Closed-ended question

Abstract: Two quite different reasons for using open-ended as opposed to closeended questions can be distinguished. One is to discover the responses that individuals give spontaneously; the other is to avoid the bias that may resultfrom suggesting responses to individuals. However, open-ended questions also have disadvantages in comparison to close-ended, such as the need for extensive coding and larger item non-response. While this issue has already been well researched for traditional survey questionnaires, not much research has been devoted to it in recently used Web questionnaires. We therefore examine the differences between the open-ended and the closeended question form in Web questionnaires by means of experiments within the large-scale RIS 2001 Web survey. The question "What is the most important, critical problem the Internet is facing today?" was asked in an open-ended and two close-ended question forms in a split-ballot experiment. The results show that there were differences between question forms in univariate distributions, though no significant differences were found in the ranking of values. Close-ended questions in general yield higher percentages than open-ended question for answers that are identical in both question forms. It seems that respondents restricted themselves with apparent ease to the alternatives offered on the close-ended forms, whereas on the open-ended question they produced a much more diverse set of answers. In addition, our results suggest that open-ended questions produce more missing data than close-ended. Moreover, there were more inadequate answers for open-ended question. This suggests that open-endedquestions should be more explicit in their wording (at least for Websurveys, as a self administered mode of data collection) than close-ended questions, which are more specified with given response alternatives.

Abstract: This paper presents how to generate questions from given passages using neural networks, where large scale QA pairs are automatically crawled and processed from Community-QA website, and used as training data The contribution of the paper is 2-fold: First, two types of question generation approaches are proposed, one is a retrieval-based method using convolution neural network (CNN), the other is a generation-based method using recurrent neural network (RNN); Second, we show how to leverage the generated questions to improve existing question answering systems We evaluate our question generation method for the answer sentence selection task on three benchmark datasets, including SQuAD, MS MARCO, and WikiQA Experimental results show that, by using generated questions as an extra signal, significant QA improvement can be achieved