A Large-scale Empirical Study on Large Language Models for Election Prediction