase-out”: 59 BBC publications scraped from India’s first local SERP and 39 publications by the Energy Information Administration from the US’s first local SERP.

The query [dirty energy] in combination with the keywords “phase out” or “phase-out” on the first SERP of the local Google domain of the US and India was present on two occasions each: Each country had two publications by the Inter Press Service website which discussed the phase-out of energy sources, not with regards to dirty energy, but with regards to fossil fuels. However, in both cases, a link on the page of the article led users to an article called “Dirty Energy, Dirty Tactics”. Australia had the most publications discussing “Dirty Energy” with regards to “phase-out”, seven to be exact, all originating from the Environmental Health Coalition.

What should also be mentioned here, is that the scraper collected 99 Facebook links from the links harvested from the first SERP on Australia’s, Canada’s and the US’s local Google domain. The scraper revealed the description of these Facebook links, showing that the pages discussed issues of “dirty energy” together with “phase-out”. For example the Facebook page of the Labor Network for Sustainability scraped from the US’s first local SERP, which states in its description: “Washington State, a blue-green alliance is helping to phase out dirty energy while creating good jobs.“ The exact same 99 Facebook-links and descriptions can also be found in the scraped data of Australia’s and Canada’s first local SERP.

The last query of this research was [“100% renewable energy”], which was also scraped for the keywords “phase out”, “phase-out” and “100% renewable energy”. The Lippmannian device revealed that the “100% renewable energy” and “phase-out” were only used in combination in one publication, an article by Ecowatch, an ecological news-website which appeared in India’s local Google first SERP. However, while the concept of “100% renewable energy” could be found in the title of the article, the concept of “phase-out” was not present within the article itself, but in the comment section, where a user discussed the importance of phasing-out of certain energy sources. Last but not least, the query [“100 % renewable energy”] was also scraped for the keywords “phase in”, “phase-in” and “100% renewable energy”, as it is assumed that when discussing renewable energy, people would also use the concept of “phasing-into” such energy sources. However, the scraper revealed that in the four countries considered in this study, no publication made use of the two concepts together.

Google Trends & Wikipedia

Google Trends

Google trends is a tool that allows users to see various metrics regarding search traffic for one or more given search terms. In order to give a historical overview of various terms’ online popularity, we utilized Google Trends graphs.

We began by entering the search terms that we have been using throughout this course: coal phase out, fossil fuel phase out, dirty energy phase out, and 100% renewable energy. Unfortunately, not all of our queries were popular enough to include the same information. For example, not enough data was available regarding “100% renewable energy” to have mapped geographic data on the subject. After exhausting all of the terms used in our other google search (coal phase out, natural gas phase out, and dirty energy phase out) without available data, the term “phase out gas” was the most similarly related term that did indeed return results. The term “100% renewable energy” was able to be analyzed by Trends, although not exhaustively so.

The terms ‘phase out gas’ and ‘100% renewable energy’ were then entered into the compare box together, allowing for a comparative view of the terms’ “interest over time”. With these two terms overlaid onto one graph, it is also possible to zoom in on particular moments in time to get a clearer view of which particular times saw the most traffic. This micro view also allows one to view the news stories which were implicated in these spikes.

Wikipedia

Front-end and Contropedia

The English Wikipedia site was searched for entries of 4 terms of interest: (1) "Phase Out" (In this case for specificity to environmental discourse, the entry is entitled 'Fossil-Fuel Phase-Out), (2) "Coal", (3) "Fossil Fuel", and (4) "100% Renewable Energy". There was no Wikipedia entry for the term "dirty energy".

Each entry was first analyzed on the front end using a ctrl+F search for each other term of interest. For example, the entry for "Fossil Fuels" was scanned for occurrences of the terms "phase out", "100% renewable energy" and "Coal". This process was repeated for each term on each entry. It was noted whether a term occurred as a link or plain text.

Additionally, there was a backend analysis of the "Fossil-Fuel Phase Out" entry using the Contropedia Tool. The tool showed that, of our terms of interest, "fossil fuels" and "coal" each had a history of editorial controversy.

The connectivity amongst these pages was finally mapped visually using Adobe InDesign.

Back-end Analysis

The “View Page Statistics” section was examined for both of our entries. Using these graphs, we noted spikes in traffic which could then be coupled with a Google search (described below) to give further insight into what caused these traffic spikes.

Google Search

In this case, it was noted from a back-end Wikipedia analysis that the entry for “fossil-fuel phase out” saw particularly high traffic on 13 November, 2014. Google was then searched for “fossil-fuel phase out” with results limited to those published on 13 November, 2014. The top ten results in this search all mentioned the G20 in their titles, leading us to believe that news coverage of the G20 is related to increased interest of the term “fossil-fuel phase out” on that day.

Talk Page

A Wikipedia article’s talk page provides insight into controversy and edits of a page. In this study, we focused on a qualitative study of the talk page, which involved reading through the latest talk pages to identify recent controversies regarding the pages’ editing.

Wikipedia analysis findings

These results show that on the day of the Wikipedia traffic spike for “fossil-fuel phase-out” are likely a result of news regarding G20 decisions.

Two relevant findings resulted from the use of Google Trends. First, as seen below, Trends showed that both “100% renewable energy” and “phase out gas” rose to prominence drastically, and eventually maintained a steady search volume.

Terms stick, narratives change

Secondly, Trends notes top news stories based upon selected query terms. As seen below, the term “phase out” was never used in reference to a complete non-renewable phase out, but rather the phasing out of particular technologies and support surrounding non-renewables. Examples shown below include the phasing out of “fossil fuel subsidies”, “fuel tax”, and “energy inefficient light bulbs”.

These same phase outs, regarding subsidies and the like, were quoted on Wikipedia. But why? The Talk Page for ‘Fossil-fuel phase out’ provides some insight. Many users noted that full phase out does not exist in practice. Thus, when “phase out” is used concretely, it tends to refer to these partial measures. To avoid this, it would be recommended to take one of two steps: 1) Tie the term ‘phase out’ more clearly to an environmental movement (rather than environmental phenomenon [which proponents argue it is not]) that has the ultimate goal of 100% renewable energy or 2) foster the adoption of measures in practicing law and diplomacy that do, in fact, utilize ‘phasing out’ as a protocol towards 100% renewable energy.

Conclusions

For what concerns the Google Search results, it can be argued that the relation between the neutral queries and the keyword phase out was weak, while for the other two queries, [“100% renewable energy”] in particular, the narrative does not gain a significant traction in the mainstream and resonate mainly within the bubble.

Given that “fossil fuels” is the query in which most Google News links appears, it is reasonable to think that it is on this terrain that GCCA should get more traction.

Moreover, when talking about nuclear energy, in fact, on the dedicated English Wikipedia page there is an internal link to “Nuclear phase out”.

Users browse the Internet and use Search engines in order to get information, the role of educational website, such as ‘Energy Quest’ emerges as fundamental, in particular for the younger generations. This could be a field that can be penetrated by GCCA communication and be an effective vehicle of its narrative. Another significant finding was that, on the American, Canadian and Australian local Google domains, the Lippmannian device returned a significant amount of results originating from Facebook when [dirty energy] was queried, 99 in total. ON these pages, “dirty energy” was discussed with regards to “phase-out”, which suggests that a strong narrative is going on within the sphere of social media. This should be investigated further.

The query [“fossil fuels”] did not provide any results related to the keyword “phase-out”, while a different query such as “fossil fuels phase-out” provided a list of results. Thus, it can be argued that a user looking for information on fossil fuel is not likely to find results connected to a real debate on phasing-out, that is to say that the narrative of a phase-out policy appears only to a public that is searching information about that. To be considered effective, the communication strategy should enable a user to find “keywords” within the content accessed through neutral query.

To sum up, “fossil fuels phase-out” and “coal phase-out” appear as a topic that is likely to be found by people which are already interested in that particular issue, people which reasonably represent an “environmentally aware public”.

Regarding the non-neutral queries, [“dirty energy] provide results not strictly related to environmental issues – i.e. Dirty Energy bars. The negative connotation of the term would probably make it difficult for the narrative to get to the mainstream, nonetheless this terrain could be the most appealing for an effective narrative.

Finally, the association of last query [“100% renewable energy”] and the keyword “phase-in” is remarkably inconsistent, that is to say that not even within the so-called bubble this narrative works.

Discussion

These multiple research approaches aimed to capture a general view of a narrative which changed and gained traction over time. Yet narratives are inherently qualitative, elusive, and contextual. We aimed to address this challenge by emphasizing and considering the role of identifying keywords, which may more easily be tracked and quantified.

While there is much more to narrative than keywords alone, an analysis of specific terms also shed light on the importance of the terms’ very structure. For instance, “phase- out” occurred in all sorts of settings - referring to electronics, audio engineering, and sometimes non-renewables - while “100% renewable energy” was entirely found in this energy-specific context. While this may seem obvious, it suggests that the more specific a term is, the less this term may be changed by the context around it. “Phase-out” has gained traction as a term but changed in meaning, while “100% renewable energy” seems to have traction as both a term and as an intended narrative.

Finally, the Wikipedia analysis shed light on the connectivity of our terms of interest with one another. Most notably, the entry for “coal” does indeed link to “fossil-fuel phase out”, meaning that the term “phase-out” has indeed picked up traction within the mainstream to a certain extent. However, “fossil fuels” and “100% renewable energy” do not currently link to phase out. Further work could aim to tie these terms more closely and prominently together.

Bibliography

• Adger, W. Neil, Arnell, Nigel W., Tompkins, Emma L.. “Successful adaptation to climate change across scales”. In: Global Environmental Change 15. 2005.

• Hinman, Lawrence M. Searching ethics: The role of search engines in the construction and distribution of knowledge. Web search: Multidisciplinary perspectives. Eds. Spink, Amanda, and Michael Zimmer. Springer. 2008.

• Return On Now. “2013 Search Engine Market Share By Country”. 24 October 2013. Web. 10 December 2014.

• Rogers, Richard. Digital Methods. MIT Press. 2013.