Disaster SS Part B 080918 clean

SUPPORTING STATEMENT

ASSESSMENT OF THE SOCIAL AND ECONOMIC IMPACT OF HURRICANES AND OTHER CLIMATE-RELATED NATURAL DISASTERS ON COMMERCIAL AND RECREATIONAL FISHING INDUSTRIES IN THE EASTERN, GULF COAST, AND CARIBBEAN TERRITORIES OF THE UNITED STATES

OMB CONTROL NO. 0648-xxxx

B. COLLECTIONS OF INFORMATION EMPLOYING STATISTICAL METHODS

1. Describe (including a numerical estimate) the potential respondent universe and any sampling or other respondent selection method to be used. Data on the number of entities (e.g. establishments, State and local governmental units, households, or persons) in the universe and the corresponding sample are to be provided in tabular form. The tabulation must also include expected response rates for the collection as a whole. If the collection has been conducted before, provide the actual response rate achieved.

Target population

The respondent universe for this study includes a variety of sectors from the fishing industries in Maine through Texas, Puerto Rico, and U.S. Virgin Islands that can be impacted by hurricanes and other natural disasters. Types of respondents expected are commercial and recreational (for-hire) fishing vessel owners, bait and tackle shop owners and/or managers, seafood dealers, seafood processors, marina/boat repair/marine supply owners and/or managers, and aquaculture facilities. The different sectors targeted in this study were grouped into six categories identified in Table 4.

Table 4. Target population in each of the sector categories to be surveyed

Sector Categories	Target Population
Fishermen	Individuals that own fishing vessels operating commercial or recreational (for-hire) businesses
Bait & Tackle Stores	Individuals that own or manage bait & tackle stores
Marinas/ Boat Repair Yards/ Marine Supply Facilities	Individuals that own or manage marinas, boat repair yards, and marine supply stores
Seafood Dealers	Individuals that own or manage facilities dealing seafood
Seafood Processors	Individuals that own or manage facilities operating seafood processors
Aquaculture Facilities	Individuals that own or manage an aquaculture facility

Target population universe and sample size

In the context of this research, defining a numerical estimate of the respondent universe is challenging, due to the diversity of sectors that will be assessed and because there is no single source of information from which a respondent universe can be assembled. Therefore, values for calculating the respondent universe (Table 5) come from a combination of published data and information from personal communications.

The respondent universe for this study was assembled from a number of different sources including NOAA Fisheries license files, state license files, fishing industry organizations, prior NOAA Fisheries data collections, the internet, and other key informants. Specifically, published data for delimiting the number of permitted commercial and recreational (for-hire) vessels and seafood dealers came from the NOAA Fisheries Northeast Regional Office (NERO), the Southeast Regional Office (SERO), and the Atlantic Coastal Cooperative Statistics Program (ACCSP). The number of bait and tackle stores and seafood processing facilities are estimated based on previous data collection efforts by NOAA Fisheries Office of Science and Technology. The number of marina/boat repair/marine supply businesses in coastal counties is based on North American Industrial Classification System (NAICS) coded business location data from ESRI and personal communication.

The estimated respondent universe for Maine through Texas, Puerto Rico, and U.S. Virgin Islands are presented in Table 5. Note that any one disaster will affect only a limited subset of this overall population. Following a disaster, a specific frame will be assembled which will include only those individuals located in affected areas. Further it should be noted that any one disaster might only affect a part of a state or parts of multiple states. In such cases, we will use the available regional resolution (usually counties) to assemble the most appropriate frame. Sampling is only conducted among the potentially affected population the size of which is disaster specific and unknowable at this point.

To illustrate how sampling might look, samples are drawn for each sector for each state and territory in Table 5. The estimated minimum sample sizes (see table 5 below) were calculated using a 5% confidence interval and 95% confidence level for each strata, specifically each sector by state, using as basis the estimated universe population described in Table 5. The minimum statistical sample size is inflated by 20% to account for expected non-response. The minimum sample size for all sectors, states and U.S. Caribbean territories is 18,746.

Table 5. Estimated respondent universe and estimated sample by sector for each state

*Note: The sample size numbers were inflated to include a 20% non-response rate.

Expected response rate

To maximize the response rate, we will work with state and local officials and organizations engaged with different sectors of the fishing industry in each state to broadly advertise the survey prior to implementation. As mentioned earlier the extent and type of damage to infrastructure such as phone lines, electrical power and cell towers will determine the most effective method for data collection. This study will make use of four methods for data collection: telephone, fillable-online, mail, and intercept, face-to-face surveys. Precise information on expected response rates are not currently available because researchers involved in this study have not previously conducted interviews with the fishing industry applying all four methods in one effort and potential response rates for each method are expected to differ.

While response rates for internet-based surveys tend to be lower than other modes (Cook et al., 2000; Couper, 2000), Dillman et al. (2009) found that a mixed-mode strategy of one data collection followed by another can substantially increase response rates. For example, they found that web-based surveys followed up with a telephone survey can improve response rates by 35%. Dillman et al. (2009) also found that mail surveys followed by telephone contact yielded a total response rate of 82%. Depending on the communication channels available after a disaster, telephone and internet may not be available, which will require employing in-person surveys to collect the data. Extensive previous experience by the researchers involved in this study justifies the use of in-person interviews to reach recreational and commercial fishermen. The intercept method used previously by the investigators to reach fishermen in a study on job satisfaction and well-being in fishing communities in the Mid-Atlantic elicited an 85% response rate (Pollnac et al. 2014).

Based on this information, the overall response rate for this study is expected to be approximately 82% to 85%. The sample sizes described in Table 5 reflect the desirable sample sizes based on the calculation described under Section B, Question 2 below. Oversampling based on the estimated response rate may be employed to maximize the overall sample size.

Once the study is completed, we will calculate the final response rate using the appropriate American Association for Public Opinion Research (AAPOR) Response Rate Calculator.

2. Describe the procedures for the collection, including: the statistical methodology for stratification and sample selection; the estimation procedure; the degree of accuracy needed for the purpose described in the justification; any unusual problems requiring specialized sampling procedures; and any use of periodic (less frequent than annual) data collection cycles to reduce burden.

Commercial and recreational fishermen, the largest number of respondents in these data collections, will be contacted predominantly through the telephone survey. Unlike the other businesses in this study whose work place is stationary, fishermen are generally hard to locate as they work at sea, often out of cell phone range, and under conditions that would make interviews unsafe.

Following a disaster, communities most affected by the disaster and most dependent on fishing will be visited by NOAA Fisheries social scientists to conduct general reconnaissance, conduct ethnographic research and collect qualitative information. Ports will be systematically selected using indices of community dependence on commercial and/or recreational fishing developed using factor analysis (Jepson and Colburn 2013). As part of these visits, in-person surveys (of the quantitative economic information) might be conducted with selected, location-based respondents (dealers, processors, etc.).

The estimated sample sizes (see table 5 above) were calculated using a 5% confidence interval and 95% confidence level for each strata, specifically each sector by state, using as basis the estimated universe population described in Section B, Question 1 above. The required statistical sample size is inflated by 20% to account for expected non-response. The sample selection process will be a random sample approach in each stratum. In other words, each individual commercial and for-hire vessel owner, bait and tackle store owner/manager, dealer, marina, boat repair and marine supply store owner/manager and aquaculture facility owner/manager is considered one respondent unit and each one, in the fishing industries of each state and/or U.S. Caribbean Territory will have an equal chance of being selected within each stratum. Note that due to the selected statistical precision, in many smaller strata we will in fact be conducting a census.

No unusual problems are expected; therefore, specialized sampling procedures will generally not be needed. An exception might be, if a disaster involves a particularly large population (with substantially different damage profiles in different areas, as is usual), we might stratify the population further based on expected disaster impact (e.g., high impact, category 5 storm impact areas vs. lower impact, category 4 impact areas). This will allow us to adjust our sampling intensity by impact-strata to a) stay in budget while b) still achieving adequate coverage of the high impact areas.

This is an as-needed data collection intended to capture information regarding the impacts of hurricanes and other natural disasters either shortly after the event or one-year, post impact.

3. Describe the methods used to maximize response rates and to deal with nonresponse. The accuracy and reliability of the information collected must be shown to be adequate for the intended uses. For collections based on sampling, a special justification must be provided if they will not yield "reliable" data that can be generalized to the universe studied.

Various steps will be taken to maximize response rates. NOAA Fisheries will work with state agencies to coordinate press releases notifying the public of the survey, its purpose, and the different ways it will be administered. To maximize response rates survey administrators will conduct the survey in four ways: over the phone, online, mail and in-person. Name, address and phone number of potential respondents will be assembled from existing sources including federal and state agencies, fishing businesses, and fishing organization membership lists. For the rapid assessment, the telephone will be used as the primary way to survey the entire sample population with a telephone number. An online survey will be made available to potential respondents that do not want to complete the survey over the telephone. In-person interviews will be conducted in conjunction with site visits; when telephone service is not available, or to specifically target respondents who are not responding to the other methods of contact. For the long-term assessment, telephone in conjunction with mail will be used to survey the entire sample population with an address and phone number. In this study, a mixed mode survey approach will be used because there is evidence that response rates will increase if a respondent who did not complete a survey with one mode is offered a different mode (de Leeuw 2005: 233-255).

For the telephone interview, each potential respondent will be called up to five times before he/she is recorded as a non-respondent. Following the Pew Research Center’s approach, the calls will be staggered over times of day and days of the week (including at least one daytime call) to maximize the chances of making contact with a potential respondent. Interviewing will also be spread as evenly as possible across the survey period. The number of calls where contact was made, a survey was successfully completed, and refusals will be recorded (Pew Research Center 2013). Telephone respondents will also have the opportunity to complete the survey online rather than over the phone if they prefer.

To decrease the potential for nonresponse, the survey instrument has been carefully designed to ensure that questions are posed in simple and straightforward language and are as brief as possible without compromising the quality of information obtained. Moreover, prior to the implementation of the survey, interviewers will explain that the survey is anonymous, participation is voluntary and that the interview can be stopped at any point. It will also be explained that participants can skip questions they do not want to answer.

In the face of an unexpected and significant frequency of nonresponse that could lead to potentially biased results, the data in-hand on respondents and non-respondents will be compared to investigate differences that could indicate biased results. If bias is suspected, demographic and other relevant information about the specific target sectors, available prior to contact and obtained through the surveys, will be used to adjust weights for non-response. This approach has been extensively used to address non-response bias (Carlson and Williams 2001, Little and Vartivarian 2003). The type and extent of information that is readily available on the target populations as well as information that will be obtained during the data collection are considered appropriate to adjust the weights of respondents presenting similar characteristics to non-respondents if such approach is necessary. If a strong bias is suspected, a brief non-response telephone survey might be conducted to roughly quantify the impact of the bias.

Contact has been made with key members of NOAA Fisheries, academia, and industry to better understand the study universe.

4. Describe any tests of procedures or methods to be undertaken. Tests are encouraged as effective means to refine collections, but if ten or more test respondents are involved OMB must give prior approval.

A review of the study description, the study methodology, and the survey instrument has been undertaken. NOAA Fisheries personnel in the Northeast and Southeast regions have reviewed the survey tool and provided comments on both the survey tool and the study.

The survey questions in this project are based on a Hurricane Sandy one-year assessment (OMB Control No. 0648-0686), which was tested and implemented in 2013-2014. A total of 952 interviews were conducted with commercial and recreational fishermen/vessel owners (N=522), seafood dealers (N=87), Bait and tackle stores (N=94), Marina owners/managers (N=235), and aquaculture facilities (N=14). The results of the Hurricane Sandy assessment (Colburn et al. 2015 NOAA Tech Memo; Clay, Colburn, & Seara 2016; Seara, Clay, & Colburn 2016) were used to improve the clarity of questions for both the rapid and long-term surveys of the proposed study.

Statistical tests employed in the proposed assessments are expected to be similar to those used in the above Colburn et al. 2015 NOAA Tech Memo. For example, the Mann-Whitney U statistic was used for all mean value comparisons between two independent groups involving total value of physical damages/losses and percent revenue lost. Comparisons involving multiple groups were conducted using Kruskal-Wallis one-way analysis of variance. Non-parametric tests were chosen in order to account for non-normality of data distribution and the presence of outliers. In addition, where appropriate, the total and average value of impacts by sector may be reported.

5. Provide the name and telephone number of individuals consulted on the statistical aspects of the design, and the name of the agency unit, contractor(s), grantee(s), or other person(s) who will actually collect and/or analyze the information for the agency.

The internal NOAA Fisheries design, development, and review team including statistical analysis includes the Principal Investigators Dr. Mathew McPherson (Southeast Fisheries Science Center; 646-289-2235), Dr. Michael Jepson (Southeast Regional Office; 727-551-5756) and Dr. Lisa L. Colburn (Office of Science and Technology; 401-782-3252).

The primary individuals expected to collect the data will be NOAA Fisheries social scientists and contractor social scientists from the Southeast Fisheries Science Center, Southeast Regional Office and Northeast Fisheries Science Center.

REFERENCES

Carlson, B. L. and S. Williams (2001). “A comparison of to methods to adjust weights for non-response: propensity modeling and weighting class adjustments”. In Proceedings of the Annual Meeting of the American Statistical Association, August 5-9, 2001. https://www.amstat.org/sections/srms/Proceedings/y2001/Proceed/00111.pdf.

Clay, P.M., L.L. Colburn, & T. Seara (2016). “Social bonds and recovery: An analysis of Hurricane Sandy in the first year after landfall.” Marine Policy ??: 1-7.

Colburn, L.L., P.M. Clay, T. Seara, C. Weng, & A. Silva (2015). “Social and economic impacts of Hurricane/Post Cyclone Sandy on the commercial and recreational fishing industries: New York and New Jersey one year later.” U.S. Department of Commerce, NOAA Technical Memorandum NMFS-F/SPO-157, 68p., August 2015.

Cook, C., F. Heath, & R.L. Thompson (2000). “A meta-analysis of response rates in web- or Internet-based surveys.” Educational and Psychological Measurement 60: 821-826.

Couper, M.P. (2000). “Web surveys: A review of issues and approaches.” Public Opinion Quarterly 64: 464-494.

de Leeuw E.D. (2005). “To mix or not to mix data collection modes in surveys”. Journal of Official Statistics 21(2):233–255.

Dillman, D.A., G. Phelps, R. Tortora, K. Swift, J. Kohrell, J. Berck, & B.L. Messer (2009). “Response rate and measurement differences in mixed-mode surveys using mail, telephone, interactive voice response IVR) and the Internet.” Social Science Research 38: 1-18.

Jepson, M. and L. L. Colburn (2013). “Development of Social Indicators of Fishing Vulnerability and Resilience in the U.S. Southeast and Northeast Rergions. U.S. Department of Commerce, NOAA Technical Memorandum NMFS-F/SPO-129, 64p., April 2013.

Little, R. J. and S. Vartivarian (2003). “On weighting the rates in non-response weights”. Statistics in Medicine 22: 1589-1599.

Marshall, N.A., and P.A. Marshall (2007). "Conceptualizing and Operationalizing Social Resilience within Commercial Fisheries in Northern Australia." Ecology and Society 12: 14.

Melillo, J.M., T.C. Richmond, & G.W. Yohe, Eds., 2014: Highlights of Climate Change Impacts in the United States: The Third National Climate Assessment. U.S. Global Change Research Program, 148 pp.

PEW Research Center. “Our Survey Methodology in Detail.” Available at http://www.people-press.org/methodology/our-survey-methodology-in-detail/. Accessed on October, 2013.

Pollnac, R.B., T. Seara, & L.L. Colburn (2014). “Aspects of Fishery Management, Job Satisfaction, and Well-Being among Commercial Fishermen in the Northeast Region of the United States.” Society and Natural Resources 01:1-18.

Seara, T., P.M. Clay, & L.L. Colburn (2016). “Perceived adaptive capacity and natural disasters: A fisheries case study.” Global Environmental Change 38: 49-57.

APPENDIX A

Table 6: Hurricane direct hits on the mainland U.S. coastaline and for individual sates 1851-2017 by Saffir/Simpson Cateogry.

From: Blake, E.S., E.N. Rappaport, J.D. Jarell, & C.W. Landsea, 2005: "The Deadliest, Costliest, and Most Intense United States Hurricanes from 1851 to 2004 (and Other Frequently Requested Hurricane Facts.) NOAA Technical Memorandum NWS-TPC-4, 48 pp. and Jarell, J.D., B.M. Mayfield, E.N. Rappaport, & C.W. Landsea, 2001: "The Deadliest, Costliest, and Most Intense United States Hurricanes from 1900 to 2000 (and Other Frequently Requested Hurricane Facts.) NOAA Technical Memorandum NWS-TPC-3, 30 pp.

NOAA Hurricane Research Division: http://www.aoml.noaa.gov/hrd/tcfaq/E19.html