| Acknowledgements | 第1-9页 |
| Abstract | 第9-11页 |
| 内容摘要 | 第11-13页 |
| Chapter One Introduction | 第13-19页 |
| ·Background to the study | 第13-15页 |
| ·Purpose | 第15页 |
| ·Research questions and expected goal | 第15-16页 |
| ·Organization of the thesis | 第16-19页 |
| Chapter Two Literature Review | 第19-38页 |
| ·Some Achievement in the field of language testing | 第19-22页 |
| ·Widening Scope of Inquiry | 第19页 |
| ·Nature of language ability | 第19-21页 |
| ·Affecting factors | 第21-22页 |
| ·Greater Methodological Diversity and Sophistication | 第22页 |
| ·Differing Epistemologies and Opportunities in Challenges | 第22-26页 |
| ·Some differing epistemologies | 第22-23页 |
| ·Opportunities and Challenges | 第23-24页 |
| ·Model parameter estimations | 第24-26页 |
| ·Markov Chain Monte Carlo(MCMC) | 第26-38页 |
| ·Two approaches to statistical inference | 第27-31页 |
| ·Bayesian computation | 第31-34页 |
| ·Looking to the future | 第34-38页 |
| Chapter Three Methodology | 第38-44页 |
| ·Model hypothesis | 第38-40页 |
| ·Hypothesis of relevance and independence | 第38页 |
| ·Hypothesis of homogeneity | 第38-39页 |
| ·Hypothesis of quality in the evaluation on different test approaches | 第39-40页 |
| ·Subjects | 第40-42页 |
| ·Approach subjects | 第40-41页 |
| ·Candidate subjects | 第41-42页 |
| ·Subjects of testing paper | 第42页 |
| ·Data collection and data conversion | 第42-44页 |
| ·Data collection | 第42页 |
| ·Data conversion | 第42-44页 |
| Chapter Four Model constructing and Estimation of parameters | 第44-52页 |
| ·Constructing of Stochastic Volatility Model(SV) | 第44-45页 |
| ·Observation equations | 第44页 |
| ·Hidden Markov models | 第44页 |
| ·Stochastic volatility model(SV) | 第44-45页 |
| ·Bayesian analysis of SV models | 第45-48页 |
| ·Bayesian analysis of SV models | 第45-46页 |
| ·Calculating of MCMC for SV | 第46-47页 |
| ·The Gibbs Sampling Algorithm | 第47-48页 |
| ·Implementing SV model in the WinBUGS | 第48-52页 |
| ·Directed acyclic graph(DAG) | 第48-49页 |
| ·Bayesian inference using BUGS | 第49-51页 |
| ·Diagnosis to convergence of MCMC process | 第51-52页 |
| Chapter Five Results and Analysis | 第52-67页 |
| ·MCMC simulation to the growth rate | 第52-61页 |
| ·MCMC simulation for candidate(AA1) | 第52-53页 |
| ·MCMC simulation for candidate(BA2) | 第53-54页 |
| ·MCMC simulation for candidate(AC3) | 第54-56页 |
| ·MCMC simulation for candidate(BB4) | 第56-58页 |
| ·MCMC simulation for candidate(AC5) | 第58-59页 |
| ·MCMC simulation for candidate(BC6) | 第59-61页 |
| ·MCMC simulation results | 第61-63页 |
| ·Volatility of modal volatility by audio-alone approach is higher | 第61页 |
| ·Values of persistence in the volatility are approximate | 第61页 |
| ·Values of volatility of log-volatilities are approximate | 第61-63页 |
| ·Comparison of results from SV model | 第63-64页 |
| ·Discussion | 第64-67页 |
| Chapter Six Conclusion | 第67-69页 |
| ·Findings | 第67页 |
| ·Implications | 第67页 |
| ·Limitations | 第67-69页 |
| References | 第69-75页 |
| Appendix A Samples of the listening tests | 第75-78页 |
| Appendix B Raw Scores of the Subjects in the Listening Tests | 第78-80页 |
