Multiversal methods are used for assessment of misspecification risk in the scientific research. In this study, the Relative Risk Reduction after full COVID- 19 vaccination is estimated in populations subjected to lockdown measures. While the multiverse model suggests a positive estimate of the effect of vaccination on risk reduction, apparently trivial alternative analyses in model specification reduce the estimate. Absurdly, some model specifications support a significant negative esti- mate. This is regarded as a case of Janus effect. Causes of Janus effect are identified in adoption of both Oxford Stringency Index as control variable and Negative Bi- nomial instead of Quasi Poisson model. In some literature, these alternatives are characterised as equivalent. The multiversal methodology is further proposed to enhance countermeasures of Janus effect and risk of misspecification.

Multiversal methods for model selection: COVID-19 vaccine coverage and relative risk reduction.

Tomaselli V.
Primo
;
Cantone G. G.
2022

Abstract

Multiversal methods are used for assessment of misspecification risk in the scientific research. In this study, the Relative Risk Reduction after full COVID- 19 vaccination is estimated in populations subjected to lockdown measures. While the multiverse model suggests a positive estimate of the effect of vaccination on risk reduction, apparently trivial alternative analyses in model specification reduce the estimate. Absurdly, some model specifications support a significant negative esti- mate. This is regarded as a case of Janus effect. Causes of Janus effect are identified in adoption of both Oxford Stringency Index as control variable and Negative Bi- nomial instead of Quasi Poisson model. In some literature, these alternatives are characterised as equivalent. The multiversal methodology is further proposed to enhance countermeasures of Janus effect and risk of misspecification.
9788891932310
multiverse methods, specification curve analysis, relative risk reduc- tion, COVID-19 vaccine coverage.
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.11769/533064
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact