SECTION 1 INTRODUCTION
1.0 Introduction
Very little is known about the scale of sea angling and its geographical spread across Scotland, or the economic impact of sea angler expenditure on coastal communities. Because the impact is largely hidden, it is possible that sea angling could be over-looked when fisheries, tourism and coastal developments are being considered. The Scottish Government is seeking to rectify this by funding this study which will be the first assessment of the contribution made by sea angling to employment and income in Scotland as a whole and regions within Scotland 12.
1.1 Objectives
The aims of the study, as stated by the Scottish Government, were as follows:
a) What is the economic impact of recreational sea angling in Scotland? Economic impact should be measured in terms of income and employment effects arising from the expenditure by those undertaking recreational sea angling ( RSA). Where possible, estimates should include not just direct effects but indirect and induced multiplier ones as well. It is essential that displacement effects should be explicitly estimated and taken into account so that both gross and net impacts are estimated.
b) Where are the important local centres for sea angling in Scotland? Measures of relative importance could be presented in terms of the level of participation, level of expenditure and the volume of catches by RSAs.
c) What are the main competing areas both within and outwith Scotland? Centres should be described and compared also in terms of the type of sea angling operations and the species caught so that it is clear why some areas are more important than others. This may require an assessment of other factors such as the availability of slipways, charter boats, other attractions and facilities. It should be noted that there may well be important variations between different areas in terms of the species that are targeted. The west coast tends to specialise in sharks, skate, spurdog, dogfish and tope whilst the species on the east coast are predominantly cod and ling.
d) What are the principal characteristics of the recreational sea angling sector? This should be considered in terms of both the numbers involved, their age, sex and socio-economic status, the frequency of their participation; the level, type and location of their expenditure; whether they fish from the shore or boats (chartered or their own); their area of residence and length of stay. Area of residence will reveal how many are stay or day visitors to an area.
e) What are the key trends in the sector? What factors seem to explain the past absolute and relative performance of the sector in different areas?
f) What are the future prospects for the sector? Are there areas, and if so where, that appear to offer the best prospects for future expansion and what are the principal opportunities and constraints likely to affect future development?
1.2 Background Considerations
With the above objectives and the paucity of secondary data, extensive primary data needed to be collected from sea anglers and other stakeholders. There are however particular problems associated with the collection of primary data on sea angling in Scotland. These are introduced below.
1.2.0 Estimating Scottish population totals
In Scotland, the angling population totals are unknown because a) there is no licensing system and b) no reliable household omnibus survey has yet been undertaken to determine Scottish participation rates in angling; either generally or in relation to particular types of angling, such as sea angling. Moreover, there are no lists of sea anglers' contact details. Whilst a similar situation exists in England and Wales with respect to sea angling, there has been some general household survey work in England and Wales directed at determining participation levels in sea angling (eg NOP, 1971; Moon and Souter 1995; Dunn and Potten, 1994). Yet, even then, rather than utilising prior estimates of population participation rates, a more recent assessment of the economic value of sea angling in England and Wales (Drew Associates 2004) believed it necessary to participate in a household omnibus survey.
Since Scottish sea angler contact details were not available, sea anglers could not be randomly selected in order to distribute telephone or postal questionnaires. It is of course feasible for researchers actively to seek sea anglers, perhaps by visiting sea angling marks, or by contacting clubs, or by leaving self completion questionnaires at suitable locations. A similar strategy is to issue a general invitation for anglers to supply their contact details, or to complete on-line questionnaires.
The problem is that, to a greater of lesser degree, all these approaches are inherently biased towards the more enthusiastic angler. For example, with on-site interviews, anglers who fish more regularly are more likely to be contacted. A similar problem exists with distributing questionnaires through angling clubs or inviting sea anglers to provide contact details or to complete on-line questionnaires. Exclusive reliance on these instruments would generate estimates of, for instance, mean sea angler days which would be hopelessly biased (unless observations were weighted according to the probability of the angler being sampled). It was therefore not appropriate for this evaluation to rely exclusively on either contacting sea anglers on-site or inviting them to complete questionnaires.
The twin problems of an unknown population and the potential bias in survey instruments have been tackled in a number of ways. For example, studies of freshwater angling in Scotland (Mackay Consultants, 1989; Fisheries Research Management, 2000; Deloitte and Touche, 1996; Radford et al, 2004; Riddington et al, 2004) have tackled the issue of an unknown population by aggregating estimates of the number of angler days provided by individual fishery owners. Unfortunately, since sea angling is open access, this option was not available.
Another theoretical possibility would be to try and count the population of anglers around the coast. It would of course be prohibitively expensive to constantly monitor every sea angling site, let alone the broader expanse of the open sea. Inevitably, sampling would be necessary, with the sample stratified by types of sites, times of the day, days of the week etc. On site sampling would also raise the problem that the keener anglers are more likely to be sampled on-site. It was concluded that on-site enumeration could not produce reliable population totals.
The most robust and cost effective approach to the identification of the population of Scottish sea anglers is to undertake a general telephone omnibus survey of individuals in Scotland, stratified by regions to establish regional participation levels in sea angling.
Given the objectives of the study, unbiased estimates were also required of other key parameters relating to the total population of sea anglers. These include the distribution of total angler effort (measured in sea angler days) across (i) shore angling, (ii) own or friend's boat and (iii) charter boats. As explained above, the available angler survey instruments are biased and, by themselves, cannot be relied upon to generate reliable population totals for (i), (ii) or (iii). Consequently, the omnibus survey also has to provide unbiased estimates of how often sea anglers fish, and the relative importance of shore and boat angling trips.
1.2.1 Generating data on average angler expenditure and other variables
A requirement of the study is to estimate the economic impact of angler expenditure, and to estimate expenditure impacts specific to regions within Scotland. A large number of observations are required to provide sufficient observations on expenditure per angler day covering all the regions and expenditure categories. It was not feasible to use the omnibus survey for this purpose. This is because, assuming 2% of the population engage in sea angling, 10,000 telephone contacts would only yield 200 observations and we estimate that 450-500 would be required. The omnibus survey would therefore have to be expanded to 25,000 telephone contacts and many more questions would have to be asked of each sea angler. This option was prohibitively expensive and the complex questioning would have placed a considerable burden on the telephone questionnaire itself.
It is therefore necessary to deploy, where possible, more cost effective instruments to generate information on the activity and the views and expenditures of a sufficient number of individual anglers. A recent approach has been the use of on-line questionnaires. This was used successfully in a Scottish angling context by Radford (2004) and by Riddington (2004) and in England and Wales by Radford et al (2007).
Provided that there is sufficient advertising, promotion and support from within the sea angling community, it has been our experience that an on-line survey is the most cost effective way to generate observations on large numbers of anglers.
As discussed above, by inviting anglers to complete an on-line questionnaire, there is a potentially serious problem of self-selection bias. There are two points to be made about this. First, the inherent bias in the angler survey instruments is not too serious provided the extent of the bias is known. If the omnibus survey (and the United Kingdom Tourist Study) is unbiased, and it should be, then it can be used to correct for the inherent bias in the angler survey instruments. Indeed, this becomes a key role of the omnibus survey.
Second, there are some key variables that will be largely unaffected by self selection bias. For example, an important variable is the average angler expenditure per day. There is no reason to suppose that the keener angler necessarily spends more per day. Similarly, whilst Radford et al (2007) found bias in the amount of visiting angling activity, it could not find any bias in the distribution of that activity. To clarify, whilst the "biased" internet respondent from Glasgow might spend 10 days, on average, fishing outside his home area, compared to 2 for the unbiased "omnibus" respondent, the likelihood of these being spent in Skye would not be significantly different. Thus, provided unbiased estimates of the total sea angling days of visiting anglers from each region (including outside Scotland) are available, we were able to estimate the number of sea angler days in each region.
A further problem arises. A Scottish omnibus telephone survey will not pick up visitors from outside Scotland. The issue of non-Scottish sea anglers is discussed below.
1.2.2 Assessment of non Scottish sea anglers
As outlined above, it would have been prohibitively expensive to undertake an on-site count to identify the number of non-Scottish visitors participating in sea angling. Although far fewer are expected, the importance of non-Scottish visitors is magnified in that they are potentially the least likely to retain their spending in Scotland and hence have a larger than normal economic impact. Fortunately the UK Tourist Study ( UKTS) is a very large survey that covers all households in the UK and asks questions both about visit destinations but also about activities undertaken during these visits.
A major problem is that angling in the UKTS covers both sea and freshwater angling. The number of freshwater anglers identified cannot be simply subtracted from the total number of anglers found by the UKTS because an unknown, and possibly significant number, undertake both sea angling and freshwater angling. TNS(2007) surveyed some 1427 anglers and identified the proportions that went game, coarse and sea angling and combinations thereof. Although the sample was not random nor exclusively from the RUK (Rest of the UK), the proportions were, at least in part, verified by the internet study and consequently allowed an estimate of sea anglers.
1.2.3 Assessment of Young Anglers
For both practical and legal reasons, the population of young anglers under 18 cannot be obtained via a telephone survey. For a number of reasons discussed later, this group is not thought to be critical in determining economic impact. However they are potentially important if trying to identify policy because of the high economic value that might be associated with the high number of participants in this category. In this case participation estimates have been extracted from the archived data of the very large UK Leisure Survey and the ratio of young to adult rates calculated. That ratio has then been applied to the participation rates found in the omnibus survey.