Abstract

This study focuses on the ways that UK environmental publishers utilise Facebook to engage their users with sustainability. The research explores users' engagement with posts shared by Greenpeace UK, WWF UK and Guardian Environment between 1st November 2015 and 31st October 2016. By mapping the number of shares, comments and reactions posts receive to features in the posts, we investigate how different features can help maximise user engagement. Analyses include the type of media accompanying posts; textual features such as emojis and hashtags; thematic analysis of post content including mapping categories to the pillars of sustainability (Gibson, 2006); and identifying the purpose of posts in line with Saxton and Waters' (2014) I–A–C Framework. We recommend that publishers always use accompanying text with posts to provide context and inform users. By asking questions in post text, publishers can encourage comments and build community, whilst embedding links can inspire action. Visual media, particularly photos, are effective, and can be used to entertain users. Publishers should also include greater diversity of content themes to reconcile the three pillars of sustainability. Although each publishers' audience will vary, these common features could encourage their users to engage with sustainability.

Sustainable Development Goals: Partnerships for the Goals (17)

Introduction

Media and the Environment

The public's understanding of the natural world and humans' role in the world are shaped by communication (Carvalho, 2009). The relationship between audiences and the media has traditionally been criticised for being biased in the direction of media channels (Happer and Philo, 2016). Power dynamics can lead to some organisations being able to speak out about issues, whilst others are limited (Carvalho, 2009), leading to the media influencing the prominence and advancement of certain environmental issues over others (Hansen, 1991). The dynamic between media and its audiences is evolving to embrace the potential for real–life engagement with people's lives whilst offering greater interactivity (Happer and Philo, 2016; Walker and Starosielski, 2016). As social media platforms can be accessed by any organisation, sites like Facebook could counter these power dynamics and increase the transparency of the media.

Although the media does not dictate people's behaviour or how they should think, it can frame public knowledge through the information it shares and the methods used (Happer and Philo, 2016). The media has evolved to become interactive, with people being invited to participate in discussions about the things that happen locally, nationally and globally, including the reactions people should have and the way they should think about what is happening in the world (Brinkmann, 2012). This could mean people feel pressurised into reacting in particular ways by the social movement and countermovement organisations that compete for media access. In the case of climate change, public concern about the issue has been reduced due to the 'climate denier' movement (Brulle et al., 2012; McCright and Dunlap, 2011). Psychological barriers can also cause individuals to fail to act and engage with environmental problems (Gifford, 2011) leading to sociological ambivalence (Carolan, 2010). Debates about how 'fake news' and misinformation spread through social media (Kucharski, 2016) could further undermine trust in genuine news and information (Lazer et al., 2017).

Social media and Facebook

'Social media' incorporates social networking sites, content sharing platforms, collaborative websites, and chat applications (Mangold and Faulds, 2009). Everyday activities, such as knowledge sharing, have been changed by social media (Brinkmann, 2012). Use of social media in awareness campaigns can also create behaviour change, although this is not seen across all subject areas (Happer and Philo, 2016). Facebook interactions between publishers and users therefore have the potential to increase knowledge sharing and change behaviour.

Facebook has become a popular social communication tool for building social connections, sharing content and identities, and for social investigation (Gruzd and Haythornthwaite, 2008; Joinson, 2008). Individuals can use Facebook to connect with other users, share status updates and content, join groups, and follow pages to receive publishers' news and status updates (Caers et al., 2013). Facebook has the highest number of active user accounts with 1,871 million global users, over half of whom access the platform daily, and 42 million users in the UK (Kemp, 2017). Expanding features, such as Messenger and Facebook Live (Zuckerberg, 2016), meet users' needs, offer opportunities for organisations to communicate with users, and prevent environmental numbness through the use of creative media (Gifford, 2011). Businesses and organisations can also use Facebook to communicate with customers and stakeholders, share organisational updates, and cultivate and manage relationships (Caers et al., 2013; Mangold and Faulds, 2009; Nah and Saxton, 2016; Saxton and Waters, 2014).

Social media has created fragmented groups of users that identify with specific interests (Greenberg and MacAulay, 2009). These groups enable end users to share information rapidly, shifting the way that information is distributed away from dominant media, towards users themselves (Greenberg and MacAulay, 2009). This evolution in information–sharing has generated a need for organisations to change how they communicate with stakeholders; instead of being selective, communications should be open and conversational to inspire engagement and participation (Greenberg and MacAulay, 2009). Photographs, videos and blogs encourage content sharing, increase participation and collaboration (Greenberg and MacAulay, 2009), and are media that conservationists and environmental activists use to communicate with their audiences (Blewitt, 2011). Environmental publishers could therefore use Facebook's media sharing functionality to inspire action and invite collaboration through their online networks.

Social media and identity

Users' profiles, and the media shared through them, can either be private or public, enabling users to edit themselves and project a representation of self to contacts and the public. People present themselves as exhibitions (Horgan, 2010) and are curators of content (Brinkmann, 2012). Individuals also externalise the self through textual and visual representations (Illouz, 2007), suggesting that the self is made into a commodity which is displayed publicly. The desire to be seen in a certain way could cause social media users to edit and frame the content they share to fit with the personal exhibition they choose to present to other users (Brinkmann, 2012). In the case of Facebook, the media users share is the result of the curation process and is outside publishers' control. Being aware of factors that can limit or increase the likelihood of content being shared by users could be useful for publishers. By understanding their audience, publishers could identify users' motivations, tailoring content to achieve engagement. As noted by Blewitt (2011), knowledge of spectatorship and aesthetics, and their impact on users' perceptions of content, is necessary to achieve this understanding.

Environmental organisations and Facebook

Environmental Non–Governmental Organisations (ENGOs) play roles including lobbying, raising public awareness about issues and causes, and being involved in global environmental politics (Finger and Princen, 2013). Causes promoted by ENGOs include deforestation, marine pollution, and international development (Finger and Princen, 2013), climate change (Katz–Kimchi and Manosevitch, 2015), and corporate social responsibility (Hendry, 2006). The type and scope of activities ENGOs employ depends on their size, membership, and the mobility of their networks (Zito and Jacobs, 2009). ENGOs' activities can include forming lobbying alliances, creating accreditation schemes (Hendry, 2005), running workshops, producing educational documentation, and direct action campaigns (Zito and Jacobs, 2009). As social media could provide an informal platform for educating the public about sustainability (Andersson and Öhman, 2016; Ehrlich, 2011), ENGOs could use Facebook to meet their objectives.

Despite the potential for social media to raise awareness of brands and open dialogues with new stakeholders, environmental organisations have been criticised for failing to maximise the opportunities social media presents (Bortree and Seltzer, 2009; Greenberg and MacAulay, 2009; Nulman and Özkula, 2016). Environmental organisations are beginning to realise the potential for, and benefit of, direct engagement with users through Facebook. ENGOs use Facebook pages to share information relevant to their causes and to connect users to additional environmental resources (Stoddart and MacDonald, 2011). Facebook has also reduced ENGOs' dependence on conventional media, such as press releases, providing greater independence and increasing ENGOs' ability to communicate directly with users (Greenberg and MacAulay, 2009).

ENGOs use social media to mobilise their followers. New media technologies enable collective action (Stoddart and MacDonald, 2011), whilst Facebook features such as 'like', 'comment' and 'share' can be used to organise and mobilise users, as in Greenpeace's climate change campaign (Katz–Kimchi and Manosevitch, 2015). Greenpeace's 'e–tactics' demonstrated how the ENGO mobilises users to pressurise businesses to achieve action (Hendry, 2005). Katz–Kimchi and Manosevitch (2015) highlight the potential for ENGOs to use Facebook to mobilise their fans and use 'low–cost activism' to increase their communicative capacity (Stoddart and MacDonald, 2011).

Previous research

Waters et al. (2009) analysed the content of publishers' Facebook profile pages, but research into the content of publishers' posts is limited (Saxton, 2014). Textual analysis of posts and the ways end users react to and participate in Facebook communications (Saxton and Waters, 2014) are key areas for exploration. Analysis of stakeholder engagement could also answer questions about the effectiveness of social media communication strategies, as previous research has focused on what organisations publish rather than the results of their activities.

Saxton and Waters (2014) attempted to identify the types of Facebook posts that engage users. Engagement was measured according to the number of 'likes', 'shares' and 'comments' posts received. The researchers proposed an I–A–C Framework (information, action, community) to map posts according to their purpose. 'Information' posts published facts or news to increase users' knowledge; 'action' posts asked users to complete a task, such as signing a petition; whilst 'community' posts were more conversational, aiming to build the publishers' network of users. Saxton and Waters (2014) provided insight into whether post purpose influences user engagement.

Research aim

This research aims to ascertain whether environmental publishers use specific techniques in their content to engage Facebook users with sustainability. Through analysing posts shared by Greenpeace UK, WWF UK and the Guardian Environment on their public Facebook pages between 1st November 2015 and 31st October 2016, the research aims to identify features which can increase or limit the likelihood of users engaging with content. Greenpeace and WWF were selected as examples of ENGOs that engage in environmental activism and conservation. The Guardian Environment (the arm of the Guardian newspaper that focuses on environmental content) was included to explore whether more traditional media publishers that focus on environmental content could also contribute to citizenship and environmental literacy alongside ENGOs. Textual analysis was used to identify the type of media accompanying posts, post features, and themes in content, to build on previous research. Saxton and Waters' (2014) I–A–C Framework was also used, and adapted, to explore the roles that environmental publishers play, and the capacity these roles have for engaging Facebook users with sustainability. The research findings were used to create a checklist of recommendations for environmental publishers that could help increase users' engagement with sustainability.

Methodology

Research design

The research used an inductive approach to build theories based on analysis of the data collected (Bryman, 2012). The research paradigm was a combination of interpretivism, as the aims of the research related to understanding human behaviour (Bryman, 2012), and pragmatism, which is concerned with problem solving and the human experience (Morgan, 2014). Pragmatism is appropriate for this research as it considers the actions of the environmental organisations who publish information and the users who interact with it. A mixed–methods approach was applied, combining qualitative and quantitative research methods (Bryman, 2012; Patel, 2015). Inductive data analysis was carried out using data mining and discourse analysis to identify themes, before developing theories in line with grounded theory.

Ethical consideration

Social research taking place in internet contexts faces ethical dilemmas. Considerations include concerns about extracting personal information from social media sites through application programming interfaces (APIs) (Lomborg and Bechmann, 2012) and difficulties in gaining informed consent from participants who are located around the world (Markham and Stavrova, 2016). Platforms such as Facebook provide APIs to enable interaction with, and extraction of, data (Lomborg and Bechmann, 2012). As all data exported from Facebook for use in this research was publicly available and no individual Facebook users would be identifiable, obtaining consent was not necessary, however practitioners adapting this method should ensure their own approach adheres to ethical standards.

Population and sample

The target population for this research was UK environmental publishers with pages on Facebook. A purposive sample was strategically selected from the population to enable the data to answer the research questions (Bryman, 2012) based on publishers sharing at least one post per day and having more than 75,000 users following their page. The aim was to generate a dataset that would include sufficient data to identify trends in publishers' posting behaviours, and be big enough to achieve data and theoretical saturation, yet small enough to make a deep analysis possible (Onwuegbuzie and Collins, 2007, in Bryman, 2012). The sample period was from 1st November 2015 to 31st October 2016. The aim was to increase the 2–month sampling period used in Saxton and Waters' (2014) study, without generating too much data to analyse within the timeframe of the research.

The three sampled publishers were Greenpeace, WWF, and Guardian Environment. Blewitt (2011) discusses conservationists' and environmental activists' approaches to engaging the public with environmental subjects, including through use of visual media, whilst Happer and Philo (2016) discuss the role of the media in people's understanding of environmental issues. WWF and Greenpeace were selected as ENGOs who fulfil conservation and activism roles. Guardian Environment was chosen to discern whether they communicate about environmental issues through social media differently to the ENGOs, and whether they experience different levels of interaction from users.

Data collection

In line with Saxton and Waters' (2014) research, data was collected using a computer script to extract data from Facebook's database. Saxton and Waters (2014) used original Python code to download status updates from their target organisations' Facebook pages, including the 'likes', 'shares' and 'comments' posts received. Using scripts enables the visible traces of users' interactions to be analysed by researchers to study how social realities are displayed (Markham and Stavrova, 2016), making this approach appropriate for achieving the aims of the research project.

A custom script extracted data from Facebook including post comments and replies, reactions, shares, and accompanying text. The script included the suite of 'reactions' that enable users to specify emotions other than 'like', and facilitated textual analysis. Not all collected data was used, however additional comment and sentiment analysis could be carried out using this data. The code was written in Node.js; a JavaScript runtime program that can be used to build scalable network applications that perform specific tasks on an event–loop and call back designated data (Node.js Foundation, 2017). To extract the required data for the research project, the Node.js script communicated with Facebook's API to find and retrieve specific post attributes stored in Facebook's database. The Node.js script exported the data from Facebook into Postgres – an open source database which can be queried using structured query language (SQL). SQL queries enabled publisher data to be grouped before exporting it to Microsoft Excel for cleansing and interrogation.

To verify the accuracy of data extraction, manual comparisons of 50 status updates per publisher were made against the original Facebook posts in line with a trial completed by Saxton and Waters (2014). The test ensured that no posts were omitted during data extraction, as Facebook limits the number of requests that can be made through its API in one session (Lomborg and Bechmann, 2012). Although the script was designed to overcome the problem of omitting data, the manual check prevented inaccuracies in the dataset.

Data analysis

The majority of analysis was qualitative content analysis, using coding and thematic analysis to explore data and generate theories. The posts that received the highest and lowest engagement levels were identified for each publisher, with engagement measured according to the number of interactions that posts received as in Saxton and Waters (2014). The 20 most commented–on, reacted–to and shared posts were analysed and compared to the 20 least commented–on, reacted–to and shared posts. The sample size for 'high' and 'low' engagement posts was based on the same premise as the complete dataset, enabling data and theoretical saturation whilst facilitating deep analysis within the timeframe of the research (Brinkmann, 2012; Onwuegbuzie and Collins, 2007, in Bryman, 2012).

When generating the 'low engagement' sample, some publishers had more than 20 posts with no interactions. Random numbers were therefore generated in Excel using the RAND() function with values assigned to each post and the lowest values forming the sample. Some posts also appeared in multiple 'high' or 'low' engagement categories (comments, reactions or shares), indicating that the sampling process could have caused lower numbers of duplicate 'low engagement' posts to be selected for analysis.

Saldaña (2009) discusses the appropriate amount of data to code from the data corpus, suggesting that emphasis should be on the quality of data over quantity. Coding was therefore limited to identifying textual features and post themes in the message text, and grouping posts in line with Saxton and Waters' (2014) I–A–C Framework. 'First cycle coding' identified 'sub–themes' which were grouped into categories during 'second cycle coding'. Categories were grouped further under the pillars of sustainability – environmental, social and economic (Gibson, 2006).

Posts were classified according to the I–A–C Framework (Saxton and Waters, 2014) depending on whether their main purpose was sharing information, inspiring action or community–building. This classification enabled links between post purposes and the roles environmental publishers adopt through Facebook to be explored. The analysis showed that publishers frequently combine these purposes, leading to an adapted I–A–C Framework being applied in the research (e.g. I+A posts).

Tests for statistical significance were carried out on I–A–C data. Anderson–Darling tests showed normal distribution of interactions. Two–sample t–tests were used to ascertain whether interactions for framework categories were statistically significant between 'high' and 'low' engagement samples (Chew, no date). One–way ANOVA tested for significant differences (Laerd Statistics, 2013) in the number of reactions received by all 'high engagement' posts, most commented–on posts, most shared posts, and most reacted–to posts, according to their I–A–C framework category.

Results

Total posts and interactions

The number of posts published by each publisher varied considerably (Table 1). Greenpeace published the most posts with an average of 152 posts per month. Guardian Environment published a mean of 133 monthly posts, whilst WWF published an average of 64 posts per month. The number of monthly posts published by WWF was the most consistent, with a range of 25, whilst Guardian Environment's range was highest at 275. Greenpeace consistently published more posts than WWF and Guardian Environment between November 2015 and July 2016. Guardian Environment's post frequency increased from June 2016, peaking at 333 posts. Greenpeace's and WWF's post frequency peaked in October 2016 at 187 posts and 77 posts respectively.

Total posts published

Mean number of monthly posts

Minimum number of monthly posts

Maximum number of monthly posts

Range of monthly posts

Greenpeace UK

1,825

152

113

187

74

WWF UK

768

64

52

77

25

Guardian Environment

1,598

133

58

333

275

Table 1: The total, mean, minimum, maximum and range of monthly Facebook posts published by Greenpeace, WWF and Guardian Environment between 1st November 2015 and 30th October 2016.

Figure 1 also shows the mean number of interactions publishers' posts received each month. Guardian Environment's engagement peaked in December 2015 with 41,350 total interactions. The increase in post frequency from July causes Guardian Environment's mean number of interactions–per–post to drop despite receiving three of their highest monthly interaction rates in August, September and October. Between February and September, WWF's average number of interactions–per–post drop, reaching 1,529 in July. WWF's mean number of interactions–per–post peak in September at 3,981 before dropping back to 1,603 in October. Greenpeace have the highest average interactions–per–post in all months except September, although Greenpeace's total interactions still exceeded WWF's, with the mean number of interactions–per–post peaking at 4,361.

Figure 1: The number of Facebook posts published by Greenpeace UK (solid black line) WWF UK (solid dark grey line) and Guardian Environment (solid light grey line) between 1st November 2015 and 31st October 2016 mapped to the mean number of interactions each post received (dotted lines in same colour).

The posts from each publisher's peak month for total interactions were identified to enable analysis of content responsible for increased engagement levels. One main post received comparably higher interactions in WWF and Guardian Environment's peak months, whilst Greenpeace's peak month included multiple posts with more than 29,000 interactions. Coding of accompanying text and media revealed themes and post purposes that could influence engagement.

Post features

Types of media

Posts were accompanied by different types of media: events, links, photos, statuses and videos. Events and statuses were not used frequently but did occur in the full dataset. 'Events', 'photos', and 'videos' are self–explanatory; 'links' shared content from external websites, whilst 'statuses' were text–only updates.

In the entire dataset, links were the most common media for Guardian Environment and Greenpeace (Figure 2a) and photos were most common in WWF's posts. Photos accompanied more 'high' than 'low' engagement posts (Figure 2b and 2c). Videos accompanied a higher percentage of 'low' than 'high' engagement posts for WWF and Greenpeace, but Guardian Environment differed, with one video in their 'low engagement' sample, and 8 of their 14 video posts in the 'high engagement' sample. Further analysis of video posts showed that many Greenpeace videos were shared from other publishers' pages. The original content received 230–422,983 shares, which could indicate that shares were being attributed to the original publishers. WWF's least shared videos were mainly uploaded to their Facebook video section and re–posted, which reduced the interactions recorded in the data.

Statistical tests on the types of media accompanying posts generated some p–values that were close to the 95% confidence interval, but there were no statistically significant differences between group means as determined by one–way ANOVA in each instance.

Figure 2: Percentages of posts published by Greenpeace UK, WWF UK and Guardian Environment on Facebook between 1st November 2015 and 31st October 2016. (a) Percentage of total Facebook posts published by each sampled publisher between 1st November 2015 and 31st October 2016 accompanied by each type of media. n=1825, 768, and 1598 for Greenpeace UK, WWF UK, and Guardian Environment respectively. (b) Percentage of 'high engagement' posts published by each sampled publisher between 1st November 2015 and 31st October 2016 accompanied by each type of media. n=60 (c) Percentage of 'low engagement' posts published by each sampled publisher between 1st November 2015 and 31st October 2016 accompanied by each type of media. n=60

Textual features

Six textual features were identified – emojis, hashtags, embedded links, questions, quotes, and accompanying text (Figure 3). Greenpeace and WWF use a wider range of textual features than Guardian Environment, with most features appearing in both 'high' and 'low' engagement samples. Posts with no accompanying text only appear in the 'low engagement' sample for all publishers. Quotes are present in Greenpeace's 'high engagement' posts and more of Guardian Environment's 'high' than 'low' engagement posts.

Emojis, hashtags and embedded links appear across both 'high' and 'low' engagement samples for Greenpeace and WWF but do not appear in the Guardian Environment samples (Figure 3c). Questions were used by all publishers, although Guardian Environment made minimal use of this feature.

Figure 3: Percentages of 'high' and 'low' engagement posts published on Facebook by Greenpeace UK, WWF UK and Guardian Environment between 1st November 2015 and 31st October 2016 that include each textual feature broken down by interaction type (comments, reactions, and shares). (a) Percentage of Greenpeace UK's 'high' and 'low' engagement Facebook posts for each interaction type that include each type of textual feature. (b) Percentage of WWF UK's 'high' and 'low' engagement Facebook posts for each interaction type that include each type of textual feature. (c) Percentage of Guardian Environment's 'high' and 'low' engagement Facebook posts for each interaction type that include each type of textual feature.

Links inspiring action

In both 'high' and 'low' engagement samples, embedded links were associated with calls to action as shown in Table 2. The Guardian is not included in Table 2 as their sampled posts contained no embedded links.

Publisher

'Action'

Number of Posts

High Engagement

Greenpeace

Sign petition

16

Pledge / Join movement

3

WWF

Adopt animal

7

Support campaign

1

Total

27

Low Engagement

Greenpeace

Find out information / how you can help

11

Sign petition

5

Send email / letter

2

Call

2

Tweet

1

Vote

1

Pledge / Join movement

1

WWF

Support campaign (Wear it Wild)

6

Download app

3

Do sponsored event

1

Make something and share

1

Attend event

1

Send email

1

Total

36

Table 2: Actions users were asked to perform through embedded links in message text in 'high' and 'low' engagement posts published by Greenpeace UK and WWF UK between 1st November 2015 and 31st October 2016. Guardian Environment's 'high' and 'low' engagement samples included no embedded links.

Canvassing users

Questions were used in the accompanying text to 'canvass' users, encouraging interaction through 'comments' by asking for users' ideas or opinions (Table 3).

Publisher

Question

Comments

Greenpeace

Do you think disposable coffee cups should be taxed?

853

Greenpeace

Many of the laws that protect our oceans, keep our air cleaner and our climate safer were made by the EU. Now that we've voted to leave the EU, what do you think we should do to protect these things? Pollution and climate change don't take note of borders so we reckon we still need to work together.

1,266

Greenpeace

It's World Penguin Day! What's your penguin name?

7,992

WWF

Caption time! What's your caption for this panda chilling at the top of a tree?

1,112

WWF

Aaaaand... WEEKEND. Can you come up with a caption for this leopard? Tell us your suggestions in the comments. #LoveNature

363

Table 3: Examples of Facebook posts from the 'high engagement' sample published by Greenpeace UK and WWF UK between 1st November 2015 and 31st October 2016.

Thematic analysis

Post categorisation

Thematic analysis generated 42 sub–themes for 'high engagement' and 61 sub–themes for 'low engagement'. Grouping sub–themes into categories gave 15 'high' and 17 'low' engagement categories (Figure 4). 'Technology' and 'Organisation News' only appeared in the 'low engagement' sample. The number of categories differed between publishers, with WWF content covering the fewest categories. Higher numbers of post categories appeared in all publishers' 'low engagement' samples.

WWF had the highest number of posts in a single category in both 'high' and 'low' engagement samples. 'Natural Environment', 'Conservation and Biodiversity' and 'Organisation Activities' were the most common categories for both WWF's 'high' and 'low' engagement samples. Greenpeace's two most common 'high engagement' categories – 'Energy and Resources' and 'Natural Environment' – were also most frequent in the 'low engagement' sample. The most frequent categories in Guardian Environment's 'high' and 'low' engagement samples differed; 'Consumption and Waste', 'Energy and Resources' and 'Environmental Exploitation' posts were most common in their 'high engagement' sample, whilst 'Natural Environment' was their most frequent 'low engagement' category.

Figure 4: Number of posts published by Greenpeace UK, WWF UK and Guardian Environment between 1st November 2015 and 31st October 2016 that align with each content category. (a) The combined most commented–on, reacted–to, and shared posts for each publisher mapped to each content category. (b) The combined least commented–on, reacted–to, and shared posts for each publisher mapped to each content category. n=60 for each publisher.

Peak categories

Peaks in 'Natural Environment' and 'Organisation Activities' posts highlighted these categories for deeper analysis. Table 4 shows that images from the natural world were used to entertain.

Publisher

Message Text

I–A–C Category

High Engagement

Greenpeace

It's World Penguin Day! What's your penguin name?

I C

WWF

Cheer yourself up this morning with a caption comp! Get your thinking cap on, and see if you can come up with a genius caption for this polar bear. Write your suggestions in the comments below.

C A

WWF

Happy Valentines' Day! [Photo of elephants]

C

Low Engagement

Greenpeace

Some of the animal kingdom's dads are strict, some are protective, and some are even big into gender equality! Which animal dad is your father most like? Take our Father's Day quiz to find out

A C

Greenpeace

Time to reawaken your sense of wonder [National Geographic Photographs]

C

WWF

Wow! Look at these wonderful gorilla twins swinging from vines and playing in the jungle. #DYK that although gorillas are excellent climbers, only 3% of their daily activity takes place in trees? We hope you're having a lovely weekend with family too.

I C

Table 4: Examples of 'high' and 'low' engagement posts from the 'Natural Environment' category which demonstrate that publishers use animals to entertain users. Posts are mapped to the adapted I–A–C Framework (Saxton and Waters, 2014), according to whether posts share information (I), ask users to complete an action (A), aim to build community (C), or perform a combination of these roles.

Table 5 shows a breakdown of 'Organisation Activities' posts published by WWF. This category comprises the 'fundraising' and 'activism' sub–themes, and used events like Mother's Day to engage users.

Publisher

Message Text

Theme

High Engagement

WWF

Tomorrow the world will be switching off... are you ready? http://po.st/EH2016Signup From 6.30am UK time tomorrow, the wave of switch–offs will begin as Earth Hour celebrations start spreading around the world. Don't miss your chance to be part of it. Simply switch off your lights for one hour at 8.30pm local time tomorrow night – add your name to be counted! http://po.st/EH2016Signup

Activism

WWF

A beautiful way to say thank you – give an animal adoption as a Mother's Day gift: http://po.st/UOXxv7 Snow leopard cubs are born in spring or early summer, and their mother keeps them safe in a well–concealed den. They'll stay by her side until they become independent at around 18–22 months. To say a special thank you to your mum for all her love and care, adopt a snow leopard for her this Mother's Day. She'll receive an adoption pack and regular updates about our work to protect snow leopards. Adopt today: http://po.st/UOXxv7

Fundraising

Low Engagement

WWF

Up for a challenge? We're creating a whole flock of origami birds for a very special reason – to protect a treasured World Heritage site. Reckon you can help us?

Activism

WWF

Calling all wild things! Wear it Wild on Friday 27th May and help raise funds for amazing species http://po.st/Wild2016 Time to stick your neck out and have some fun!

Fundraising

Table 5: Examples of 'high' and 'low' engagement posts from the 'Organisation Activities' category, which demonstrate that WWF uses Facebook to encourage users to engage with their objectives.

Pillars of sustainability

Mapping categories to the pillars of sustainability showed that 'environmental' posts were most common in both 'high' and 'low' samples when all data is combined. 'Economic' posts were more frequent in the 'high' than 'low' engagement sample for all data, whilst more 'social' posts appeared in the 'low' than 'high' engagement sample (Figure 5). When the individual publishers' posts are considered, 'economic' posts are most frequent in the 'high engagement' sample for Greenpeace and Guardian Environment. 'Economic' posts did not appear in WWF's 'high engagement' sample; 'environmental' themes dominated their content with 'social' posts as the remainder. 'Economic' posts were least common in WWF's 'low engagement' posts, whilst 'environmental' posts accounted for approximately one third, and 'social' themes made up nearly two thirds of posts. 'Environmental' posts accounted for over half of Greenpeace and Guardian Environment's 'low engagement' samples, but less of their 'high engagement' posts. 'Social' posts comprised 20% and 28.3% of Greenpeace and Guardian Environment's 'low engagement' posts respectively, whilst 28.3% and 20% of posts were 'economic'.

ANOVA showed there were no statistically significant differences between group means for the percentage of posts that aligned with each pillar of sustainability ('high' engagement – F–value = 2.21; p–value = 0.190; 'low' engagement – F–Value = 2.15; p–Value = 0.197).

Figure 5: Percentage breakdown of 'high' and 'low' engagement Facebook posts published by Greenpeace UK, WWF UK and Guardian Environment between 1st November 2015 and 31st October 2016 grouped according to the three pillars of sustainability: social, economic and environmental.

I–A–C framework

Categorising posts according to Saxton and Waters' (2014) framework highlighted the need to include new categories to show grouped framework functions. The grouped categories accounted for approximately half of the sampled posts, with 49.4% of 'high engagement' and 50.0% of 'low engagement' posts combining multiple I–A–C categories.

Figure 6 shows that 'information' posts were most common in both 'high' and 'low' engagement samples when data for the three publishers is combined. Action posts were least common for 'high' and 'low' engagement posts, with twice as many 'low engagement' 'action' posts. There were 1.7% fewer 'high engagement' than 'low engagement' posts under the 'community' category. The 'high engagement' sample contained 5.6% more 'information and action' and 2.8% more 'information and community' posts than the 'low engagement' samples, whilst 7.2% more 'action and community' and 1.7% more 'information, action and community' posts were in the 'low engagement' sample than the 'high engagement' sample. P–Values in 2–sample t–tests showed that the differences between 'high' and 'low' engagement for each framework category were not statistically significant.

Figure 6: Comparison of 'high' and 'low' engagement Facebook posts published by Greenpeace UK, WWF UK and Guardian Environment between 1st November 2015 and 31st October 2016 mapped according to posts being information based (I), requesting users complete an action (A), attempting to build community (C), or a combination of these roles, in line with the adapted I–A–C Framework (Saxton and Waters, 2014).

Figure 7 provides a breakdown of 'high' and 'low' engagement posts per publisher mapped to the adapted I–A–C Framework. WWF posts appear across all categories and combinations, and they are the only publisher with 'high engagement' posts in the 'action and community' and 'information, action and community' categories (Figure 7a). 'Information and action' is Greenpeace's most frequent 'high engagement' category, whilst 'action' posts are present but the least frequent. Guardian Environment posts are only found in three categories, with 76.7% 'information', 21.7% 'information and community', and 1.7% 'information and action' posts. Guardian Environment's 'low engagement' posts also encompass three categories, comprising 75% 'information', 1.7% 'action', and 23.3% 'information and community' posts (Figure 7b). Greenpeace and WWF 'low engagement' posts appear in all categories (Figure 7b). The highest percentage of Greenpeace's 'low engagement' posts are 'information and action' posts at 31.7%. WWF's largest 'low engagement' category is 'action and community' at 25%, whilst information posts account for 20% of their 'low engagement' posts.

Figure 7: Percentage of Facebook posts published by Greenpeace UK, WWF UK, and Guardian Environment between 1st November 2015 and 31st October 2016 that align with each category of the adapted I–A–C Framework (Saxton and Waters, 2014) according to whether posts share information (I), ask users to perform an action (A), build community (C), or perform a combination of these purposes. (a) Percentage of each publishers' 'high engagement' Facebook posts mapped to the adapted I–A–C framework categories. (b) Percentage of each publishers' 'low engagement' Facebook posts mapped to the adapted I–A–C framework categories.

Figure 8a shows the percentage of interactions each publisher's 'high engagement' posts received. Of Guardian Environment's interactions, 51.8% were for 'information' posts, 47.4% were for 'information and community' posts, and 0.7% were for 'information and action' posts. WWF also received their highest proportion of interactions for 'information' posts, with 'information and action' posts comprising 15.1% of interactions, and 'information, action and community' posts accounting for 13.9% of interactions. WWF received interactions for all combinations of categories. 'Information' posts were Greenpeace's second–highest category, receiving 26.9% of their interactions, whilst 'information and action' posts accounted for 40.1% of Greenpeace's engagement. Greenpeace received 2.9% of their interactions for 'action' posts and 11.5% of their interactions for 'community' posts; the two lowest framework categories in the 'high engagement' sample. 'Information and community' posts received 18.7% of Greenpeace's interactions.

Figure 8b breaks down individual interactions by I–A–C category. Over 70% of Guardian Environment's comments and reactions were for 'information' posts. 'Information and community' posts received 65.1% of shares, with 1.4% of shares being for 'information and action' posts and the remaining 33.6% of shares for 'information' posts. Over 50% of WWF's comments and shares and 46.9% of reactions are for 'information' posts. WWF combine 'action' posts with other functions, with the collective combinations making up 32.7%, 30.6% and 33.6% of their comments, reactions and shares respectively. Greenpeace received more interactions for combined categories, with 'information and action' and 'information and community' posts making up 67.4% of comments, 53.6% of reactions, and 64% of shares.

Figure 8: Percentage of interactions that 'high engagement' posts published by Greenpeace UK, WWF UK and Guardian Environment on Facebook between 1st November 2015 and 31st October 2016 for each adapted I–A–C Framework category (Saxton and Waters, 2014). Posts were categorised according to whether their purpose was to share information (I), ask users to complete an action (A), to build community (C), or a combination of these purposes. (a) The combined percentage of each publishers' 'high engagement' sample that aligns with each adapted I–A–C Framework category. (b) Each publishers' 'high engagement' posts broken down by engagement type (comments, reactions and shares) mapped to the adapted I–A–C Framework categories.

ANOVA showed there were no statistically significant differences between group means for each adapted I–A–C framework category (Table 6).

Framework Category

t–Value

p–Value

I

0.17

0.877

A

–1.07

0.345

C

–0.29

0.793

I + A

0.40

0.719

I + C

0.37

0.735

A + C

–0.79

0.486

I + A + C

–0.20

0.852

Table 6: P–values and t–values from 2–sample t–tests carried out on 'high' and 'low' engagement samples to ascertain whether the number of interactions received by posts in each adapted I–A–C Framework category (Saxton and Waters, 2014) was statistically significant.

Discussion

Unique posts

The appearance of duplicate posts in 'high engagement' samples could have been caused by Facebook's News Feed algorithm. Users' News Feeds are populated by content shared by friends and pages (Patel, 2016), whilst Facebook's algorithm attempts to show users posts that are most relevant to them (Boland, 2014). Users' interactions with posts feed into the algorithm, affecting posts' visibility and organic reach (Patel, 2016). Posts receiving higher engagement are exposed to wider audiences, giving more opportunities to achieve further engagement, whilst lower performing posts can struggle to attract interactions. Facebook's algorithm can also create 'filter bubbles' that are biased towards topics users previously engaged with, preventing users from getting a 'bigger picture' of news and events (Adee, 2016).

Total posts and interactions

Post frequency and overall engagement levels

It is important that publishers are active on social media, publishing fresh, interesting and interactive content (Kaplan and Haenlein, 2010). Although Greenpeace shared the most posts and received the most interactions, post frequency alone does not determine engagement. On some occasions, Guardian Environment's interactions increased when post frequency increased, however WWF's interaction numbers show that fewer posts can elicit higher engagement. Other factors could therefore influence engagement more than post frequency.

Peaks in engagement

Peaks in mean monthly interactions could indicate content that is more likely to engage users, or that posts receive consistently high engagement levels. Each publisher's most interacted–with posts are explored below.

WWF's most engaged–with post celebrated the result of conservation efforts and was accompanied by a photo of a giant panda. This 'information' post (Saxton and Waters, 2014) maps to the 'environment' pillar of sustainability and uses the commodification of the panda (Castree, 2003) to promote WWF's news. Promoting the positive news and result could inspire users to contribute to conservation efforts by adopting other species, enabling WWF to achieve their offline objectives. Whether WWF deliberately intended to encourage adoptions without explicitly directing users to sponsor animals is unknown.

Guardian Environment's most engaged–with post was published ahead of the COP21 climate negotiations. Relating to indigenous people's rights, this post fell under the 'social' pillar of sustainability. The post used the forest to underpin the publisher's message so could also be classed as commodifying nature (Castree, 2003). The post fits the 'information and community' category of the I–A–C Framework (Saxton and Waters, 2014) and performs the educational role often adopted by environmental organisations (Andersson and Öhman, 2016; Zito and Jacobs, 2009).

Greenpeace's most engaged–with post from their peak month celebrated the publisher's campaign against Tesco stocking John West tuna. The post was categorised as 'environmental exploitation' and grouped under the 'environmental' pillar of sustainability (Gibson, 2006). Greenpeace's photograph of a turtle in a clean ocean commodifies nature (Castree, 2003) to promote their cause; the publisher could have used a photo of tinned tuna but chose a 'happy' turtle in an unpolluted environment. This 'action and information' post (Saxton and Waters, 2014) may have been designed to encourage further action, as the accompanying text urged users to pressurise Sainsbury's too. The post is an example of Greenpeace using communication strategies to inform their networks about companies to target through campaigns (Hendry, 2005). Greenpeace's second–most engaged–with post from their peak month was a 'social', 'information and community' (Saxton and Waters, 2014) post, which showed solidarity with victims of the Baghdad terror attack. Greenpeace's most engaged–with post from the whole sample featured fracking and was categorised as 'Energy and Resources', grouped under the 'economic' pillar of sustainability, and performed an 'information and action' role (Saxton and Waters, 2014).

Peak engagement summary

The most engaged with posts i) are accompanied by the three main types of media (photos, videos and links), ii) align with all three pillars of sustainability, and iii) include a combination of 'information', 'information and action', and 'information and community' posts. The combination of environmental, social, and economic posts demonstrates ENGOs attempt to address all three pillars of sustainability (Alcock, 2008) and indicates that publishers are meeting the challenge of reconciling these differences in ways that Facebook users respond to.  

Influence of post features

Accompanying media

Images and videos accompanied many of WWF and Greenpeace's posts, supporting the literature (Blewitt, 2011; Greenberg and MacAulay, 2009; Malhotra et al., 2013; Saxton and Waters, 2014), however Guardian Environment made minimal use of visual media. The number of videos in Guardian Environment's 'high engagement' posts suggest that more videos could increase their interactions. Based on the number of photographs accompanying Greenpeace and WWF's 'high engagement' samples, Guardian Environment could trial more image–based content rather than focusing on publishing links to news articles.

The high numbers of videos in Greenpeace and WWF's 'low engagement' samples were unexpected, although the additional investigation provided an explanation for lower engagement with video content. Greenpeace and WWF could therefore be receiving higher engagement for video content than Figure 2 shows. As Patel (2016) highlights that Facebook's algorithm tracks the amount of time users spend engaging with content as well as the number of 'click' interactions posts receive, video content is particularly worthy of consideration for publishers. To mitigate against the problem of contributing to other publishers' data, publishers could create their own video content which other publishers would share, offsetting the lost interactions.

Textual features

All publishers have 'low engagement' posts with no accompanying text whilst all 'high engagement' posts are accompanied by text. Including text alongside posts enables publishers to embed other features and engage with users (Bortree and Seltzer, 2009; Campbell et al., 2014; Kaplan and Haenlein, 2009) making 'accompanying text' an essential recommendation.

Emojis and hashtags could be classed as informal textual features. As content is shared in an online environment, users may expect social media posts to include these features. Hashtags were originally used on Twitter to increase keyword visibility in searches (Twitter, 2017). Hashtags are now also used in Facebook posts, perhaps initially through organisations coordinating their media messages across online platforms and offline activities as recommended by Kaplan and Haenlein (2009). Using informal features enables publishers to avoid being too professional (Kaplan and Haenlein, 2009; Malhotra et al., 2012), helping to build rapport with users. Guardian Environment could try incorporating emojis and hashtags to ascertain whether their users engage with these common features of social media posts.

There could be two main explanations for the appearance of quotes in more 'high' than 'low' engagement posts. Pull quotes from articles could entice users to read the full article by adding more detail than headlines offer. Quotes could also show expert opinion or celebrity endorsement. Branigan and Mitsis (2014) suggest that celebrity endorsement can help promote causes to Generation Y (the population born between 1977 and 1995). Stoddart and MacDonald (2011) also discuss the use of celebrity and experts to lend credibility to conservation activities.

Embedded links in message text shows publishers' attempting to use Facebook to promote activism. Stein (2009) discusses environmental organisations' motivations when using new media; emphasis is placed on communicating information, generating resources and mobilising users – actions promoted through the embedded links in Table 2. The calls–to–action in the 'high engagement' embedded links are actions that can easily be carried out online, and could be classed as 'clicktivism'. The 'low engagement' sample on the other hand includes activities that require more effort from users. This could support criticism that 'clicktivism' erodes offline activism, with users believing they have done their part by sharing posts with their network (Karpf, 2010; White, 2010).

De Vries et al. (2012) promote the use of questions as a mechanism for engaging users on Facebook, especially to increase comments. Canvassing questions were included in posts that received 16,058 comments, with 40% of Greenpeace's and 65% of WWF's most commented–on posts using canvassing. The results therefore support De Vries et al. (2012), showing that questions are valuable for increasing dialogic exchanges with users. Although Cvijikj and Michahelles (2013) recommended avoiding questions in favour of photos, the sampled publishers' questions tended to be accompanied by photos of either entertaining or provocative subjects. As photos provide more 'vivid' content (Cvijikj and Michahelles, 2013), combining questions with images could be responsible for higher engagement than if questions were text–only status updates.

Thematic analysis

Content categories and pillars of sustainability

Thematic analysis highlighted many sub–themes across the three publishers' 'high' and 'low' engagement posts. The fact that many categories appeared in both 'high' and 'low' engagement samples shows that the most common 'low engagement' categories should not be avoided completely. Categories that appear most frequently could simply be favoured by publishers based on their organisational objectives.

Grouping posts according to the pillars of sustainability highlighted differences between publishers' content. Greenpeace and Guardian Environment receiving higher engagement for 'economic' content could support Hendry's (2005) findings that Greenpeace mobilises networks to target economic concerns. The lack of 'economic' content in WWF posts could indicate that fewer posts in their entire dataset were 'economic' themed, whilst social posts were mainly organisation–centric. As noted by Zito and Jacobs (2009), WWF should cover a wider breadth of issues in line with their broad aims, whilst the data supports Alcock's (2008) findings that ENGOs can struggle to reconcile the three pillars of sustainability. The difficulty for publishers is that although they control the content they share, they may favour content that users are likely to engage with. The messages that environmental publishers communicate could therefore become moderated (unintentionally) by users.

Commodifying nature to entertain users

The 'Natural Environment' posts in Table 4 show how environmental publishers commodify nature (Castree, 2003) to entertain their users. Playing the role of 'entertainer' could be part of publishers' strategy to increase the reach of posts which promote the organisations' activities. Through 'reactions', 'shares' or 'comments', users can promote content to their own networks, potentially generating 'page likes' for publishers. By creating more engaging content that users react to, publishers could cause their users' Facebook algorithms to adapt, making their future posts more likely to appear in their users' news feed (Boland 2014; Patel, 2016). Although the commodification of nature can be criticised for assigning economic value to the environment (Liverman, 2004; McCauley, 2006), it could be a valuable technique for environmental publishers to increase the visibility of their content.

Using events to promote participation

Table 5 shows how WWF's 'Organisation Activities' posts used events to encourage fundraising and activism. By linking their products to events, WWF are disguising their appeal for users to contribute by making their fundraising attempts more indirect. Greenpeace's inclusion of links to petitions with informational posts could be classed as a similar strategy; both approaches aim to increase engagement with the ENGOs' objectives without directly asking users to perform a specific action. If the strategies convert to engagement from users, criticisms about 'clicktivism' could be countered (e.g. Karpf, 2010; White, 2010) (Halupka, 2014).

I–A–C framework

Adapting Saxton and Waters' (2014) I–A–C framework enabled exploration of the more nuanced nature of publishers' content. Although around half of posts were single–purpose, Figure 7 showed that the number of interactions received for each framework category varied by publisher. Interactions for combined categories were higher for Greenpeace, indicating that Greenpeace should consider greater use of combined categories. WWF received fewer interactions for combined categories, with 'information' posts accounting for the highest percentage of their interactions. Publishing more 'information' posts could increase WWF's engagement levels. Although Guardian Environment shared more 'information' posts, the number of interactions received for 'information and community' posts showed that more 'community–building' posts could help increase their engagement levels, particularly in terms of 'shares'.

The three main categories of the I–A–C Framework link to roles that publishers play to engage their users with sustainability:

'Information' as education

All publishers share 'information' content, whether as single or combined category posts. The use of 'information' posts indicates that publishers play an educational role through Facebook, supporting suggestions that social media could increase environmental literacy (Ehrlich, 2011) and deliver education and citizenship (Andersson and Öhman, 2016). Although the effectiveness of communications with these aims would depend on the receptiveness of publishers' audiences, the engagement levels for 'information' posts suggests that Facebook users interact with sustainability information, participate in discussions on sustainability themes, and share sustainability content with their networks.

'Action' for activism

'Action' posts are among the least frequent in either sample, however when combined with 'information' and/or 'community' purposes, are more abundant. 'Action' posts enable publishers to ask users to sign petitions, fundraise, or join campaigns. Publishers play the role of activists through 'action' posts, encouraging their networks to contribute to campaigns (Hendry, 2005; Katz–Kimchi and Manosevitch, 2015).

'Community–building' through entertainment and canvassing

By adopting 'entertaining' and canvassing' roles, publishers take an active approach in building their community and eliciting engagement through Facebook's commenting feature. Canvassing involves users in debates (De Vries et al, 2012) and creates open discussions which can encourage participation (Greenberg and MacAulay, 2009), building online environments where like–minded users can debate sustainability (Bortree and Seltzer, 2009).

Recommendations for publishers

Many features could affect user engagement with publishers' Facebook posts making it impossible to know categorically what will convert to higher interactions. Trends in the data have informed the following recommended strategies however, which aim to improve environmental publishers' Facebook posts to successfully promote sustainability.

  1. Post quality over quantity
  2. Accompanying text and visual media attract users' attention
  3. Hashtags and emojis do not hurt
  4. Self–promotion can disengage users
  5. Canvassing invites comments, providing opportunities for dialogue
  6. Entertainment supports audience growth
  7. Information posts can educate
  8. Action posts can inspire activism
  9. Community–building can be achieved using canvassing and entertainment

Limitations

Various limitations could present opportunities for future research. The first limitation was caused by the need to complete qualitative research to sufficient depth (Onwuegbuzie and Collins, 2007, in Bryman, 2012). The three sampled publishers were major publishers from the UK with established networks of users, which could make the recommended features and strategies, such as canvassing, less practical for publishers with smaller followings. As the sample also focused on UK publishers, recommendations may not transpose to international and cultural contexts.

The 12–month sampling period could be a further limitation; collecting data over a longer period could determine how post frequency and interactions have changed since each publisher joined Facebook. As Facebook features evolve however, the value of collecting data over a longer period is debatable; old features may not be comparable to new functionality.

A final limitation is that data was limited to 'click' engagements. Social media users vary in activity levels, some never participating actively (Lomborg and Bechmann, 2012) making their engagement difficult to measure. Publishers' Facebook Insights could show post reach, enabling analysis of the total number of users who saw content rather than just those who interacted.

Future research

The research highlighted additional areas that future research could explore:

External influences

Textual analysis could identify whether external events and current affairs influence publisher content, and whether the level of engagement that current affairs content receives differs to content created by publishers.

Coding visual media

Videos and photographs could be coded to identify trends in media content (Emmison, 2016; Saldaña, 2009).

Sentiment analysis

Sentiment analysis (Pang and Lee, 2008) could provide insight into how posts influence users' reactions. Message text, 'reactions' and user comments could be coded to assess whether content receives positive or negative reactions.

Comment analysis

Comment analysis could identify whether publishers enter into dialogic conversations with users (Bortree and Seltzer, 2009; Meredith, 2016) and the resulting effect on interactions.

Conclusions

The research identified ways that environmental publishers use Facebook to engage users with sustainability. The content of posts included variables that could influence the amount of engagement publisher's posts received. A number of factors could therefore be responsible for the variance in interactions between the sampled publishers, however the following conclusions can be drawn from the research.

'Page likes' affect reach

The higher a publisher's number of 'page likes', the larger the pool of potential Facebook users who could see the publisher's posts in their News Feed. Increasing the number of 'page likes' therefore facilitates exposure to new audiences and gives more opportunities to build networks of like–minded users (Bortree and Seltzer, 2009). Creating content that users engage with is therefore key to providing more opportunities for publishers to expand their overall network, and visibility, through Facebook.

Features impact potential engagement

Facebook is a visual platform making images and video effective for engaging users; Guardian Environment in particular could make better use of visual media in their posts. Message features, including the text accompanying posts, are another crucial way to attract users' attention and provide context for content. Without message text, publishers can lose opportunities to open conversations, increase participation and activism, and build their virtual community. Emojis and hashtags increase the conversational nature of posts, show sentiment and enable alignment with other on– and offline campaigns. Quotes can demonstrate endorsement for campaigns or entice users to access additional content, boosting engagement levels. Post features should therefore be carefully considered, both in terms of conversion to interactions, and in achieving the wider aims of publishers' communication strategies.

Connecting the 'Three Pillars of Sustainability'

Although publishers' audiences can be more receptive to particular topics making it tempting for publishers to focus on 'safe' categories, balance between influencing and being influenced by users is important. Increasing the number of 'social' posts that focus on themes such as equality and equity could increase literacy around such topics, whilst 'economic' posts, especially for WWF, could develop users' awareness beyond 'environmental' themes.

Publishers commodify nature

All publishers showed examples of commodifying nature in their posts, whether as 'products' to raise money through adoption, causes to protect, or as subjects of entertaining images and videos. Where commodification is used to achieve organisational aims relating to sustainability, this practice could be less harmful. Publishers should be mindful of how they commodify nature to ensure it increases rather than decreases the perceived value of nature.

Posting with purpose

The need to adapt Saxton and Waters' (2014) I–A–C Framework showed there are more nuances between post purposes than being limited to single categories of 'information', 'action' or 'community–building'. Higher interaction levels can also be achieved by combining post purposes. Being mindful of post purpose can highlight the roles publishers play; in particular, 'information' posts correspond to educational roles, 'action' posts promote the 'activism' role of organisations, whilst 'community–building' posts are achieved through the adoption of 'canvassing' and 'entertaining' roles.

In summary, it is not possible to state conclusively that some post variables are more conducive to high engagement than others based on this research alone. Publishers can adopt strategies which increase the potential for users to engage with their content however, creating an environment that fosters discussion, participation and knowledge–sharing. By employing these strategies, publishers can make better use of Facebook as a platform to engage users and achieve their organisational aims.

Acknowledgements

The corresponding author would like to thank Dr Lisa Lau of the School of Geography, Geology and the Environment at Keele University for her advice regarding ethical considerations for this research. The data acquisition script used in the research has been made available as an Open Source repository at: https://github.com/jooldesign/nodejs–fb–scraper

Competing Interests

SJB declares that they have no competing interests.

AGC declares that they have no competing interests.

CB declares that they have no competing interests.

References

Adee, S. (2016) 'How can Facebook and its users burst the 'filter bubble'?' New Scientist, [Online] Available at: https://www.newscientist.com/article/2113246–how–can–facebook–and–its–users–burst–the–filter–bubble/

Alcock, F. (2008) 'Conflicts and Coalitions Within and Across the ENGO Community', Global Environmental Politics, 8 (4), pp 66–91

Andersson, E., Öhman, J. (2016) 'Young people's conversations about environmental and sustainability issues in social media', Environmental Education Research, 0 (0), pp 1–21

Blewitt, J. (2011) 'Critical practice and the public pedagogy of environmental and conservation media', Environmental Education Research, 17 (6), pp 719–734

Boland, B. (2014) 'Organic Reach on Facebook: Your Questions Answered', Facebook Business, [Online] Available at: https://www.facebook.com/business/news/Organic–Reach–on–Facebook–APAC

Bortree, D.S., Seltzer, T. (2009), 'Dialogic strategies and outcomes: An analysis of environmental advocacy groups' Facebook profiles', Public Relations Review, 35, pp 317–319

Branigan, E., Mitsis, A., (2014) 'Reach for Generation Y: using celebrity endorsement to communicate about nonprofit causes with young people in Australia', International Journal of Nonprofit and Voluntary Sector Marketing, 19, pp 314–321

Brinkmann, S. (2012) 'Qualitative Enquiry in Everyday Life', London: Sage

Brulle, R.J., Carmichael, J., Jenkins, J.C. (2012) 'Shifting public opinion on climate change: an empirical assessment of factors influencing concern over climate change in the U.S., 2002–2010', Climatic Change, 114 (2), pp 169–188

Bryman, A. (2012) 'Social Research Methods', 4th Edition, Oxford: Oxford University Press

Caers, R., De Feyter, T., De Couck, M., Stough, T., Vigna, C., Du Bois, C., (2013) 'Facebook: A literature review', New Media & Society, 15 (6), pp 982–1002

Campbell, D.A., Lambright, K.T., Wells, C.J. (2014) 'Looking for Friends, Fans, and Followers? Social Media Use in Public and Nonprofit Human Services', Public Administration Review, 74 (5), pp 655–663

Carolan, M. (2010) 'Sociological ambivalence and climate change', Local Environment, 15 (4), pp 309–321

Carvalho, A. (2009) 'Communication for sustainable policy: connecting science, society and government', Science for Environment Policy: Environmental Communication, European Commission

Castree, N (2003) 'Commodifying what nature?', Progressive Human Geography 27 (3), pp 273–297

Chew, J.C. (no date) 'Making Sense of the Two–Sample T–Test', Six Sigma, [Online] Available at: https://www.isixsigma.com/tools–templates/hypothesis–testing/making–sense–two–sample–t–test/

Cvijikj, I.P., Michahelles, F. (2013) 'Online engagement factors on Facebook brand pages', Social Network Analysis and Mining, 3 (4), pp 843–861

De Vries, L., Gensler, S., Leeflang, P.S.H., (2012) 'Popularity of Brand Posts on Brand Fan Pages: An Investigation of the Effects of Social Media Marketing', Journal of Interactive Marketing, 26 (2), pp 83–91

Ehrlich, P.R. (2011) 'A personal view: environmental education—its content and delivery', Journal of Environmental Studies and Sciences, 1 (1), pp 6–13

Emmison, M. (2016) 'Visual Inquiry: Issues and Developments', In: Silverman, D., ed, Qualitative Research, London: Sage, pp 297–310

Finger, M., Princen, T., (2013) 'Environmental NGOs in World Politics: Linking the Local and the Global', Oxford: Routledge

Gibson, R.B. (2006) 'Beyond the Pillars: Sustainability Assessment as a Framework for Effective Integration of Social, Economic and Ecological Considerations in Significant Decision–Making', Journal of Environmental Assessment Policy and Management, 8 (3), pp 259–280

Gifford, R. (2011) 'The Dragons of Inaction: Psychological Barriers That Limit Climate Change Mitigation and Adaptation', American Psychologist, 66 (4), pp 290–302

Greenberg, J., MacAulay, M. (2009), 'NPO 2.0? Exploring the Web Presence of Environmental Nonprofit Organisations in Canada', Global Media Journal – Canadian Edition, 2 (1), pp 63–88

Gruzd, A., Haythornthwaite, K. (2008) 'The Analysis of Online Communities using Interactive Content–based Social Networks', Proceedings of the American Society for Information Science and Technology, [Extended Abstract]

Halupka, M. (2014) 'Clicktivism: A Systematic Heuristic', Policy & Internet, 6 (2), pp 115–132

Hansen, A. (1991) 'The media and the social construction of the environment', Media, Culture and Society, 13, pp 443–458

Happer, C., Philo, G., (2016) 'New approaches to understanding the role of the news media in the formation of public attitudes and behaviours on climate change', European Journal of Communication, 31 (2), pp 136–151

Hendry, J.R. (2006) 'Taking Aim at Business What Factors Lead Environmental Non–Governmental Organizations to Target Particular Firms?', Business and Society, 45 (1), pp 47–86

Hendry, J.R. (2005) 'Stakeholder Influence Strategies: An Empirical Exploration', Journal of Business Ethics, 61, pp 79–99

Horgan, B. (2010) 'The presentation of self in the age of social media: Distinguishing performances and exhibitions online', Bulletin of Science, Technology & Society, 30, pp 377–386

Illouz, E. (2007) 'Cold Intimacies: The Making of Emotional Capitalism', Cambridge: Polity Press

Joinson, A.N. (2008) 'Looking at, looking up or keeping up with people? motives and use of Facebook', Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, pp 1027–1036

Kaplan, A.M., Haenlein, M. (2010) 'Users of the world, unite! The challenges and opportunities of Social Media', Business Horizons, 53 (1), pp 59–68

Karpf, D. (2010) 'Online Political Mobilization from the Advocacy Group's Perspective: Looking Beyond Clicktivism', Policy & Internet, 2 (4), pp 7–41

Katz–Kimchi, M., Manosevitch, I. (2015) 'Mobilizing Facebook Users against Facebook's Energy Policy: The Case of Greenpeace Unfriend Coal Campaign', Environmental Communication, 9 (2), pp 248–267

Kemp (2017) 'Digital in 2017 – Global Overview', We Are Social, [Online] Available at: https://wearesocial.com/uk/blog/2017/01/digital–in–2017–global–overview

Kucharski, A. (2016) 'Post–truth: Study epidemiology of fake news', Nature, 540, [Online] Available at: https://www.nature.com/nature/journal/v540/n7634/full/540525a.html

Laerd Statistics (2013) 'One–way ANOVA', [Online] Available at: https://statistics.laerd.com/statistical–guides/one–way–anova–statistical–guide.php

Lazer, D., Baum, M., Grinberg, N., Friedland, L., Joseph, K., Hobbs, W., Mattsson, C., (2017) 'Combating Fake News: An Agenda for Research and Action', [Online] Available at: https://shorensteincenter.org/wp–content/uploads/2017/05/Combating–Fake–News–Agenda–for–Research–1.pdf

Liverman, D. (2004) 'Who Governs, at What Scale and at What Price? Geography, Environmental Governance, and the Commodification of Nature', Annals of the Association of American Geographers, 94 (4), pp 734–738

Lomborg, S., Bechmann, A. (2014) 'Using APIs for Data Collection on Social Media', The Information Society, 30, pp 256–265

Malhotra, A., Kubowicz Malhotra, C., See, A. (2013) 'How to Create Brand Engagement on Facebook', MIT Sloan Management Review, [Online] Available at: http://sloanreview.mit.edu/article/how–to–create–brand–engagement–on–facebook/

Mangold, W.G., Faulds, D.J. (2009) 'Social media: The new hybrid element of the promotion mix', Business Horizons, 52 (4), pp 357–365

Markham, A., Stavrova, S., (2016) 'Internet/Digital Research', In: Silverman, D., ed, Qualitative Research, London: Sage, pp 229–243

Martens, K. (2002) 'Mission Impossible? Defining Nongovernmental Organizations', Voluntas: International Journal of Voluntary and Nonprofit Organizations, 13 (3), pp 271–285

McCauley, D. (2006) 'Selling out on nature', Nature, 443 (7107), pp 27–28

McCright A.M, Dunlap R.E. (2011) 'The politicization of climate change and polarization in the American public's views of global warming 2001–2010', The Sociological Quarterly, 52 (2), pp 155–194

Meredith, J. (2016) 'Using Conversation Analysis and Discursive Psychology to Analyse Online Data', In: Silverman, D., ed, Qualitative Research, pp 261–276, London: Sage

Morgan, D.L. (2014) 'Pragmatism as a Paradigm for Social Research', Qualitative Inquiry, 20 (8), pp 1045–1053

Nah, S., Saxton, G.D. (2012) 'Modelling the adoption and use of social media by nonprofit organizations', New Media & Society, 0 (0), pp 1–20

Node.js Foundation (2017) 'About Node.js', [Online] Available at: https://nodejs.org/en/about/

Nulman, E., Özkula, S.M. (2016) 'Environmental nongovernmental organizations' digital media practices toward environmental sustainability and implications for informational governance', Current Opinion in Environmental Sustainability, 18, pp 10–16

Pang, B., Lee, L., (2008), 'Opinion Mining and Sentiment Analysis', Foundations and Trends in Information Retrieval, 2 (1–2), pp 1–135

Patel, N. (2016) 'Facebook's Algorithm Revealed: How to Remain Visible in The Cluttered News Feed?', [Online] Available at: http://neilpatel.com/blog/facebooks–algorithm–revealed–how–to–remain–visible–in–the–cluttered–facebook–news–feed/

Patel, S. (2015) 'The research paradigm – methodology, epistemology and ontology – explained in simple language', [Online] Available at: http://salmapatel.co.uk/academia/the–research–paradigm–methodology–epistemology–and–ontology–explained–in–simple–language

Saldana, J. (2009) 'The Coding Manual for Qualitative Researchers', London: Sage

Saxton, G.D., (2014) 'How Organizations Use Social Media: Engaging the Public', Social Metrics, [Online] Available at: http://social–metrics.org/how–organizations–engage–with–social–media/

Saxton, G.D., Waters, R.D., (2014) 'What do Stakeholders Like on Facebook? Examining Public Reactions to Nonprofit Organizations' Informational, Promotional, and Community–Building Messages', Journal of Public Relations Research, 26, pp 280–299

Stein, L. (2009) 'Social movement web use in theory and practice: A content analysis of US movement websites', New Media & Society, 11 (5), pp 749–771

Stoddart, M.C.J., MacDonald, L., (2011) '"Keep it Wild, Keep it Local": Comparing News Media and the Internet as Sites for Environmental Movement Activism for Jumbo Pass, British Columbia', Canadian Journal of Sociology, 36 (4), pp 313–335

Twitter (2017) 'Using hashtags on Twitter', Help Centre, [Online] Available at: https://support.twitter.com/articles/49309

Walker, J., Starosielski, N. (2016) 'Introduction: Sustainable Media', In: Starosielski, N., Walker, J., (eds), Sustainable Media: Critical Approaches to Media and Environment, New York: Routledge

Waters, R.D., Burnett, E., Lamm, A., Lucas, J. (2009) 'Engaging stakeholders through social networking: How nonprofit organizations are using Facebook', Public Relations Review, 35, pp 102–106

White, M. (2010) 'Clicktivism is ruining leftist activism', The Guardian, [Online] Available at: https://www.theguardian.com/commentisfree/2010/aug/12/clicktivism–ruining–leftist–activism

Zito, A.R., Jacobs, J.E., (2009) 'NGOs, the European Union and the case of the environment', In: Joachim, J., Locher, B., eds, 'Transnational Activism in the UN and the EU: A Comparative Study', pp 101–115, Oxford: Routledge

Zuckerberg, M. (2016) 'Today we're launching Facebook Live for everyone', Facebook, [Online] Available at: https://www.facebook.com/zuck/posts/10102764095821611