Justification for change - recall study

JUSTIFICATION FOR CHANGE

MRIP FISHING EFFORT SURVEY

OMB CONTROL NO. 0648-0652

The MRIP Fishing Effort Survey (FES) was designed to be a more accurate and cost effective alternative to the Coastal Household Telephone Survey (CHTS, OMB Control no. 0648-0652), the methodology historically used to collect recreational fishing effort data. We have completed a series of pilot studies testing design alternatives and examining potential sources of error in the FES design. Specifically, we have examined the impact of cash incentives and different versions of the survey instrument on survey response and survey measures and assessed the potential for bias resulting from nonresponse and under-coverage. These studies have resulted in improvements to the FES design and will help explain any observed differences between CHTS and FES estimates as we transition to the new design. We propose one additional, supplemental information collection that will further assess sources of error, specifically measurement error, in the FES design. This information collection, which is described below, will result in further improvements to the FES design and will help us anticipate and explain the impacts of transitioning from the CHTS to the FES.

The anticipated burden for this request is approximately 10,080 respondents, 10,080 responses and 1,680 hours¹. We will accommodate this additional burden by uniformly reducing FES sampling levels among all states and reference waves such that there will be no net increase in overall burden, and no additional respondents contacted.

Recall Study

The FES is a self-administered mail survey that collects recreational fishing effort data from samples of residential addresses. The survey asks household residents to report recreational fishing activity the occurred during a two-month reference wave. The sampling and data collection designs for the Recall Study will be nearly identical to the FES. However, the reference period will be reduced to one month – respondents will be asked to report fishing activity that occurred during a one-month reference wave. We will assess recall error in the FES by comparing FES estimates to Recall Study estimates that have been aggregated to coincide with FES waves (e.g., we will compare one two-month FES estimates to the sum of two one-month Recall Study estimates). Differences between FES and Recall Study estimates can be attributed to differential recall error between the two reference periods and will be an indication that FES estimates are susceptible recall bias. The Recall Study will also examine the impact of one month waves on the precision and timeliness of recreational fishing effort estimates.

The Recall Study will overlap with the MRIP Fishing Effort Survey (FES) for a period of six months, July-December (three FES waves). The design of the recall study will be nearly identical to that of the FES. For each reference month, a sample of 4,000 residential addresses will be selected from a comprehensive addresses frame maintained by the United States Postal Service. Sampled addresses will be augmented by matching the addresses to state databases of licensed saltwater anglers. This matching screens the sample prior to data collection and allows households with and without anglers to be sampled at different rates. Saltwater fishing activity, including zero-trip data, will be collected from all occupants of the sampled addresses though a self-administered mail survey design. The data collection period for each month will begin one week prior to the end of the month with an initial survey mailing that includes a cover letter stating the purpose of the survey, a survey questionnaire, a post-paid business reply envelope (BRE), and a prepaid cash incentive. The initial mailing will be followed by a thank you/reminder postcard one week following the initial mailing, and a follow-up mailing, including a second questionnaire, a nonresponse conversion letter, and a BRE, three weeks after the initial mailing. Based upon the results of the FES, we expect this design to yield approximately 1,680 completed household interviews each month. Recreational fishing effort estimates generated through the Recall Study will be compared to FES estimates for the same time period. Differences in effort estimates between the studies will be an indication that the different recall periods result in differential recall error. In addition, we will compare the precision of Recall Study and FES estimates to document the impact of recall period (e.g., temporal stratification) on precision.

Note: This subset of unduplicated respondents, and associated responses, was already part of the previously approved burden for this OMB Control No. That is, fewer people will have been contacted for the full study, and that same number, instead, will participate in the recall study.

1 The estimated time per response for the recall study is approximately 10 minutes.

1