to put on his forehead. He was then allowed to enter the polling station when the second body temperature measurement dropped to 32.5°C. While this incident may not reflect the situation in all polling stations, it still shows that risk-mitigation measures were not always adequate. Everyone in the polling stations, including polling staff, voters, party representatives and election observers, was required to wear a face mask at all times. According to the guidelines, the voters should be given a face mask if they did not wear one. Polling staff sometimes took extra steps by distributing face masks and face shields to every voter, even if they already had one. However, the compliance to the face mask rule was marred by a number of polling staff who were found not wearing a face mask or wearing it incorrectly while carrying out their duty in some polling stations. While the majority of the polling stations visited by ANFREL clearly indicated the marks on the ground for a six-feet social distancing as stated in the MOHS guidelines, the real practice of social distancing was poorly enforced and maintained in many places. In many polling stations, the marks on the ground were not sufficient to accommodate the long queue of voters, and those who queued behind failed to keep a safe distance between each other. High turnout at some polling stations with a huge number of voters, especially in the morning, made social distancing difficult or even impossible to practise. For instance, in Yangon’s Mingaladon township, four polling stations with nearly 7,000 voters were located in a school compound, causing a massive overcrowding problem. The MOHS guidelines also stated that hand sanitiser should be readily available inside the polling stations, and different doors should be arranged for entrance and exit to avoid overcrowding. These measures were well implemented in nearly all polling stations visited by ANFREL on Election Day. Counting & Tabulation of Results Polling stations closed at 4 PM, although in some locations, long queues of voters still waiting to cast their ballots delayed closing time by one hour or more. Closing was orderly, and polling stations were immediately converted into counting stations. Across the 22 locations visited by ANFREL observers, the counting of ballot papers was found to be slow and sometimes disorganised, but overall highly reliable. Record numbers of advance votes often delayed the process, as these ballots needed to be removed from their envelopes and assigned to the correct hluttaw. Five hours after vote counting had begun, some polling stations were still processing advance votes and had not yet moved on to Election Day ballots. This explains why many poll workers were up counting ballots until the early hours of the morning, displaying impressive dedication and professionalism while doing so. No substantial errors were reported by ANFREL observers during the counting process. Ballots were properly called and the few mistakes made by UEC staff in assessing ballots were usually corrected immediately by the party agents present. The corrections included in the UEC’s 30 October announcement regarding issues with advance voting were also properly implemented, and no ballots were found to be unjustly discarded. The incidence of invalid ballots was low overall, although it could be further reduced through voter education efforts. In the polling stations monitored by ANFREL, the main causes of invalid votes were ballot papers either not stamped at all or stamped for two or more candidates. Most polling stations also displayed at least one ballot paper cast for the dissolved Union Democratic Party (UDP). This shows that the announcements posted at polling stations were insufficient to inform voters, and the UEC should have undertaken greater communication efforts to avoid voters being disenfranchised. 9