Dear all,
I am working with the following dataset, stacked by id and domain
I am estimating several mixed-effects regression models with random intercepts at the country level, for each domain.
My dependent variable is voted_party_repgap, which measures the distance between each individual voter and her voted party on several domains. That is, this variable is created by matching data derived from two different sources (party-level and individual-level). Additionally, some of the IVs are party-level, some are party-system level, and some are individual-level.
My main IV is voted_party_blur100. Since it could be argued the case for reverse causality, I would like to address this possibility. One way (I guess) to do so is to lag my main IV (at time t) and regress my DV at time t+1 on it. The methodological issues I see are several, however. Most importantly, my DV is generated my matching party-level data that are panel data, with individual level-data that are taken from repeated cross-sections. That is, I have different individuals for each wave of the election study. Additionally, if I follow this path, shall I lag also the other variables in the model? Again, if so, I would encounter similar issues as above with individual-level predictors such as pol_sophistication, feelclosest_is_closest, and perform_gov, which come from different individuals.
What is the best way to deal with reverse causality in such a case? What is the most parsimonious solution? Are there alternatives?
Sorry in advance if the question will come out as confusing.
Sincerely
Mattia
I am working with the following dataset, stacked by id and domain
Code:
* Example generated by -dataex-. For more info, type help dataex clear input str26 countryname float id long domain float(dataset pol_sophistication) byte perform_gov float(my_issue voted_party_pervote voted_party_extremism voted_party_blur100 voted_party_repgap voted_party_age_party polarization n_parties) "Austria" 31032 2 2019 8.75 1 0 3.8 1.0375171 1.4077452 4.037517 32 4.217497 4.085089 "Austria" 30699 2 2019 5 2 0 3.8 1.0375171 1.4077452 3.037517 32 4.217497 4.085089 "Austria" 822 2 2014 8.823529 2 0 23.991 .9172006 18.314589 2.0827994 65 5.33847 5.189692 "Austria" 30636 2 2019 9.375 1 0 25.966 3.363714 56.53221 1.6362858 69 4.217497 4.085089 "Austria" 30266 2 2019 3.75 1 . 31.47 3.320255 55.51868 3.320255 69 4.217497 4.085089 "Austria" 30454 2 2019 9.375 1 0 31.47 3.320255 55.51868 2.3202553 69 4.217497 4.085089 "Austria" 30708 2 2019 5 2 0 26.859 1.2898846 1.5980145 3.2898846 69 4.217497 4.085089 "Austria" 275 2 2014 8.235294 1 . 23.991 .9172006 18.314589 3.9172006 65 5.33847 5.189692 "Austria" 30324 2 2019 8.75 . 0 5.296 1.5868378 13.084457 2.586838 5 4.217497 4.085089 "Austria" 677 2 2014 9.411765 1 . 23.991 .9172006 18.314589 1.0827994 65 5.33847 5.189692 "Austria" 30567 2 2019 4.375 2 0 25.966 3.363714 56.53221 3.363714 69 4.217497 4.085089 "Austria" 30753 2 2019 3.125 1 0 31.47 3.320255 55.51868 3.320255 69 4.217497 4.085089 "Austria" 30310 2 2019 6.25 1 0 25.966 3.363714 56.53221 6.363714 69 4.217497 4.085089 "Austria" 31050 2 2019 1.875 1 0 25.966 3.363714 56.53221 4.363714 69 4.217497 4.085089 "Austria" 336 2 2014 5.294117 1 . 26.819 3.550965 0 2.550965 65 5.33847 5.189692 "Austria" 301 2 2014 5.294117 2 1 20.506 2.83029 46.62173 2.83029 65 5.33847 5.189692 "Austria" 538 2 2014 10 2 0 23.991 .9172006 18.314589 .9172006 65 5.33847 5.189692 "Austria" 30695 2 2019 6.25 2 1 31.47 3.320255 55.51868 3.320255 69 4.217497 4.085089 "Austria" 30174 2 2019 8.125 2 0 31.47 3.320255 55.51868 6.320255 69 4.217497 4.085089 "Austria" 30775 2 2019 8.125 1 0 26.859 1.2898846 1.5980145 .28988457 69 4.217497 4.085089 "Austria" 576 2 2014 7.647059 1 1 12.416 1.5764472 3.320789 1.5764472 28 5.33847 5.189692 "Austria" 30938 2 2019 10 2 0 26.859 1.2898846 1.5980145 1.7101154 69 4.217497 4.085089 "Austria" 30259 2 2019 4.375 2 0 26.859 1.2898846 1.5980145 3.7101154 69 4.217497 4.085089 "Austria" 30895 2 2019 1.25 2 0 5.296 1.5868378 13.084457 4.586838 5 4.217497 4.085089 "Austria" 30456 2 2019 1.875 . 0 31.47 3.320255 55.51868 3.320255 69 4.217497 4.085089 "Austria" 30441 2 2019 8.75 1 0 26.859 1.2898846 1.5980145 3.7101154 69 4.217497 4.085089 "Austria" 30808 2 2019 3.125 2 0 25.966 3.363714 56.53221 1.3637142 69 4.217497 4.085089 "Austria" 30382 2 2019 3.75 2 0 31.47 3.320255 55.51868 4.3202553 69 4.217497 4.085089 "Austria" 687 2 2014 8.823529 2 1 4.964 2.581272 42.51263 .41872835 1 5.33847 5.189692 "Austria" 30411 2 2019 5 1 0 25.966 3.363714 56.53221 1.6362858 69 4.217497 4.085089 "Austria" 30936 2 2019 6.25 2 0 26.859 1.2898846 1.5980145 3.7101154 69 4.217497 4.085089 "Austria" 30995 2 2019 4.375 2 0 26.859 1.2898846 1.5980145 2.7101154 69 4.217497 4.085089 "Austria" 77 2 2014 8.235294 1 . 23.991 .9172006 18.314589 2.0827994 65 5.33847 5.189692 "Austria" 30318 2 2019 5 2 1 25.966 3.363714 56.53221 7.363714 69 4.217497 4.085089 "Austria" 17 2 2014 4.7058825 1 . 26.819 3.550965 0 .4490347 65 5.33847 5.189692 "Austria" 30225 2 2019 5 2 0 26.859 1.2898846 1.5980145 1.7101154 69 4.217497 4.085089 "Austria" 1001 2 2014 5.294117 1 . 20.506 2.83029 46.62173 2.83029 65 5.33847 5.189692 "Austria" 30303 2 2019 2.5 1 0 31.47 3.320255 55.51868 3.320255 69 4.217497 4.085089 "Austria" 574 2 2014 3.529412 . 0 20.506 2.83029 46.62173 2.83029 65 5.33847 5.189692 "Austria" 30704 2 2019 .625 1 1 26.859 1.2898846 1.5980145 6.289885 69 4.217497 4.085089 "Austria" 197 2 2014 8.823529 2 1 23.991 .9172006 18.314589 1.0827994 65 5.33847 5.189692 "Austria" 30459 2 2019 8.125 1 0 25.966 3.363714 56.53221 1.3637142 69 4.217497 4.085089 "Austria" 163 2 2014 8.823529 1 . 26.819 3.550965 0 1.4490347 65 5.33847 5.189692 "Austria" 553 2 2014 3.529412 2 1 20.506 2.83029 46.62173 7.83029 65 5.33847 5.189692 "Austria" 588 2 2014 10 2 0 26.819 3.550965 0 1.5509653 65 5.33847 5.189692 "Austria" 30928 2 2019 8.125 1 0 25.966 3.363714 56.53221 2.3637142 69 4.217497 4.085089 "Austria" 236 2 2014 8.235294 1 1 26.819 3.550965 0 2.550965 65 5.33847 5.189692 "Austria" 30822 2 2019 7.5 1 0 26.859 1.2898846 1.5980145 .28988457 69 4.217497 4.085089 "Austria" 30467 2 2019 6.25 2 0 26.859 1.2898846 1.5980145 .28988457 69 4.217497 4.085089 "Austria" 944 2 2014 7.058824 2 . 20.506 2.83029 46.62173 5.83029 65 5.33847 5.189692 "Austria" 722 2 2014 8.235294 2 0 20.506 2.83029 46.62173 5.83029 65 5.33847 5.189692 "Austria" 804 2 2014 8.823529 2 1 23.991 .9172006 18.314589 5.917201 65 5.33847 5.189692 "Austria" 30687 2 2019 8.125 1 0 25.966 3.363714 56.53221 1.3637142 69 4.217497 4.085089 "Austria" 30291 2 2019 9.375 1 0 25.966 3.363714 56.53221 4.363714 69 4.217497 4.085089 "Austria" 30274 2 2019 1.25 2 0 25.966 3.363714 56.53221 7.363714 69 4.217497 4.085089 "Austria" 319 2 2014 4.7058825 2 . 12.416 1.5764472 3.320789 .42355275 28 5.33847 5.189692 "Austria" 30835 2 2019 9.375 1 0 25.966 3.363714 56.53221 3.363714 69 4.217497 4.085089 "Austria" 30780 2 2019 5.625 1 0 25.966 3.363714 56.53221 .3637142 69 4.217497 4.085089 "Austria" 485 2 2014 6.470588 2 0 12.416 1.5764472 3.320789 .42355275 28 5.33847 5.189692 "Austria" 253 2 2014 7.058824 2 0 23.991 .9172006 18.314589 .9172006 65 5.33847 5.189692 "Austria" 686 2 2014 2.3529413 2 1 12.416 1.5764472 3.320789 .42355275 28 5.33847 5.189692 "Austria" 30434 2 2019 2.5 2 0 25.966 3.363714 56.53221 3.363714 69 4.217497 4.085089 "Austria" 30437 2 2019 9.375 1 0 31.47 3.320255 55.51868 .3202553 69 4.217497 4.085089 "Austria" 568 2 2014 5.294117 2 . 26.819 3.550965 0 .5509653 65 5.33847 5.189692 "Austria" 30660 2 2019 7.5 . 0 31.47 3.320255 55.51868 8.320255 69 4.217497 4.085089 "Austria" 30934 2 2019 3.125 1 1 26.859 1.2898846 1.5980145 4.2898846 69 4.217497 4.085089 "Austria" 30351 2 2019 3.75 . 0 25.966 3.363714 56.53221 .3637142 69 4.217497 4.085089 "Austria" 30081 2 2019 7.5 2 0 3.8 1.0375171 1.4077452 3.962483 32 4.217497 4.085089 "Austria" 187 2 2014 9.411765 1 1 23.991 .9172006 18.314589 .9172006 65 5.33847 5.189692 "Austria" 30908 2 2019 9.375 1 1 26.859 1.2898846 1.5980145 .7101154 69 4.217497 4.085089 "Austria" 30962 2 2019 5 2 0 31.47 3.320255 55.51868 2.3202553 69 4.217497 4.085089 "Austria" 589 2 2014 8.823529 2 0 23.991 .9172006 18.314589 5.917201 65 5.33847 5.189692 "Austria" 30496 2 2019 8.75 . 0 25.966 3.363714 56.53221 8.363714 69 4.217497 4.085089 "Austria" 68 2 2014 9.411765 1 0 26.819 3.550965 0 2.550965 65 5.33847 5.189692 "Austria" 902 2 2014 7.647059 1 0 23.991 .9172006 18.314589 3.9172006 65 5.33847 5.189692 "Austria" 30563 2 2019 3.75 1 0 25.966 3.363714 56.53221 1.3637142 69 4.217497 4.085089 "Austria" 900 2 2014 4.7058825 2 1 20.506 2.83029 46.62173 5.83029 65 5.33847 5.189692 "Austria" 30101 2 2019 3.75 1 0 31.47 3.320255 55.51868 4.3202553 69 4.217497 4.085089 "Austria" 30536 2 2019 5.625 2 0 31.47 3.320255 55.51868 2.3202553 69 4.217497 4.085089 "Austria" 30833 2 2019 4.375 1 0 26.859 1.2898846 1.5980145 1.7101154 69 4.217497 4.085089 "Austria" 30940 2 2019 5.625 1 0 31.47 3.320255 55.51868 6.320255 69 4.217497 4.085089 "Austria" 30698 2 2019 4.375 . 0 25.966 3.363714 56.53221 1.3637142 69 4.217497 4.085089 "Austria" 169 2 2014 7.058824 1 . 26.819 3.550965 0 2.550965 65 5.33847 5.189692 "Austria" 1012 2 2014 8.235294 1 . 26.819 3.550965 0 1.4490347 65 5.33847 5.189692 "Austria" 30308 2 2019 5 1 1 31.47 3.320255 55.51868 .6797447 69 4.217497 4.085089 "Austria" 30914 2 2019 8.125 1 0 26.859 1.2898846 1.5980145 3.7101154 69 4.217497 4.085089 "Austria" 30573 2 2019 2.5 1 0 25.966 3.363714 56.53221 4.363714 69 4.217497 4.085089 "Austria" 501 2 2014 5.882353 1 1 12.416 1.5764472 3.320789 2.423553 28 5.33847 5.189692 "Austria" 581 2 2014 7.058824 2 . 23.991 .9172006 18.314589 1.0827994 65 5.33847 5.189692 "Austria" 30430 2 2019 3.125 . 0 25.966 3.363714 56.53221 8.363714 69 4.217497 4.085089 "Austria" 30285 2 2019 5.625 2 0 26.859 1.2898846 1.5980145 4.2898846 69 4.217497 4.085089 "Austria" 928 2 2014 2.9411764 1 1 23.991 .9172006 18.314589 4.0827994 65 5.33847 5.189692 "Austria" 203 2 2014 5.882353 . 0 26.819 3.550965 0 7.550965 65 5.33847 5.189692 "Austria" 848 2 2014 8.823529 1 . 23.991 .9172006 18.314589 4.917201 65 5.33847 5.189692 "Austria" 30112 2 2019 8.125 1 0 31.47 3.320255 55.51868 .3202553 69 4.217497 4.085089 "Austria" 648 2 2014 7.058824 2 . 26.819 3.550965 0 5.550965 65 5.33847 5.189692 "Austria" 156 2 2014 9.411765 2 . 12.416 1.5764472 3.320789 1.4235528 28 5.33847 5.189692 "Austria" 30972 2 2019 9.375 . 0 31.47 3.320255 55.51868 3.320255 69 4.217497 4.085089 "Austria" 803 2 2014 5.294117 2 1 23.991 .9172006 18.314589 2.0827994 65 5.33847 5.189692 "Austria" 31009 2 2019 2.5 1 0 25.966 3.363714 56.53221 7.363714 69 4.217497 4.085089 end label values domain domain3 label def domain3 2 "Redistribution", modify label values perform_gov QPP20_1 label def QPP20_1 1 "Approve", modify label def QPP20_1 2 "Disapprove", modify
Code:
mixed voted_party_repgap voted_party_blur100 pol_sophistication feelclosest_is_closest perform_gov voted_party_pervote voted_party_extremism voted_party_age_party polarization n_parties dataset if domain==2 || country:
My main IV is voted_party_blur100. Since it could be argued the case for reverse causality, I would like to address this possibility. One way (I guess) to do so is to lag my main IV (at time t) and regress my DV at time t+1 on it. The methodological issues I see are several, however. Most importantly, my DV is generated my matching party-level data that are panel data, with individual level-data that are taken from repeated cross-sections. That is, I have different individuals for each wave of the election study. Additionally, if I follow this path, shall I lag also the other variables in the model? Again, if so, I would encounter similar issues as above with individual-level predictors such as pol_sophistication, feelclosest_is_closest, and perform_gov, which come from different individuals.
What is the best way to deal with reverse causality in such a case? What is the most parsimonious solution? Are there alternatives?
Sorry in advance if the question will come out as confusing.
Sincerely
Mattia
Comment