simardcasanova’s avatarsimardcasanova’s Twitter Archive—№ 26,284

  1. …in reply to @Patrick_M_Econ
    @Patrick_M_Econ @colette_salemi @TimonScheuer @gdosi1 @AndreaRoventini @giorgiofagiolo Reproducibility of computer simulations requires the code to be publicly available and *also* to specify which software was used to run the code in question Both things are usually extremely lacking in ABM literature
    simardcasanova’s avatar Mood 0
    1. …in reply to @simardcasanova
      @Patrick_M_Econ @colette_salemi @TimonScheuer @gdosi1 @AndreaRoventini @giorgiofagiolo My initial point was not to say that ABM are fundamentally impossible to replicate (I don’t see why it would be the case) It’s rather that the norms in the research community around their replicability are currently not good/strong enough
      Permalink simardcasanova’s avatar Mood -1 🙁
      1. …in reply to @simardcasanova
        @Patrick_M_Econ @colette_salemi @TimonScheuer @gdosi1 @AndreaRoventini @giorgiofagiolo If I had to summarise my criticism in one argument, it would be that ABMs in economics usually lack documentation. In too many instance we don’t even know which software was used!
        Permalink simardcasanova’s avatar Mood -4 🙁