{"doc_desc":{"idno":"DDI_EAP_2021-2023_PLMS-W1_v01_M","producers":[{"name":"Development Data Group","abbr":"DECDG","affiliation":"World Bank Group","role":"Documentation of the survey"}],"prod_date":"2024-12-16","version_statement":{"version":"Version 01 (December 2024)"}},"study_desc":{"title_statement":{"idno":"EAP_2021-2023_PLMS-W1_v01_M","title":"Pacific Labor Mobility Survey 2021-2023","sub_title":"Wave 1","alternate_title":"PLMS 2021-23"},"authoring_entity":[{"name":"Dung Doan","affiliation":"The World Bank"},{"name":"Matthew Dornan","affiliation":"The World Bank"},{"name":"Ryan Edwards","affiliation":"The Australian National University"}],"production_statement":{"producers":[{"name":"The World Bank","abbr":"WB","affiliation":"","role":""},{"name":"Development Policy Centre,  The Australian National University","abbr":"ANU","affiliation":"","role":""}],"copyright":"The World Bank","funding_agencies":[{"name":"Australian Department of Foreign Affairs and Trade","abbr":"DFAT","role":""}]},"distribution_statement":{"contact":[{"name":"Dung Doan","affiliation":"The World Bank","email":"ddoan@worldbank.org","uri":""},{"name":"Ryan Edwards","affiliation":"The Australian National University","email":"ryan.edwards@anu.edu.au","uri":""}]},"series_statement":{"series_name":"Other Household Survey [hh\/oth]"},"version_statement":{"version":"Version 01: Anonymized datasets for public distribution","version_date":"2024-09-27"},"study_info":{"abstract":"Previous surveys on labor migration from Pacific Island countries are often cross-sectional, not readily available, and focusing on one migration scheme, country, or issue and hence incompatible. Such limitation of existing data restricts analysis of a range of policy-relevant issues that present themselves over the migrants' life cycle such as those on migration pathways, long-term changes in household livelihood, and trajectory of migrants\u2019 labor market outcomes, despite the significant impacts of labor migration on the economy of the Pacific Island countries. To address these shortfalls in the Pacific migration data landscape, the PLMS is designed to be longitudinal, spanning multiple labor sending and receiving countries and collecting omnibus information on both migrants, their households and non-migrant households. The survey allows for disaggregation and reliable comparative analysis both within and across countries and labor mobility schemes. This open-access and high-quality data will facilitate more research about the Pacific migration, help inform and improve Pacific migration policy deliberations, and engender broader positive change in the Pacific data ecosystem.","coll_dates":[{"start":"2021-12-08","end":"2022-02-07","cycle":"Face-to-face segment of the household survey"},{"start":"2022-11-25","end":"2023-03-09","cycle":"Phone-based segment of the household survey"},{"start":"2022-11-21","end":"2023-03-15","cycle":"Worker survey"}],"nation":[{"name":"Australia","abbreviation":"AUS"},{"name":"Kiribati","abbreviation":"KIR"},{"name":"New Zealand","abbreviation":"NZL"},{"name":"Tonga","abbreviation":"TON"},{"name":"Vanuatu","abbreviation":"VUT"}],"geog_coverage":"Tonga: Tongatapu, \u2018Eua, Vava\u2019u, Ha\u2019apai, Ongo Niua.\nVanuatu: Malampa, Penama, Sanma, Shefa, Tafea, Torba. \nKiribati: Abaiang, Abemama, Aranuka, Arorae, Banaba, Beru, Butaritari, Kiritimati, Maiana, Makin, Marakei, Nikunau, Nonouti, North Tabiteuea, North Tarawa, Onotoa, South Tabiteuea, South Tarawa, Tabuaeran, Tamana, Teraina.","analysis_unit":"- Households in Kiribati, Tonga, and Vanuatu. \n- Temporary migrant workers from Kiribati, Tonga and Vanuatu who participated in the Pacific Australia Labour Mobility scheme in Australia and the Recognised Seasonal Employers scheme in New Zealand","data_kind":"Sample survey data [ssd]","notes":"The PLMS has two parts: a worker survey and a household survey. The scope of the questionnaires is as follow. The household survey includes modules on Household rosters, Socio-demographics, Education, Children, Employment, Income and expenditure, Housing and assets, Remittances from household and non-household members, Perception on impacts of temporary migration schemes, and Gendered impacts of temporary migration.  The worker survey includes modules on Socio-demographics, Health, Employment, Income and expenditures, Remittances, Migration history, Communication,  COVID-19, Social impacts of temporary migration, and Gender."},"method":{"data_collection":{"sampling_procedure":"Sampling frame:\nThe PLMS sample was designed based on a Total Survey Error framework, seeking to minimize errors and bias at every stage of the process throughout preparation and implementation.\n\nThe worker sample frame is an extensive list of approximately 11,600 migrant workers from Kiribati, Tonga and Vanuatu who had participated in the RSE and PALM schemes. \nDue to the different modes of interviews, sampling strategies for the face-to-face segment of the household survey in Tonga was different from the rest of the surveys implemented via phone interviews. The face-to-face segment of the household survey selected households using Probability Proportional to Size sampling based on the latest population census listing and our worker sample frame, with technical inputs from the Tonga Statistics Department. The phone-based segment of the household survey used a combination of Probability Proportional to Size sampling based on the existing sample frame and random digit dialing. The design of the sample benefited from technical inputs from the Tonga Statistics Departments and the Vanuatu National Statistics Office, as well as World Bank staff from Kiribati.\n\nAs participation in the survey is voluntary, a worker might agree to participate while their household did not, and vice versa. Because of this, the survey did not achieve a complete one-to-one match between interviewed workers and sending households. Of all interviewed respondents, 418 workers in the worker survey are linked to their households in the household survey. However, after removing incomplete interviews, 341 worker-household pairs remain. They are matched by either pre-assigned serial ID numbers or contact details collected in the household and worker surveys during the post-fieldwork data cleaning process.","sampling_deviation":"The survey was originally planned to be conducted face-to-face and was so for most of the collection of household data in Tonga. However, due to COVID-19, it was switched to phone-based mode and the survey instruments were adjusted accordingly to better suit the phone-based data collection while ensuring data quality. In particular, the household questionnaire was shortened, and sampling strategy changed to a combination of Probability Proportional to Size sampling based on the existing household listing and random digit dialing.\n\nCompared to in-person data collection, the usual caveats of potential biases in phone-based survey related to disproportional phone ownership and connectivity apply here. The random digit dialing approach provides data representative of the phone-owning population. Yet due to lack of information, it is difficult to judge whether sending households in Kiribati, Tonga, and Vanuatu are more or less likely to own a phone and\/or respond positively to survey request than non-sending households.","coll_mode":["Computer Assisted Personal Interview [capi]"],"research_instrument":"- The questionnaires were jointly designed in English by the World Bank and researchers at the Development Policy Centre, Australian National University. They were translated into Bislama, Gilbertese and Tongan, scripted into CAPI\/CATI programs, tested and piloted before being finalized. The design of the questionnaires and the samples benefited from technical inputs from the Tonga Statistics Departments, Pacific consultants, and academic experts specialized in Pacific labor mobility and remittances. \n- Enumerators are native speakers from the labor-sending countries covered in the survey and were trained to elicit information asked in the questionnaire in local languages. \n- The phone-based household questionnaire is moderately shorter than the in-person version.","coll_situation":"Computer Assisted Personal Interview (CAPI) and Computer Assisted Telephone Interview (CATI)","act_min":"Rigorous quality control measures were put in place during the implementation of the PLMS. \nEnumerators were selected based on their performance during training sessions. Their performance during survey implementation was regularly assessed; briefing, feedback and corrective actions were provided when issues were detected.\n\nThe questionnaire for the face-to-face segment of the household survey was scripted in Survey Solutions, a CAPI application developed by the World Bank. For implement the phone-based segment of the household survey and the worker survey, the questionnaires were scripted in TELSIA (Telephone Survey Integrated App), an all-in-one CATI application. Both Survey Solutions and TELSIA allow us to capture a wide range of data with ease, with features allowing enumerators to synchronize collected data to a secured server, and supervisors to monitor progress and detect issues in real time.\n\nThe scripted questionnaires were checked and tested by the core survey team. Once the first version of a scripted questionnaire was ready, the field team conducted a survey plot with about 40-50 households. It assessed the questionnaire duration, the CAPI\/CATI scripts, the survey protocols, and all operating procedures. The CAPI\/CATI applications were adjusted during and after the pilot. The enumerators were then trained again on a fully debugged application.\n\nData collection was carried out under stringent protocols to maximize response and reach the target sample size. This includes:\n1) Emphasis on the initial contact and introduction\n2) Emphasis on the voluntary nature and confidentiality of participation in the survey\n3) For phone-based data collection, for each phone contact enumerators attempted least seven calls at different times of the day on different days of the week to optimize rate of response. This was then followed by another call by the supervision team to those who refused to participate in the survey.\n\nIn terms of data management and data quality assurance, a Comprehensive Field Quality Control (CFQC) system ran over the entire data collection period to detect errors and undesirable interviewer effects, and to take corrective actions on time. The CFQC system includes:\n1) Computerized built-in validation checks in the CAPI\/CATI application detected issues in the collected data, based on a wide range of criteria on consistency, validity, and value range.\n2) A comprehensive monitoring dashboard provided daily statistics on collected data as well as performance of each enumerator.\n3) A composite Interviewer Risk Index synthesized the survey-specific data quality indicators and swiftly identified underperforming interviewers. This led to the elimination of interviews by one interviewer in Vanuatu and two interviewers in Kiribati. \n4) Non-response items were closely monitored to minimize potential nonresponse bias.\n5) Double-blind back-calls were used to verify the survey implementation by enumerators and to check the accuracy of the collected data. \n6) For face-to-face data collection, a map of all interview locations based on built-in GSP coordinates were used to monitor the consistency between fieldwork and sample design.\n7) For phone-based data collection, random audio supervision by a team of auditors monitored compliance with the calling and interviewing protocols as well as the accuracy of the registered answers.\n8) The principal investigators had direct access to all CFQC elements and collected data in real time. The investigators had weekly supervision meetings with the field team to discuss and address undesirable issues when detected.","cleaning_operations":"The published data have been cleaned and anonymized. All incomplete interview records have been removed from the final datasets. The anonymization process followed the theory of Statistical Disclosure Control for microdata, aiming to minimize re-identification risk, i.e. the risk that the identity of an individual (or a household) described by a specific record could be determined with a high level of confidence. The anonymization process employs the k-anonymity method to calculate the re-identification risk. Risk measurement, anonymization and utility measurement for the PLMS were done using sdcMicro, an add-on package for the statistical software R for Statistical Disclosure Control (SDC) of microdata.\n\nSince the household questionnaire was shortened when the survey switched from face-to-face to phone-based data collection, there face-to-face datasets and phone-based datasets are not identical, but they are consistent and can be harmonized. The mapping guide enclosed in this publication provides a guide to data users to wish to harmonize them.\n\nHousehold expenditure variables in the household dataset and individual wage variable in the household member dataset are in USD. Local currencies were converted into USD based on the following exchange rates: 1 Tongan Pa'anga= 0.42201412 USD; 1 Vanuatu Vatu= 0.0083905322 USD; 1 Kiribati dollar= 0.66942499 USD."},"analysis_info":{"response_rate":"Face-to-face segment of the PLMS household survey: not applicable. Phone-based segment of the PLMS household survey: 26%. The PLMS Worker survey: 31%"}},"data_access":{"dataset_use":{"contact":[{"name":"Dung Doan","affiliation":"The World Bank","email":"ddoan@worldbank.org","uri":""}],"cit_req":"World Bank and Australian National University (2023). Pacific Labor Mobility Survey (PLMS) 2021-2023, Waves 1. Ref ITA_2024_UKR-REF_v01_M. Washington D.C. Dataset downloaded from [url] on [date].","disclaimer":"The user of the data acknowledges that the original collector of the data, the authorized distributor of the data, and the relevant funding agency bear no responsibility for use of the data or for interpretations or inferences based upon such uses."}}},"schematype":"survey","tags":[{"tag":"DOI"}]}