Who trusts who and why? Risk communication during the Montserrat volcanic eruption.




K Haynes, Volcanic island in crisis: investigating environmental uncertainty and the complexities it brings. Australian Journal of Emergency Management (2006) 21, 21-28.


K Haynes, J Barclay and N Pidgeon, The issue of trust and its influence on risk communication during a volcanic crisis. Bulletin of Volcanology (2008) 70, 605-621.


K Haynes, J Barclay and N Pidgeon, Whose reality counts? Factors affecting the perception of volcanic risk. Journal of Volcanology and Geothermal Research (2008) 172, 259-272.



The cast of humans around a “typical” volcanic eruption has three main groups: The residents; their government; the scientists who observe the volcano and try to make sense of its activity. Obviously other players, such as the media (local and global), also form distinct groups of varying prominence. Each group has its own agenda. The residents try to get on with their lives but have to modify these to work around the chaos caused by the volcano. The government also tries to get on with its pre-eruption plans for the community but finds them progressively disrupted by the volcano. The scientists have three distinct tasks: (1) Collect factual data about the volcano and its eruption; (2) Interpret these data so as to forecast future hazards and the risks they may pose for residents; (3) Communicate their findings to either the government or, in some cases, direct to the public via meetings and/or the media.


Montserrat adds two complexities to the general pattern. Firstly, the scientists in the Montserrat Volcano Observatory (MVO) officially confine their contribution to data collection. They then pass these to an international Scientific Advisory Committee (SAC) who make periodic formal assessments of possible future events and consequent risks. Secondly, the government system for the island is about as complex as it is possible to design, thanks to its status as a British Overseas Territory. This is illustrated (Fig. 1) by a diagram in a research article by Haynes (2006) in the Australian Journal of Emergency Management. The original sources of this diagram are buried in volumes that are either inaccessible or expensive, whereas this summary is available on the internet at the click of a mouse.


Fig. 1  The diagram in Haynes (2006) summarising the web of government management and communication on Montserrat.


Forests have been felled to provide the paper used to document exhaustively the ins and outs of every aspect of the early years of the eruption, culminating in 19 fatalities in 1997 and the loss of two thirds of Montserrat beneath new volcanic deposits. Tensions ran high at the time, as is clearly shown by two memorable quotes: on the temporary evacuation shelter accommodation, “not even fit for cattle” (C M Brandt); on Montserrat’s approaches to the UK Government for more aid, “They say £10,000, double, treble, and then think of another number. It will be golden elephants next” (Clare Short).


I shall not rake further over these old coals here. Instead I shall concentrate on one specific issue: trust. Apart from those who always lived in the north of the island, most Montserrat residents have had to endure horrendous life changes brought on by the eruption, including permanent evacuation from the south and multiple temporary evacuations from Salem and surrounding districts. These evacuations have occurred as generally consensual events, rather than forceful evictions, because residents have mostly believed what they were told about the likely future behaviours of the volcano and trusted those who brought the bad news. Nevertheless, there were dissenters and – as always happens when the media are searching for “stories” – the views of such groups were well publicised worldwide during the early years of the eruption.




The hierarchy of trust on Montserrat



Against this background Hayes, Barclay and Pidgeon investigated the key questions of “Who trusts who and why?” on Montserrat during a period (2003-4) when the volcano was very troublesome to residents. In their report in the Bulletin of Volcanology in 2008 they tried to escape the trap of subjectivity that some previous studies had fallen into by using much the same techniques as those developed by Gallup and similar polling organisations to select a fully representative panel of the public to be interviewed and answer questionnaires. In addition they interviewed a similar number of assorted scientists and a smaller number of government officials of various types. They also sat in as observers on a series of meetings of the full-time MVO staff, the SAC and other groups – notably the joint Emergency Policy Group of scientists, government and officials.


Finally they used the preliminary findings from the interviews to construct carefully targeted questionnaires and then used the answers to these to generate a mass of data that could be explored by statistical analysis. This table shows one of the three questionnaires used to evaluate trust between groups; a respondent could agree/disagree with a statement on a 1-5 scale.





The qualitative results showed very high trust of scientists by the authorities. Of course there were dissenters but the risk of becoming legally liable for the consequences of expressing an alternative viewpoint tended to discourage open disagreement. A fascinating change to this eruption-long relationship took place in March 2007, when the authorities decided to not to follow an MVO/SAC report that implied a recommendation to evacuate Salem yet again. Both groups felt that rapid staff turnover in the other was a problem, with time “wasted” bringing counterparts “up to speed” on matters they needed to know.


The public overwhelmingly trusted the scientists, often because they had learned to do so the hard way. They mostly had less trust in the authorities. In addition a small vociferous group of residents had no trust at all in either scientists or government.


Within the broad term “trust” the qualitative survey looked at subdivisions (“dimensions”) including: “competence; integrity; value similarity; openness; conflicting messages of safety and danger”. Several interesting points emerged under these headings:


1. Although the scientists were generally trusted to produce facts impartially, some of the public worried that government (island, UK or both) fiddled with the interpretation of the data to suit their political purposes.


2. There was a clear view that the future of the MVO would be most effective if it were both Caribbean-managed and Caribbean-staffed. This change was made in April 2008.


3. Residents were inclined to trust individual scientists to varying extents, depending on the sociability and “amiability” of each one. This point is widely recognised in other fields – notable politics – and has important implications for organising and communicating the science. Actually it is a pity that people confuse the characters of scientists with the quality of their science because there is no logical reason why successful scientists should necessarily be sociable. From all accounts, Isaac Newton was notoriously antisocial but nobody questions his scientific achievements.


4. Some residents living around the lower Belham Valley have become determined and well organised in sending out the message that the eruption will never destroy this area. Others are particularly likely to heed their messages when the volcano is a dangerous state but combining this with little or no audio-visual huffing and puffing, as during most of 2007 and 2008 (up to 29 July).


In order to obtain quantitative results, questionnaire respondents were asked who they trusted most/least to tell the truth about the volcano, from the following groups:


Friends and family


ZJB radio (the island radio)

Emergency Operations Centre

The Montserrat Reporter (the island newspaper)

Salem Volcanic Crisis Committee

Opposition politicians on Montserrat

British Governor’s Office

Montserrat Government

World Press


The detailed numerical results are summarised in Fig. 2.



Fig. 2  Who do Montserrat residents trust for volcano/eruption information?



Residents primarily trust other residents and the scientists, followed by local emergency management teams and media. Both the Governor’s Office and Montserrat Government are only marginally trusted and the world press is firmly mistrusted. Perhaps this hierarchy of trust is inevitable. The authorities have the miserable job of forcing people out of their houses and jobs when volcanic danger looms – hardly a recipe for good relations with residents and least of all when the evacuations are temporary, so that residents perceive that “nothing happened”. Breaking down trust into such “dimensions” as “competence, reliability and openness” produced the same pecking order: scientists well ahead; Montserrat Government level with or slightly ahead of Governor’s Office.



Fig. 3  Where the greatest trust lies on Montserrat (Bob Thompson); between family and friends in homes and on the street.

Fig. 4  Two flags flying over one small island (Bob Thompson). A recipe for complicated government and communications.


The implication of this study is that the residents strongly prefer that the scientists, and not the government authorities, tell them both about the volcanic activity and what it means for their homes and livelihoods. The problem is that scientists are not specifically trained in this sort of communication, although some may be talented amateurs. Many scientists feel that their job is to report data to the authorities and then to let the latter handle communications with residents. This is the opposite of what the Montserrat residents want. Of course the balance within this communication triangle between scientists, government and residents will vary in different societies.


Fig. 5  MVO scientist collecting data (Copyright NERC). But does he/she automatically have a talent for communicating these data and their implications to both the authorities and residents, and a duty to do so?



Montserrat Volcano Observatory

Click here to reach page 2 of this topic