Volume 28, Number 8—August 2022
Research
COVID-19 Vaccination Intent and Belief that Vaccination Will End the Pandemic
, Liesbeth Claassen, Mattijs Lambooij, Ka Yin Leung, Kees Boersma, and Aura Timen
Table 3
Cumulative variance explained and partial dependence of 10 strongest determinants of COVID-19 vaccination intention in random forest model for residents of the Netherlands
| Ten strongest determinants in random forest model | Cumulative variance explained, % | Partial dependence, lowest–highest value* | Direction of relationship with vaccination intention |
|---|---|---|---|
| Vaccination, end of crisis | 54 | 3.9–4.3 | Positive |
| Vaccination, expectations of loved ones | 62 | 4.0–4.4 | Positive |
| Vaccination, developed too quickly | 70 | 4.5–4.2 | Negative |
| Vaccination, side effects well researched | 72 | 4.1–4.3 | Positive |
| Vaccination, approved therefore safe | 72 | 3.9–4.2 | Positive |
| Vaccination, good protection | 73 | 4.2–4.5 | Positive |
| Vaccination, new techniques are safe | 73 | 4.2–4.4 | Positive |
| Vaccination, live sooner without measures | 74 | 4.2–4.3 | Positive |
| Vaccination, behavior of loved ones | 74 | 4.0– 4.3 | Positive |
| Vaccination, possibility severe illness | 74 | 4.4–4.3 | Negative |
*Interpretation of the partial dependence figures, first row (end of crisis) as an example: When all other determinants are kept constant, the lowest value for this belief (1, certainly not) matches a mean vaccination intention of 3.9 and the highest value of this belief (5, certainly yes) matches a mean vaccination intention of 4.3. Because many of the beliefs correlate strongly, and the partial dependence figures are controlled for all other determinants in the model, the partial dependence ranges are small. Partial dependence figures corresponding to all of the 10 best determinants values (1–5) are shown in the Appendix.