Upon completion of interviewing for the day, interviewers must review their questionnaires for completeness and accuracy. Supervisors must review all questionnaires daily. Any apparent mistakes or inconsistencies must be checked with interviewers and, if necessary, the respondent should be contacted to clarify an answer.
Item-by-item instructions for interviewer and supervisor editing apply to the 1995 telephone household survey screening and interview questionnaires (Attachment M).
Screening Questions
Item 1 County of Residence? The county code must be a valid three digit FIPS code for a coastal county in the State being called.
Item 2 Permanent Residence? Only anglers from permanent, year-round residences are eligible to be interviewed.
Item 3a Anglers in Household? This item should have a positive response if there is any further information from this interview.
Item 3b Twelve-Month Fishing? Record the number of anglers in the household who went marine recreational finfishing in the past 12 months in their home State.
Item 4 Two-Month Fishing? Record the number of anglers in the household who went marine recreational finfishing in the past 2 months from their home State. The response must be less than or equal to the 12 month fishing response.
Interview Questions
Item 1 Two-Month Trips? - Record the number of trips made by each angler in the past 2 months from their home state. This number should match the total number of trips profiled (or remembered even if all details were not remembered).
Item 2 Date of Trip? The dates of the fishing trips for an individual angler must be in reverse chronological order. Only dates within 60 days prior to the interview date or "00" are valid.
Item 3 Mode of Fishing? Fishing trips from shore modes must be coded from "1" to "5", while fishing trips from boat modes must be coded from "6" to "8."
NOTE: Particular attention should be paid to high numbers of trips reported for the partyboat and charter boat modes. Reports of more than 3-4 trips in these modes may be due to interviewing a captain or crew member, and occasionally a commercial fisherman who did not understand the screening questions. Reports with more that 3-4 trips must be verified during the interview, and if necessary, through recontact at a later date.
Item 4 Boat Access Type? - Nothing should be coded unless Item 2 is coded as "6", "7", or "8". If public access is coded ("1"), type should be "1" to "4". If private access is coded ("2"), type should be coded "5" to "8".
Item 5 Time of Return? - Time should be between 0000 and 1200. The "am" or "pm" box should be checked. If the time is 1200, then either the "noon" or "midnight" box should be checked.
Item 6 Type of Water Fished In? Ocean, gulf or other open water must be coded with "1." All other inland water bodies should be coded from "2" to "5".
Item 7 County of Fishing? County codes must be among those on the site register for the State being dialed. Only valid three digit FIPS codes are allowed.
Item 8 Oil, Gas or Reef? Fishing trips from States north of New York must have "not applicable" responses. Atlantic coast trips from New York through East Florida must be coded "1" or "3." Boat trips from West Florida through Louisiana must be coded from "1" to "3".
Item 9 Striped Bass? - Fishing trips in States south of Virginia must have "not appropriate" responses. For fishing trips north of North Carolina, if the response in Item 8 is "1", then the response to Item 8a must be greater than 0. If the response to Item 8 is "2", then both 8a and 8b must be 0. The response to Item 8b cannot exceed the value of the response to Item 8a.
Item 10 Saltwater Tournament? - Trips must be coded "1" or "2".
Shellfishing Questionnaire
Item 1 Go Shellfishing? - This item must have a positive response if there is any further information from this interview.
Item 2 Participants? - If Item 1 had a positive response then this response must be greater than 0.
Item 3 Trips? - The response to this item must not be a number less than the number recorded in Item 2. The product also can not exceed the product obtained by multiplying the response to Item 2 by 60.
Item 4 Target Species? - The total of the trips reported for all types of shellfish must equal the number given in response to Item 3.
Telephone household data are stored in a three record hierarchial format with records linked by a unique identifier for each household. The identifier includes year, wave, subregion, state, county and household code number. Data included within each record type are:
Type 1 - Identifying information on the household and numbers of 12 month and 2 month anglers. One record per household. Non-fishing households are not key entered.
Type 2 - Information on each marine recreational angler in the household. One record per angler.
Type 3 - Information on each fishing trip taken during the previous two months. Includes detailed information on mode and area of fishing and gear used. One or more records per angler.
The format for each record type and variable descriptions for 1979-1992 are included in Attachment N.
Data entry procedures must be designed to achieve a data entry accuracy rate of at least 99.5 percent of all keystrokes. Dual data entry or other error checking data entry techniques should be employed.
Error checking must be accomplished through the use of table lookups during data entry and/or editing routines on the data set after data entry is completed for a wave. All checks described in Section 3.4.1 must be incorporated into machine edits. Every data element must be checked for data entry errors, reasonableness in falling within an acceptable range, and logic in relation to other data elements. Duplicate identification numbers must also be identified and corrected.
All of the interviews completed by each new interviewer on the first day of work must be tracked to ensure that the interviewer is following procedures correctly and has good interviewing technique. Throughout the remainder of the survey, 10 percent of each interviewer's work must be validated by interview monitoring or by re-contacting the respondent.
Interview monitoring must be carried out by supervisors. The supervisor would listen to an interview in progress and record the respondent's answers on a second questionnaire. At no time will the supervisor interrupt the interview while in progress. Following the interview, the two questionnaires would be compared and any discrepancies resolved. Significant inconsistencies would necessitate additional intensive checking and validation of work completed by that interviewer. Supervisors should also give suggestions to help improve interviewing techniques.
Re-contacts of respondents attempt to verify that: (1) the interview took place, and (2) responses were coded correctly. At a minimum, information on date, mode, area and gear from the most recent trip recorded on the interview form must be collected during the re-contact to validate the coding of original responses.
The RFP specifies the data reports needed by NMFS on a regular basis to effectively monitor the conduct of the telephone household survey. Included are such items as completion of quotas and dialing results by subregion and state. Dialing results include, but are not necessarily limited to:
Line busy,
No answer,
Answering service,
Answering machine,
Not in service/disconnected number/dead line,
Connected to wrong number,
Household previously called on different number,
Business/coin phone/time/weather/computer tone, etc.,
Institutional housing (dormitory/barracks/nursing home),
Part-year housing,
Wrong county,
Initial refusal,
Communication problem (deaf, foreign language),
Could not answer 12 month question,
Refused to answer 12 month question,
No fishing in last 12 months,
Could not answer 2 month question,
Refused to answer 2 month question,
No fishing in last 2 months,
Number first-time contacts who indicated marine fishing activity but turned out
to be non-fishing households.
Number of anglers who refused to provide trip information,
Number of anglers not available to be interviewed, and
Number of ineligible anglers (e.g., salmon anglers).
Timely submission of these performance data is necessary to maintain data collection quality through identification of problem areas and adjustment of procedures as necessary.
Also of critical importance to quality control is regular reporting from supervisors on interview validation results. Results of all supervisory activities described in Sections 3.1.3 and 3.5.1 must be fully documented. These include observations of in-progress interviews by interviewers, follow-up counseling after in-progress monitoring that identifies problem areas and improves interview technique, and counseling on problem areas following callback verification.
The number of two month anglers contacted in the telephone household survey can be relatively small in several states and waves. Unusually large reported numbers of trips, given these small sample sizes, can result in unrealistic expanded estimates of trips, and then catch. In these cases, it is appropriate to conduct analyses to reduce outliers in the data set.
The method of outlier adjustment used since 1987 involves an examination of frequency distributions of pooled data for trips per household from the previous four years for a particular state, mode and wave. Any observation in the current data set that is outside of the 95th percentile, based on the most recent five years of data (by wave, state, mode) but exclusive of the current year, is reduced to the value of the 95th percentile. Outliers in the raw data base are never changed; the outlier analysis occurs as a step in the estimation process.