Identifying Stochastic Volatility Of Audio-alone And Audio-visual Approach In English Listening Test

English listening plays an important role in the college English listening test,with a widespread and profound influence.Studying the authenticity of the current college English listening test is not only of great academic significance,but of more sociological significance.Approach of audio-visual integration test(multimedia) is frequently provided with more qualities than Audio-alone testing approach(tape method).The audio-visual integration test has attracted the majority of domestic and foreign scholars,testing institutions and others.Relevant evidence however has not been found.To qualitatively investigate that a kind of testing approach is superior to another, designing the approaches as the experimental treatment(variable) is usually deployed.In the framework of a larger sample,sampling is conducted for the purpose of statistical comparison,which leads easily to difficulties in the actual operation and Statistics:1) appropriateness and reasonableness to isolate the variable,that is,it is really hard to determine whether the difference is caused by the approach or by the impact of other relevant factors;2) sampling characteristics;3) predictability of results.For the above reasons,current studies on the language listening tests were ultimately based on the assumption of the invariable variance,which is not frequently reliable.In the present study,simulations of Listening of the College English Test Band 6 in different groups were performed in consecutive 18 weekly-listening tests,and a stochastic volatility model(SV) from listening tests with audio-alone and audio-visual approaches was developed,in which parameters were estimated with the method of Markov Chain Monte Carlo(MCMC).Two of the three parameters in the SV,which are the persistence in the volatility and the volatility of log-volatilities are approximately equivalent in the two kinds of test approaches.These suggested that the proto-phase volatilities impact on the succeeding and their external noise made no difference.It demonstrates the quality of credibility,structure-effect and authenticity from Bachman's empirical language testing principles.The values of the modal volatility of audio-alone approach,however,were significantly higher than that audio-visual approach of by 72.06 percent,which means that audio-alone approach is more vulnerable to volatility than audio-visual approach, indicating that the audio-visual approach is better connected to interaction,aftereffect and applicability from the Bachman principles.