were assigned to one of three options of the tone and position by 0, 1 method. The coding of the selected data was carried out by two teams of authors independently to run reliability testing. The first wave of coding was carried out starting on 15 January to 22 January 2021. The second wave of coding was carried out by switching the researchers and coding the same comments for a second time. A screenshot/link of each comment was included in the codebook. A second coder coded the screenshot for the second time blindly. Int. J. Environ. Res. Public Health 2022, 19, 5737 5 of 14 2.2. Sampling Technique Country selection: A convenient sampling technique was used to include the countries. The authors developed a selection strategy to include social media posts from the official pages of health authorities as noted below: In countries where the vaccine was delivered, the first post that addressed the delivery of the COVID-19 vaccine to the corresponding country was extracted and considered in the data collection process; In countries where the vaccine had not been delivered yet, the first post that addressed the COVID-19 vaccine, in general, was extracted and considered in the data collection process. For example, if Nigeria had posts related to COVID-19 delivery in the country, the authors identified the social media pages of the official health channels in the country, then scrolled down to find the first post that talked about vaccine delivery, then started collecting the comments on this post, with the following post to be the one to be analyzed. If the authors checked the official pages and couldn’t find any posts related to vaccine delivery (because the vaccine had not yet been delivered to the country), the authors looked for any posts that talked about COVID-19 vaccine in general. For Facebook, in each of the included posts, its comments were screened and collected through exploring the “most relevant” filter category in each port on Facebook. Comments were collected until reaching the determined sample size for each country. If the required number of comments wasn’t reached, the next post that addresses the COVID-19 vaccine was extracted and the comments were collected using the same method. For Twitter, systematic random sampling was applied by collecting every other reply to the desired tweet. We divided the number comments “N” by sample size “n” to calculate the sampling interval “i”. In case this value was in decimals, we rounded the figure to the nearest whole number/integer. Then, a random starting point, “r”, was chosen from where the sampling interval “i” is used to pick responses. 2.3. Statistical Analysis All data were assessed using SPSS, version 26.0. Figures were built using ggplot2 package for R software version 4.1.2. Descriptive frequencies were calculated for qualitative variables (comments and tone). Categorical variables were presented in total numbers (n) and percentages of all recorded posts. The Chi-squared test was applied for evaluation of association between country and attitude or comments towards vaccination. Kappa statistic was performed to test for interrater agreement; (values < 0 as indicating no agreement and 0–0.20 as slight, 0.21–0.40 as fair, 0.41–0.60 as moderate, 0.61–0.80 as substantial, and 0.81–1 as almost perfect agreement). p values < 0.05 were considered significant. The Kappa value for the inter-coder reliability was 0.85, which indicated a strong agreement between the coders. 3. Results 3.1. Description of the Collected Data In total, 4897 (83.5%) of comments were from Facebook, while 965 (16.5%) were from Twitter. The comments were written in English 46.0% (2269), Portuguese 15.6% (912), Arabic 20.5% (1203), German 6.6% (388), Malay 5.2% (307), Burmese 1.7% (100), Thai 1.5% (87), Spanish 1.8% (106), French 0.1(5) and Swedish 0.9% (55). Comments were collected from 24 countries: United States of America (USA) (2176), Brazil (846), Saudi Arabia (385), United Kingdom (UK) (398), Egypt (381), Germany (388), Malaysia (307), Myanmar (100), Morocco (150), Mexico (106), Thailand (87), United Arab Emirates (UAE) (398), Tunisia (36), Portugal (66), Iraq (76), Sweden (55), Libya (20), Jordan (59), Palestine (45), Kuwait (28), Oman (27), Lebanon (13), Sudan (11), and Senegal (5). Int. J. Environ. Res. Public Health 2022, 19, 5737 6 of 14 3.2. Position of Social Media Users towards COVID-19 Vaccine 3.2.1. On the Country Level Facebook Acceptance rate: The overall vaccine acceptance in Facebook was 40.3% (1975/4897) in the countries. Countries that had a low acceptance rate were the USA (22%), UK (22.3%), Mexico (22.6%), and Palestine (24.4%), followed by Egypt (48.5%), Myanmar (57.0%), Thailand (50.6%), and Iraq (56.6%), Portugal (49.2%), and Germany (38.2%). Countries that had high acceptance rates were Saudi Arabia (88.3%), UAE (76.1%), Libya (75.0%), Brazil (67.5%), and Kuwait (65.2%). The difference in the acceptance rate among countries was statistically significant (p < 0.001) (Table 1). Rejection rate: UAE (3.3%), Myanmar (4.0%), Thailand (5.7%), Oman (5.0%) had the lowest rejection rates among all countries followed by that had a low rejection rate were Saudi Arabia (9.6%), Sweden (20.0%), Brazil (21.6%), Kuwait (21.7%), Mexico (20.8%), and Libya (25.0%). Countries that had medium rejection rates were Egypt (33.6%), Portugal (36.5%), Iraq (35.5%), Tunisia (38.9%),