Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
1 Introduction Recreational fisheries are an important and growing component of many marine fisheries. Data about the numbers and kinds of fish taken through recreational fishing are essential for fisheries management. In recent years, recreational fisheries have been monitored less rigorously than commercial fisheries. But as concerns about overfishing have grown, more attention has been turned toward the possible impact of marine recreational fishing and the proportion of fish taken by each sector (e.g., National Research Council, 1999, 2000; Lucy and Studholme, 2002; Coleman et al., 2004). This, in turn, has led to greater demands for timely, accurate recreational fishing data and scrutiny of the methods used to collect these data. According to the best available estimates, approximately 14 million anglers made almost 82 million fishing trips along the Atlantic, Pacific, and Gulf coasts (excluding Texas) in 2004 (National Oceanic and Atmospheric Administration, 2005a). While each individual angler typically harvests a small number of fish, collectively these sport fisheries can take a significant fraction of the yearly catch. For example, in 1999, recreational fishing accounted for 94 percent of the total catch of spotted seatrout (Cynoscion nebulosus), 76 percent of striped bass (Morone saxatilis) and California sheephead (Semicossyphus pulcher), and 60 percent of king mackerel (Scomberomorus cavalla) (Figure 1.1). Recreational catch continues to grow for several important fisheries; the recreational catch of summer flounder and Pacific halibut increased 40 fold in less than 20 years (Coughenower and Blood, 1997). Recent scientific papers (e.g., Coleman et al., 2004; Cooke and Cowx, 2004) provide data that suggest recreational fisheries are a more significant factor in the exploitation of fish stocks than previously believed. Considering the potential contribution recreational fishing has to total 15
16 REVIEW OF RECREATIONAL FISHERIES SURVEY METHODS Striped Bass Red Drum Bluefish Dolphinfish Spotted Seatrout Summer Flounder Atlantic Croaker Sheephead King Mackerel Yellowfin Tuna 0% 20% 40% 60% 80% 100% Commercial Recreational FIGURE 1.1 Top Ten Recreational Species Versus Commercial Harvest for 2004. Comparisons between the top ten species in descending order of abundance by weight for U.S. recreational fish harvests and commercial landings. The figure does not include data for Alaska and Texas because no NMFS recreational surveys are conducted in those states (National Oceanic and Atmospheric Administration, 2005a). catch, it is clear that accurate monitoring of catch from both recreational and commercial fisheries is needed to ensure that total catch does not exceed the total allowable catch calculated to maintain a sustainable population. Nationally, recreational catch is monitored primarily (but not entirely) through the Marine Recreational Fisheries Statistics Survey (MRFSS), which was implemented by the National Marine Fisheries Service (NMFS) of the National Oceanic and Atmospheric Admini- stration (NOAA) in 1979 to obtain standardized and comparable esti- mates of participation, effort, and catch by anglers in U.S. marine waters (Essig and Holliday, 1991; National Oceanic and Atmospheric Admini- stration, 2005b). In addition, there are several state programs, operated instead of or complementary to the MRFSS, which are discussed in Chapter 3. The stated purpose of the MRFSS is to establish a reliable
INTRODUCTION 17 database for estimating the impact of marine recreational fishing on marine resources. According to the MRFSS, "[t]he program's mission is to provide accurate, precise, and timely fisheries-dependent information for U.S. marine waters through the coordination and administration of recreational fisheries surveys nationwide" (National Oceanic and Atmospheric Administration, 2005b). However, many now argue that recreational data collected through the MRFSS and other recreational fishing surveys are being used for management decisions in ways that exceed their intended design and purpose (Box 1.1). As exploitation levels of the nation's fisheries have increased, managers have begun (of necessity) to manage stocks more actively. This has resulted in the close monitoring of catch; more frequent closings of fishery areas and shortening of seasons; more stringent size limits; and for anglers, more detailed, species-specific bag limits. Compared to only 10 years ago, recreational fisheries data are now incorporated into more fisheries management plans and stock assessments of the regional fishery management councils. Currently, many recreational fisheries are man- aged on annually monitored quotas, and some are even managed using in-season quotas. This type of recreational fisheries management requires data to be collected and analyzed in the span of weeks instead of months--the amount of time it has traditionally taken to collect and process recreational data. Other recreational fisheries require monitoring and management on finer geographical scales. For example, the quota for the striped bass stock on the east coast is divided among states, and each state is required to report its respective catch. This is also true for king mackerel. As these new needs have arisen, some of the challenges have been met with changes to the MRFSS, and in some cases, new surveys have been implemented to provide some of the necessary data (see Appendix C). However, concern remains that the data currently being collected through recreational fishing surveys are not precise, robust, or timely enough for the various scales employed in fisheries management. The mismatch between the data required and the data delivered has prompted substantial commentary on and criticism of the program. In addition, there are occasions when the estimates of total catch vary widely between years, leading to significant changes in the management restrictions implemented, again eliciting questions and criticisms from recreational anglers and commercial fishermen. The data produced by these surveys are subjected to considerable scrutiny, but there is a lack of confidence within some elements of the angler community regarding the accuracy of total removal estimates of
18 REVIEW OF RECREATIONAL FISHERIES SURVEY METHODS Box 1.1 Users of the MRFSS Information The MRFSS information is used by a variety of different groups. The Magnuson Fishery Conservation and Management Act of 1976, amend- ed in 1980, and now the Magnuson-Stevens Fishery Conservation and Management Act (P.L. 94-265) gives much fishery management respon- sibility to eight regional fishery management councils. In addition, there are three marine fisheries commissions--the Atlantic States Marine Fisheries Commission, the Gulf States Marine Fisheries Commission, and the Pacific States Marine Fisheries Commission--which have vari- ous degrees of management and regulatory authority. There are four international commissions that make management recommendations to their governments, including the U.S. government, for implementation: the Pacific Salmon Commission (United States and Canada), the Inter- national Pacific Halibut Commission (United States and Canada), the Inter-American Tropical Tuna Commission (United States and 14 other member countries), and the International Commission on Conservation of Atlantic Tunas (United States and 40 other member countries). (The Great Lakes Fishery Commission [United States and Canada] does not take part in or use the MRFSS.) The commissions manage species that are recreationally important. The coastal states have fishery manage- ment agencies with jurisdiction in state waters (usually to three nautical miles offshore but nine nautical miles for Texas, Florida, and Puerto Rico). Other users include fishing organizations, environmental groups, various industry groups, and individuals. All these groups use data from the MRFSS and other sources to make decisions concerning their various activities. recreational fish catch. This has resulted in reduced angler cooperation with the data-collection surveys (because participation is voluntary) and in disagreements about proposed management actions. On some occa- sions, these disagreements have resulted in lawsuits against NMFS. Some user groups have even generated alternate surveys in attempts to illustrate shortcomings of existing surveys or to provide a more appro- priate design for local circumstances. This report will demonstrate that achieving valid and reliable recreational fishing estimates in the future will require much better cooperation between agencies; the redesign of some existing surveys; the creation of new surveys; and, most important, increased funding to make the above practicable.
INTRODUCTION 19 WHAT DATA ARE COLLECTED FROM ANGLERS Usually, recreational surveys at large spatial scales measure angler effort (E) using one survey (e.g., telephone survey) and measure catch per unit effort (CPUE) using a separate survey (e.g., angler interview). If E and CPUE are known, one can calculate total catch (C) as: C = E ×CPUE There are three different categories of catch: catch available for inspection during the onsite interview (A); catch unavailable for inspect- tion because it is filleted, discarded dead, or refused for inspection (B1); and catch released alive (B2). For each angler trip, total catch is defined as A + B1 + B2, and landed catch, or harvest (H), is calculated as A + B1. For management purposes such as quota monitoring, the total number of fish taken by anglers is recorded as harvest; catch released alive is not factored in. For stock assessment purposes, an additional fraction of the catch released alive that subsequently dies is estimated using a "hooking" mortality (MH). This enables the estimation of total recreational fishing removals (R): R = A + B1 + (B2 × MH) Catch, harvest, and total removals are estimated by species for specific spatial regions and temporal periods, with the spatial and temporal scale requirements depending on the management needs for that species. Further, catch is assigned to age or size classes based on biological sampling of the catch. HOW CATCH IS ESTIMATED FROM RECREATIONAL FISHERIES SURVEY DATA The data produced by surveys of anglers are used to provide infor- mation for stock assessments and to support management decisions. An accurate estimate of total removals is the most critical piece of infor- mation needed for marine fisheries management (Gulland, 1983; Pauly, 1998; Watson et al., 2000), but other estimates also are required, includ- ing effort expended; species targeted; geographical allocation of land-
20 REVIEW OF RECREATIONAL FISHERIES SURVEY METHODS ings; and the species, age, and size composition of the catch. In addition to quota monitoring and stock assessments, recreational data are also important for social and economic studies that help to determine current socioeconomic impacts of fishing participation and help to evaluate the effectiveness of current and potential fishery management actions. However, economic and social data generally are not collected in the same consistent, annual approach as catch and CPUE data. WHY RECREATIONAL FISHING DATA ARE DIFFICULT TO COLLECT While there are problems with commercial catch statistics, such as lack of mandatory reporting requirements for some fisheries and incorrect reporting, in general, it is more straightforward and less complex to collect catch statistics for the commercial sector than the recreational one. Commercial fisheries catches are usually landed at specific points where dealers are located. At these places, catch is weighed and recorded, providing an opportunity to monitor this catch through dockside landing receipts. Since both the dealer and the fisherman record and submit the data collected, there is a useful redundancy in the data that facilitates error checking and quality control. More specific effort and discard information is often documented through mandatory logbooks and observer programs. Furthermore, the universe of known fishermen in the commercial sector is defined through required commercial licenses. All of these facts align to provide rela- tively reliable and timely estimates from the commercial sector. Conversely, the participation base for the recreational fishery is considerably more numerous, diverse, and diffuse than that for com- mercial fisheries. In addition, anglers access fisheries in numerous modes1 so there are not limited access points analogous to processing facilities at which to intercept anglers. The nature of the participants is also different. Many anglers are infrequent participants in the fishery and may be less amenable to survey coverage. The goal of monitoring all different modes, locations, and species involved (e.g., charter boats versus private boats, offshore versus near- shore species, fisheries with temporally or spatially restricted fishing 1 The type of place or platform from which marine recreational fishing occurs, including fishing from boats sailing from harbors, marinas, and private docks and from shore on piers, jetties, and beaches.
INTRODUCTION 21 seasons) is a lofty one. Indeed, recreational surveys may be the most complex national surveys currently conducted. With so many complex- ities, accurate estimation of total catch is very expensive. Current meth- ods circumvent some of this expense by making unverified--and, in some cases, unverifiable--assumptions about angler behavior. The extent to which these assumptions adversely affect the estimates is not currently known and will need to be evaluated. The When and Where of Data Collection The current management framework for marine recreational fisheries is highly diverse, depending on species and region. The temporal spec- trum of management actions runs from in-season to annual to multi-year, all of which require different data collection and processing strategies. In-season management obviously requires the most timely data provision to avoid quota overruns. In fact, this approach has been implemented previously for some species and then abandoned due to the lack of estimates of sufficient precision at the appropriate temporal and spatial scale (e.g., red snapper in the Gulf of Mexico). The spatial scale for management extends from local to regional2 and national to, occasionally, international. In addition, species may be managed individually or jointly with co-occurring species. The adequacy of a particular survey design for supporting a particular temporal scale of management (in-season, annual, or multi-year) may require a similarly resolved spatial scale. As finer management scales are created, sampling strata that were previously combined to create a mean estimate on a larger scale must now stand alone. This means a larger sample size from each stratum is needed to achieve reasonable precision of the estimates within each stratum. This requires evaluation and revision on a continu- ing basis. 2 The committee uses the term "regional" to represent scales that are smaller than national but larger than state, and consequently, there will be some association of this scale with the jurisdictions of regional councils and multi- state commissions. In some instances, however, the context could be interpreted more broadly.
22 REVIEW OF RECREATIONAL FISHERIES SURVEY METHODS The Who of Data Collection Recreational fisheries surveys involve the interaction of two large and diverse groups of people: the fishing community and the survey organizations. The recreational fishing community can be usefully divided into two populations based on the mode of fishing. The first includes anglers who fish from shore, docks, and private (i.e., not for- hire) boats. The second, the for-hire sector, is comprised of anglers who fish from head boats, charter boats, and guided boats.3 The issues that impact sample design are different for the two populations. They differ in total numbers of participants, the ability to access the anglers for interview, the frequency of fishing activity, and the skill of each type of angler. Typically, these two populations are sampled separately with the for-hire sector often sampled using a boat directory telephone survey. Organizations tasked with gathering the data for these angler surveys vary across regions and states; this can contribute to variations in the quality and quantity of the data collected. NMFS conducts the MRFSS through contracts to private companies that complete the telephone and intercept interviews. Some state agencies have taken on the responsibility of conducting intercept interviews, using state personnel to collect the catch data for each angler, while contractors continue to conduct the telephone interviews. Still other states carry out all aspects of survey implementation (e.g., Texas, California, Oregon, Washington); in Cali- fornia, Oregon, and Washington, the survey activities receive partial funding from NMFS. The degree to which the regional fisheries commissions are involved in survey implementation and data collection varies as well. The Gulf States Marine Fisheries Commission serves as the contractor for the intercept portion of the MRFSS in that area; where- as the Atlantic States Marine Fisheries Commission is involved in survey design and data handling for its area but is not involved in the actual conduct of the data collection. The Pacific States Marine Fisheries Commission (PSMFC) coordinates sampling in California, Oregon, and 3Head boats, also called party boats, take large groups of anglers (sometimes as many as 100) on fishing trips; the groups usually are not pre-formed. Charter boats (also occasionally called party boats) take smaller groups of anglers, usually four to eight, most often in pre-formed groups. Guided trips are trips in which a guide takes one or two anglers in a smaller boat. These different categories operate under different U.S. Coast Guard and state license require- ments. Throughout this report, these sectors are collectively referred to as the for-hire sector.
INTRODUCTION 23 Washington between the states and NMFS, with a NMFS contract and state funds. The sampling programs are a mix of state and PSMFC samplers and supervisors. PSMFC maintains the catch and effort database for the three states with the online Recreational Fisheries Information Network (RecFIN). RecFIN is used by the states and the Pacific Fishery Management Council to monitor and manage the fish- eries under their purview. The RecFIN database is also available for access by recreational fisheries constituents and the public. Clearly, agency cooperation and coordination is crucial to producing consistent and useful data and estimators. The Cost of Data Collection The amount of data collected is limited most often by cost. Each telephone call made and each intercept survey conducted has a specific price tag associated with it depending on labor costs of the surveyor and the amount of training that person needs to conduct the survey properly. In 2004, the annual operating budget for the MRFSS was approximately $7.2 million. In addition, many states have chosen to supplement existing surveys with additional funds, increasing the overall sample size to achieve specific state goals. For example, North Carolina dedicates approximately $260,000 per year in additional funds to supplement the MRFSS sampling. Most often, additional money is expended because of the need for more accurate or precise estimates to better manage the resources. The cost of the data collection should always be weighed against the benefit of having more information. This committee cannot judge how much federal money should be spent on any endeavor, but one important aspect of the research it recommends in Chapter 6 would be to identify the financial costs and benefits in terms of various improvements in the survey. SCIENTIFIC ISSUES AFFECTING SURVEY DESIGN For all surveys, there are concerns that must be addressed adequately to foster confidence in the generated estimates. Primary among these are the statistical concerns of precision and bias. (See Box 1.2 for a few relevant statistical definitions; see Chapter 3 for more detailed discussion of these statistical issues.) Inadequate precision can be addressed in a straightforward manner by increasing sample size or by otherwise
24 REVIEW OF RECREATIONAL FISHERIES SURVEY METHODS increasing the efficiency of the sample design. Bias is more difficult to identify and reduce. It typically cannot be measured from the sample itself and cannot be reduced without changes in the way the sample units are selected or the way measurements are taken from those units. In an ideal survey, the target population units are identified in a frame, a sample is selected from the frame, and the selected units respond with accurate information. Unfortunately, errors can arise at each step in this process. Groves et al. (2004) classify errors as either errors of representation or errors of measurement. Errors of representation are those that arise due to problems that prevent the sample from representing the population accurately. These errors can lead to bias in the estimates if the excluded population units differ from the included ones. In fisheries surveys, these errors include the following: · Coverage error occurs when the sampling frame does not match the target population perfectly, due to duplications or under- coverage. Duplications occur when the frame lists a target unit more than once, such as a fishing household with more than one telephone line. Undercoverage occurs when the sampling frame does not include all the units belonging to the target population. This occurs when not all sampling sites are included in the frame or only coastal households are included in the telephone sampl- ing frame. · Nonresponse error occurs when some sampled units do not provide data, either because they are not located (e.g., not at home in telephone survey), or they refuse to participate (e.g., will not allow counting or measurement of fish in intercept survey). Errors of measurement in fisheries surveys include the following: · Respondent error occurs when the respondent cannot or will not supply accurate information. Examples of this in fisheries survey include prestige bias (i.e., the tendency for respondents to answer in a way that makes them feel or look better) and recall bias (i.e., the inability to accurately remember previous events). · Interviewer error occurs when interviewers introduce error into the data they collect. For example, such an error would occur if
INTRODUCTION 25 Box 1.2 Essential Statistical Properties Bias: The deviation of the expected value (mean) of a statistical esti- mator from the quantity it estimates. Bias can arise if some units in the population have no chance of entering the sample (e.g., anglers fishing from private land who are not encountered in the intercept survey.) Bias also can arise from inaccurate measurements made on those who do enter the sample, such as incorrect recall by respondents. Bias usually cannot be assessed from the sample itself. If it can be assessed, it usually requires a special study outside the sample. Variance: The expected value of the squared difference between a statistical estimator and its mean, reflecting the estimator's tendency to differ from sample to sample. An estimator with low variance is said to be precise. Variance usually can be assessed accurately from the sample itself. Estimators typically are reported with an associated variance that has been computed from the sample. Mean Square Error (MSE): The expected value of the squared difference between a statistical estimator and the quantity it estimates. MSE can also be computed as bias squared plus variance. Note that an estimator can be accurate (low bias) but not precise (high variance) or vice versa, but an estimator with low MSE is both accurate and precise. In this report, "sources of error" refers to MSE: bias, variance, or both. Accuracy: An accurate estimator has a small MSE. This implies that it has little or no bias and a small variance. Sometimes the term "accuracy" is used narrowly to imply low bias only. Note that highly inaccurate estimators are not useful, regardless of their precision. For example, an estimator of harvest that ignores the data and is given as -5 fish every year is perfectly precise (zero variance) but very inaccurate. the data collectors incorrectly identified a fish species in the intercept sample. Other issues that need to be addressed in assessing estimates made from a sampling design are whether the sampling and measurement are being conducted as designed and whether the estimation procedure is matched appropriately to the sample design. For example, are the inter-viewers in the intercept surveys being trained and monitored so that they collect data from a probability sample of anglers? Are the data collected appropriately weighted and combined to take into account the complex- ities of the probability design?
26 REVIEW OF RECREATIONAL FISHERIES SURVEY METHODS EXISTING SURVEYS There are several different survey methods currently being employed throughout the United States to monitor marine recreational fishery catch and effort. Primarily, these efforts are conducted and funded by NMFS, but there are many federalstate cooperative programs as well. While the MRFSS has been the predominant survey nationally, some states have never been a part of this program. For example, as alternatives to the MRFSS, Texas has conducted the Texas Marine Recreational Fishing Survey since 1974, and Alaska conducts the Alaska Sport Fish Statewide Harvest Survey. These states generally compile recreational fishing data for summary reports, but the data are not submitted to the MRFSS for inclusion in the national database. And while the MRFSS is still a major survey for some regions, fisheries managers in each region (e.g., Atlantic, Gulf, and Pacific coasts; Hawaii; Alaska) all use a different combination of methods. Currently, there are at least 13 component surveys conducted by federal or state agencies (Figure 1.2 and Appendix B), largely funded through the MRFSS, that produce data that are compatible with the overall MRFSS goals but that may have significantly different methodologies and statistical properties. The number of surveys has grown as more or better data are needed for a particular sector or geographical area. For example, all states along the west coast have implemented individual surveys that replace the national MRFSS program due to their needs for better assessment data, such as for managing groundfish on an in-season basis. Generally, the MRFSS has worked with partner agencies to develop these supplemental surveys. Because data must be comparable across all surveys for national estimates, NMFS works with the states and the regional fisheries commissions to plan new surveys, test new designs, benchmark data, and analyze and manage collected data. As mentioned previously, most recreational surveys are designed with two separate components that are then combined to estimate total catch. For the MRFSS, these two complementary surveys are the access- point angler intercept survey used to determine catch rate and species composition and the coastal household telephone survey used to deter- mine fishing effort. The Puget Sound Sampling Program in Washington, the Shore and Estuary Boat Survey in Oregon, and the California Rec- reational Fisheries Survey (CRFS) are also telephoneaccess surveys; however, the details of survey design are quite different. Alternately, the
INTRODUCTION 27 FIGURE 1.2 Many different surveys nationwide are used to estimate recreational catch and catch per unit effort (CPUE) data. The MRFSS is used along the eastern seaboard and the Gulf coast, but this survey has been replaced in each of the west coast states with surveys that are more tailored to state management needs. The MRFSS includes an access-point (onsite) intercept angler survey, a telephone survey (offsite), and the For-Hire Survey, which includes on- and offsite components, and occasionally, observers. An overlap of surveys exists in several states and regions. For example, the For-Hire Survey and the Party Charter Survey are both used in California to measure the for-hire sector. The Large Pelagic Survey is administered along the mid- and north Atlantic coast, as well as the For-Hire Survey and the MRFSS. (Since this map does not show all existing surveys, a more complete list is available in Appendix B.) Oregon Recreational Boat Survey and the Ocean Sampling Program (Washington) can be categorized as an accessaccess survey, as are the private boat and artificial structure mode sampling components of CRFS, where boat counts at a specific location are used to determine effort and
28 REVIEW OF RECREATIONAL FISHERIES SURVEY METHODS the access-point intercept is used to gather catch rate and species infor- mation. Due to the diversity of fishing modes and the nature and motivations of the angler population participating in each of these modes, no single survey methodology is likely capable of adequately capturing all the needed data. Recognizing these issues, both the MRFSS and state agencies have augmented the telephoneaccess surveys with more focused surveys to deal with these other fishing modes. For example, the MRFSS For-Hire Survey was designed in 1995 and has now been implemented nationwide (except for Texas, Alaska, Oregon, Wash- ington, and Hawaii) to provide adequate coverage of the guided trip, head boat, and charter boat sectors. Some species of interest, such as salmon, halibut, bluefin tuna, and marlin, require catch card reporting, which results in catch information for those species, depending in part on how many cards are returned. Further, other rare event species that were typically undersampled in the MRFSS are now targeted with the Large Pelagic Survey. NMFS, the regional fisheries commissions, and the states have made many advances that have improved recreational fishing surveys in recent years. However, many of the improvements have been made piecemeal as issues or demands arise. Because the context for marine recreational fisheries management has changed for the reasons noted earlier and because these other advances have occurred and other survey approaches have been developed, it is appropriate now to critique the entire concept of a MRFSS or MRFSS-like effort. The goal is to provide recommen- dations, developed in the following chapters, to facilitate future progress. COMMITTEE APPROACH AND REPORT ORGANIZATION To address questions about the MRFSS and other surveys, NMFS asked the National Academies to assemble a committee to (1) review current survey methods used by existing federal and state cooperative programs to estimate marine recreational fisheries effort, CPUE, and catch and (2) make recommendations for possible improvements to current surveys and possible implementation of alternative approaches. (See Box S.1 for the committee's full statement of task.) The Committee on the Review of Recreational Fisheries Survey Methods, composed of experts in survey design and statistics, biological statistics, fisheries management, and recreational fisheries economics and sociology, met five times over the course of the study (March 1011,
INTRODUCTION 29 2005, in Washington, DC; May 1921, 2005, in San Francisco, CA; July 79, 2005, in New Orleans, LA; September 2224, 2005, in New York, NY; and October 2628, 2005, in Tampa, FL). At each meeting, there were one or more public sessions during which the committee heard from federal and state government officials (including NMFS officials), anglers, representatives of fishing organizations, representatives of other nongovernmental organizations, and members of the public. (These presenters are listed in the acknowledgments section in the front of this report.) The presentations and written information provided at those meetings and at other times, along with various sources of published and unpublished literature, were taken into account as the committee arrived at its findings and recommendations. The committee took the approach of discussing some of the major design issues or data needs associated primarily with the MRFSS but that also occur in some of the other regional or state surveys. The report structure is defined by these issues and not by the surveys themselves. The report focuses on marine (and not freshwater) recreational fishing because that is the interest and purview of NMFS. This report presents general issues and problems with current survey methods and designs in Chapter 2, and Chapter 3 reviews sample design concerns and presents alternative survey methodologies that can be used to improve recreational fisheries data. Subsequent chapters discuss how recreational data are used in stock assessments and some of the issues that need to be addressed if these data continue to be an essential information source for future assessments (Chapter 4) and how new survey methods are needed to collect better quality or more useful data on the human dimensions of recreational fisheries (Chapter 5). Chapter 6 discusses methods for establishing a program structure that may allow for continued improvements in recreational fishing survey methods. The report concludes with a discussion of methods to improve communi- cation and enhance confidence in the national and state survey programs (Chapter 7). Additionally, case studies of three recreational species and their associated survey and management methods are used throughout the report to highlight specific issues; the full details of these examples can be found in Appendix C.