TVSEP Covid-19 Survey in Thailand completed

The household survey implemented during November and December 2020 to measure the impact of Covid-19 in Northeast Thailand was completed on Christmas Day, 25 December. The survey team of the TVSEP Data Collection Center at Ubon Ratchathani University held their final meeting in the Province of Nakhon Phanom. The conduct of the survey has been a big challenge, because the TVSEP team from LUH was not able to go to Thailand for the supervision of the survey, as was the case during the many TVSEP survey waves in the past. All training and supervision had to be done via online formats.

In Thailand, the interviews were implemented through visits of enumerator teams in the TVSEP panel households. Of course, interviews were conducted in compliance with the government rules for protection against Covid-19. Two of the TVSEP provinces were surveyed in parallel. Four teams interviewed the 980 TVSEP households in Ubon Ratchathani and three teams the 820 households in Buri Ram. Thereafter, all teams went to Nakhon Phanom to interview the 400 TVSEP households in the smallest among the three TVSEP provinces. Every day, each team of 2-3 enumerators conducted interviews in a TVSEP village with 10 households and the village head, by means of computer assisted personal interviews (CAPI) using tablets. The questionnaires were uploaded to the TVSEP server at LUH immediately after an interview was completed and subsequently checked by Data Checking Assistants who were student assistants and other TVSEP research staff.

Overall, the interviews of the 2.200 rural household in the 220 villages in the three TVSEP provinces went smooth and without major problems. Considering that the survey was implemented under the conditions of the pandemic, attrition was remarkably low. Only 59 households could not be interviewed, out of which just 8 respondents refused to be interviewed. This is an amazing result considering that the TVSEP household panel in Thailand is now in its 14^th year since its inception in 2007. TVSEP is very grateful to its panel households for this exceptional loyalty.

The most remarkable result of the survey that none of the respondents has reported any Covid-19 infections in her household. On the other hand, the survey team was lucky to complete the survey as a new Covid-19 hot spot had just emerged in Samut Sakhon province, Central Thailand, reportedly sparked by illegal migrants from neighbouring Myanmar. Since this outbreak has now evolved into Thailand’s second wave of Covid-19, TVSEP will capture its effects in the next regular household survey during 2021.

The dataset of this survey offers unique insights of the direct impact of Covid-19 on rural households during and after the first wave of the pandemic with only minimal memory bias. The data contain interesting and scientifically useful information which can show, for example, the short and medium term effect of the Thai Government’s Covid -19 support measures. The careful development of the survey instrument with the help of scientists at LUH and local experts from Thailand, the extensive testing of the questionnaire, the dedicated work by enumerators and their local supervisors, the close monitoring and supervision of the survey process and the timely online monitoring of the questionnaires has contributed to the success of the survey. The mission also showed, what can still be done in empirical research, under the difficult conditions of a global pandemic.

Over 50 persons have been involved or have contributed in various ways to the Covid-19 survey in Thailand. These includes the TVSEP partners and staff at Ubon Ratchathani University, who administered the survey in Thailand, researchers at LUH and elsewhere, who contributed to the development and refinement of the survey instrument, the Data Checking Assistants, who performed real-time online checking of all interviews and the survey management and coordination team, who tried to “keep things going” over a distance of some 10.000 km. In the following, persons who contributed in one way or another to the research are named:

The quality of the data collected is demonstrated by some selected statistics which are derived from the survey para-data (see Table 1). Except for the last week, between 400 and almost 500 households were interviewed by the 20 enumerators per week. This is around 3 households per interviewer and day. On average, an interview lasted between 44 and 62 minutes. As can be expected interview time declined as the survey was going on but did not fall the below the three quarters of an hour expected by the survey management team, based on prior testing of the questionnaire.  The effect of the work of the Data Checking Assistants is also revealed from Table 1. While during the first survey week 72 % of the questionnaire (mostly with only minor issues) were returned to the tablets of the enumerators no later than 24 hours after the interview, this has declined to 38 % during the last week. The same can be observed for the second check carried out by TVSEP headquarters which declined from 24 % during the first week to just 1 % during week 5.  On average, 53 % of the questionnaires required some first clarifications and 11 % had to pass through a second round of corrections which was implemented by means of an automated simultaneous data cleaning procedure. In a few cases TVSEP headquarter requested further clarifications which enumerators had to provide with the help of the provincial team leaders. As a by-product of this rigorous checking procedure, the comments made by the enumerators provide valuable additional information, which facilitates a deeper understanding of the household situation. These comments will be provided to data users with the final dataset.

A particular challenge for the interviews is the ageing of the rural population in Thailand. Mostly older people and young children stay in the villages, while the second generation often works as migrants in urban areas. Inevitably, many of our respondents, normally the head of household, were over 60 years old and most of the respondents were female (see Figure 1). Our eldest respondent was a 93-year old man. This may raise questions as to his ability to participate in an interview. However, his fitness for interview was confirmed by the enumerator upon receiving a warning message triggered by the tablet program in case a respondent was over 80. Also, on a less serious note, ex-chancellor Helmut Schmidt gave his last interview at the age of 96. Indeed, we found in past TVSEP surveys, the health of advanced-aged people in rural Thailand very often being quite good.

On the last day of the survey, TVSEP organized a final survey meeting in Nakhon Phanom Province. The meeting was chaired by the Dean of the Faculty of Agriculture of UBU and TVSEP Director in Thailand, Dr. Narintorn Boonbrahm, who thanked the survey team and presented certificates of appreciation to the participants of the survey.  Mr. Niels Wendt and Prof. Waibel participated online and delivered farewell speeches and closing remarks and thanked the team in Thailand for their work. Some selected photos illustrate this event.