Hypothesis Testing for Dependent Variables with Unbalanced Data // Implements hierarchical Bayesian inference, robust frequentist inference, and distribution-free inference for dependent and unbalanced data under strong-mixing conditions. Supports triangular-array, weighted-sum and mixingale convergence regimes with Whittle and composite likelihoods, heteroskedasticity-and-autocorrelation-consistent variance estimation, block bootstrap with automatic block length, fixed- bandwidth HAR inference, adaptive conformal prediction, Bayesian decision under Region of Practical Equivalence, bridge-sampling Bayes factors, and predictive comparison via the Widely Applicable Information Criterion and leave-future- out cross-validation. Methods follow Andrews (1991) doi:10.2307/2938229, Kiefer and Vogelsang (2005) doi:10.1017/S0266466605050565, Patton, Politis and White (2009) doi:10.1080/07474930802459016, Vehtari, Gelman and Gabry (2017) doi:10.1007/s11222-016-9696-4, Kruschke (2018) doi:10.1177/2515245918771304, and Gibbs and Candes (2021) doi:10.48550/arXiv.2106.00170.