adaptation_specialization_material/code/entered.stan

functions {
    vector z_scale(vector x) {
        return (x-mean(x))/sd(x);
    }
}

data {
    int<lower=1> N;
    int<lower=1> K;
    vector<lower=0>[N] soc_cap;
    vector<lower=0>[N] soc_div;
    vector<lower=0>[N] int_div;
    vector[N] res_soc_div;
    vector<lower=0>[N] age;
    array[N] int<lower=0,upper=1> m;
    matrix<lower=0,upper=1>[N,K] x;
    vector[N] stable;
    array [N] int<lower=0,upper=K-1> primary_research_area;
}

transformed data {
    vector[N] z_soc_cap = z_scale(soc_cap);
    vector[N] z_soc_div = z_scale(soc_div);
    vector[N] z_int_div = z_scale(int_div);
    vector[N] z_res_soc_div = z_scale(res_soc_div);
    vector[N] z_age = z_scale(age);
}

parameters {
    real beta_soc_cap;
    real beta_soc_div;
    real beta_int_div;
    real beta_stable;
    real beta_age;
    vector[K] beta_x;

    real mu;
    real<lower=0> tau;
    real<lower=1> sigma;
}

model {
    vector[N] beta_research_area;
    for (k in 1:N) {
        beta_research_area[k] = tau*beta_x[primary_research_area[k]+1];
    }

    beta_soc_cap ~ normal(0, 1);
    beta_soc_div ~ normal(0, 1);
    beta_int_div ~ normal(0, 1);
    beta_x ~ double_exponential(0, 1);
    beta_stable ~ normal(0, 1);
    beta_age ~ normal(0, 1);

    mu ~ normal(0, 1);
    tau ~ exponential(1);
    sigma ~ pareto(1,1.5);

    m ~ bernoulli_logit(beta_soc_cap*z_soc_cap + beta_soc_div*z_res_soc_div + beta_int_div*z_int_div + beta_stable*stable + beta_age*z_age + beta_research_area + mu);
}

generated quantities {
    
}