Home
Up
Study parameters
Sample Design
Interview Methods
Response Rates

Interview Methods

The international team held on workshop in Murmansk Russia in February 2003 in part devoted to the development of common interview methods. Our intent is to minimize differences in the meaning of responses given to core questions in the international questionnaire. Common elements of our interview methods include:

Based common interview methods on a common training manual. See Training Manual.
Interview control procedures including use of cover sheets, sample address summary forms, and interview and self-administered form envelopes, all linked with control numbers.
Formatting rules for questionnaires so that all interviewers can be trained to follow the same instructions for reading questions, following skip patterns, and recording responses.
Use of cue cards to efficiently present respondents with response choices.
Training of interviewers to record verbatim responses.
Discussion of question-by-question objectives with interviewers and training of interviewers how to handle respondent questions (including a prohibition of explaining the meaning of questions).
Extensive use of practice interviews so that interviewers are intimately familiar with the questionnaire.
Post-interview editing procedures
Common field control procedures and materials

The training manual describes in detail the above common procedures and materials. Below are links to field sampling and control materials:

Listing sheet
Sample Address Summary Sheet
Interview Disposition Codes
Completed Interview Log
Non-Response Log
Vacancy/Non-Inupiat/Slip Log

 

National Interview Method Experiences

Alaska
Canada

Alaska

Community Approval

The actions of the Alaska Native Management Board constitute the primary component of Native approval of project elements. In addition, we wrote to the traditional council of each selected village to request permission to come to the community to conduct interviews. We also asked the council to select a resident of the village to work on the project during interviewing. We followed these letters up as warranted. In several cases, we traveled to the community in advance of interviewing at the request of a council. No community chose not to participate in the project.

Field Work Organization

Field work was designed as a collaborative effort of four Native organizations (Maniilaq, Kawerak, Ukpeagvik Iņupiat Corporation, and Alaska Inuit Circumpolar Conference) and ISER under the direction of the Alaska Native Management Board. The grant award placed budgets for interviewers, field directors, associated travel costs, and respondent payments within subcontracts to the Native organizations. Our intent was to maximize the involvement of Alaska Natives in all components of the study.

We also recognized that it could be difficult to recruit people to work as interviewers and field directors given generally good job opportunities for people who would also make good interviewers and field directors. We therefore provided ourselves with the flexibility to hire interviewers and field directors through ISER if necessary.

Our intent was to apply an approach to field work successfully used by ISER (and the U.S. Bureau of the Census) over many years. This approach involves the hiring and training of people to serve as team leaders and in turn the hiring of a resident in each small village to be trained by a team leader as an interviewer. This approach combined the objectives of creating a core professional team of interviewers (team leaders) and involving small village residents in the field effort as well.

Recruitment of Field Staff

Due to budget constraints, we divided the field work across two years: 2002 and 2003. We chose to interview in each year during the months of January and February. We have found that these months offer the best opportunity for finding respondents with time for a lengthy interview. In 2002 we interviewed in the Northwest Arctic region. Maniilaq staff, working through village councils and local contacts, tried to recruit one interviewer from each of the five villages and six interviewers from Kotzebue. We hoped that village interviewers could have a week of training in Kotzebue and then remain there working with the Kotzebue-based interviewers to complete the Kotzebue sample in two weeks. Under this plan, the best available Kotzebue interviewers would then accompany the village interviewers to complete the village sample.

Unfortunately our plan was complicated by a delay in processing of the grant award and subcontract with Maniilaq. Maniilaq did not have a subcontract from the University for field work until the second week of January 2002. Interviewer training began January 14. The short time available for recruiting interviewers, coupled with weather-related travel disruptions, meant fewer interviewers attended the training. We supplemented the local interviewers with one Iņupiat from the region who worked for ICC, another Iņupiat from the region who worked for ISER, and four non-Iņupiat interviewers who worked for ISER.

In 2003 we used the same interviewing strategy as we had the year before: a combination of regional center residents, village residents, and ISER interviewers. In the Bering Straits communities it worked much as it had the year before in Kotzebue with a combination of local and non-local interviewers. On the North Slope it was again a combination of local and non-local interviewers, with many more interviews completed in Barrow and the villages by Iņupiat residents.

Training

ISER research staff led training sessions in each of the regional hubs and in the ISER Anchorage office. These session lasted five days, including extensive practice interviewing. The first topic in each training session was about the importance of protecting the respondents' rights. During the training session interviewers read a prepared statement on interviewer ethics. The meaning of ethical conduct, potential harm if anything about or from the interview was discussed, and possible pitfalls were discussed as a group. After questions were clarified, interviewers read and signed a pledge of confidentiality. Each interviewer kept a copy of the pledge and ISER kept the original signed Pledge of Confidentiality.

Through a combination of classroom instruction based on the Interviewer Training Manual, paired interviewer practice, one-on-one mentoring, open discussions, and team field exercises we covered the following aspects of field methods: Interviewer Ethics, Purpose of Survey, Purpose of Training, Steps in Conducting a Sample Survey, Community Introductions, Listing, Selecting Households and Respondents, Getting Ready to Interview, Guidelines for Conduct, General Rules for Interviewing, Introductions, Special Interview Circumstances, Disposition of Interview, Record Keeping, and Question by Question Objectives. Interviewers spent the last two days of training going out in pairs to complete interviews in selected households. Interviewers continued interviewing in pairs until they felt comfortable going alone.

Field Sampling

Field sampling for the regional centers took place in the field office using household listings. In the villages, team leaders worked with local residents to perfect the sample frame, eliminating non-occupied structures and non-Native households.

Using random start and interval selection procedures, housing units were selected and their descriptions entered on the Sample Address Summary Form and Cover Sheets

In the initial days of mapping, listing, and interviewing, two supervisors and sometimes three were in the office. Maps and listing sheets were reviewed, sample assignment sheets were completed, and interviewer assignments were distributed. Interviewer assignments need to be distributed efficiently so that all interviewers always have work available. In the early stages of interviewing each interviewer has a limited geographic area. Not only is it inefficient to walk a long distance from one interview to another, but, in January and February it is cold. Before an interviewer left the office with a new assignment, we checked the map to be sure the interviewer knew how to find the new block. In the beginning questions always arise: a housing unit listed as a single family dwelling turns out to be a three-unit apartment; the selected respondent is away for the entire study period; or how to proceed when a mean-looking dog is chained to the front door. While the training manual covers most situations, interviewers forget or a new twist appears.

Cover sheets for each potential interview are logged-out to specific interviewers. Interviewer assignments are also logged on a copy of the Sample Address Summary Form kept in the survey office. 

The interviewers used the household listing map, cover sheets, and sample address summary form to locate the sampled housing units. They use the cover sheet has a place to record the time, date, and results of each visit to the house. 

Interviewing

The Cover Sheet also has a brief description of the project and instructions for selecting the respondent. The interviewer knocks on the door of the selected housing unit and after a brief description of the project the interviewer asks, "Is there at least one person 16 years old or older living in this household who considers themselves to be Inupiat or Yupik?" If more than one person in the housing unit was eligible to be interviewed, the interviewer asks to interview the person with the next birthday. Studies have demonstrated that this is an unbiased technique for selecting among eligible respondents.

After the appropriate respondent has been determined, the interviewer will interview that person immediately or make an appointment to return when the respondent wishes. The interviewer completes the questionnaire. The interviewer then edits the questionnaire to ensure that all questions have been asked and that all responses, including verbatim responses, are legible and complete before turning it in to the office. 

Each cover sheet is always accounted for in some way. The potential interview becomes a completed interview or it becomes a non-interview of some type as described on the Disposition Code Sheet.  A special type of non-interview is a refusal. Interviewers are trained to write down everything that happens in a refusal. Previous studies have shown that most refusals are circumstantial and have nothing to do with the content of the survey or the specific interviewer. It is more likely that the interviewer has arrived when the baby has just dumped her food on the floor, when the respondent has had a disagreement with her spouse, or generally has had a bad day. A supervisor reviews the notes for clues to the situation to determine the next steps. Frequently the supervisor will wait a week and have a different interviewer return at a different time on a different day of the week and the respondent will cheerfully complete the interview. In Nome, after supervisors discovered there were a number of refusals on Friday evenings, they stopped all interviewing on Friday evenings.

Overall, the combination of interviewers worked well. We did find that respondents were uncomfortable being interviewed by someone they knew, whether or not this person lived in the same community. We adjusted by advising interviewers working in the same village to review their sample assignment and avoid interviews with people they knew. Note that this problem would be severe if a local interviewer had the complete sample assignment.

Interviews took, on average, 90 minutes to complete but varied from under an hour to repeated interview sessions over a three day period.

Interview Processing

Field office staff assigned each completed questionnaire a four-digit study number. The self-administered section, the pages containing the most sensitive questions, is assigned the same study number as its corresponding questionnaire. Once the study number is on the questionnaire and self-administered section, the study number becomes the sole identifier and all other identifiers are removed from the data file. 

The completed questionnaires, self-administered forms, and cover sheets are kept in a locked filing cabinet, accessible only to authorized project staff, until the initial analysis is complete. These materials are kept in case questions arise during the analysis that can be resolved by referring to the original interview materials. Once this stage of the analysis is complete, the cover sheets and self-administered forms are destroyed.

A completed interview is checked for completeness, accuracy, legibility, and logic before the interviewer turns it in. An interviewer's initial interviews are edited by a supervisor and returned to the interviewer immediately so mistakes and points of confusion can be clarified before very many interviews are completed. Supervisors review every interview as soon as possible after it has been completed. If there are questions about the interview, the supervisor returns it to the interviewer immediately for clarification while the interview is still fresh in the interviewer's mind. When the interviewer can't resolve the question, the respondent may be contacted for more information.

In the office, the supervisor maintains a bulletin board that lists all appointments. The board lists appointments and the interviewer who will complete it. Sometimes the interviewer who made the appointment can't keep it; another interviewer will sign-up on the bulletin board to conduct the interview. Because it is a convenient time for the respondent, interviewers will double-book appointments knowing that another interviewer will keep the appointment. Interviewers are instructed to stop in the office once a day. Not only is this a safety protocol for the interviewers, but they can check the appointment board too. It is also a time to replenish supplies, hand in completed interviews, and obtain new assignments. While in the office, a supervisor may review the sample assignment sheet to make certain the interviewer has been visiting the selected housing unit at different times of the day and on different days of the week. This technique increases the chances of finding someone at home.

Canada

Please see the Canada Concepts and Methods Report.

Please continue with Response Rates.