Editorial

This issue of the Journal contains three interesting articlesabout central aspects of public opinion research, as well asan equal number of research notes that reflect the range oftopics that interest all public opinion researchers. Toshio Takeshita takes a turn at reconceptualizing agenda-settingstudies, using a taxonomy     

Organizational Power and the Institutional Review Board

In this essay, we respond to the narratives and essays in this issue by examining organizational power issues related to researchers and institutional research boards (IRBs). Rather than taking a singular perspective on power, we use multiple perspectives and suggest that the cumulative analyses provide the best understanding of IRBs and organizational power.     

Setting the Students’ Bearings and Balance in Appraisal

The author, a professor at the University of Toronto, touches briefly on the extensive and rich archival literature that supports the teaching of macroappraisal, but notes that this is not the only educational material she offers her students when teaching appraisal theory. She discusses the usefulness to archivists of literature from the fields of ethnography, organizational knowing, records in history, personal documentary behaviour, memory, and communications, noting that the use of texts from these fields can encourage students to reflect on their own presumptions and to develop a taste for the wide reading and research that must support appraisal.     

Analysis of Statistical Question Classification for Fact-Based Questions

Question classification systems play an important role in question answering systems and can be used in a wide range of other domains. The goal of question classification is to accurately assign labels to questions based on expected answer type. Most approaches in the past have relied on matching questions against hand-crafted rules. However, rules require laborious effort to create and often suffer from being too specific. Statistical question classification methods overcome these issues by employing machine learning techniques. We empirically show that a statistical approach is robust and achieves good performance on three diverse data sets with little or no hand tuning. Furthermore, we examine the role different syntactic and semantic features have on performance. We find that semantic features tend to increase performance more than purely syntactic features. Finally, we analyze common causes of misclassification error and provide insight into ways they may be overcome.     

Assessing the Benefits of Content Notes Enhancements for Digital Audio Resources

Our library has previously investigated the potential impact the addition of contents notes on circulation. In a subsequent study, we found contents notes were the second metadata element most heavily linked to circulation (title was first). Since there has been a marked decline in the use of print resources at our library and a corresponding increase in digital resources, we wanted to know if the addition of contents notes to records for online streaming audio files would result in a similar increase in usage. After enhancing the records, we tracked usage for a period of three months.     

Some Aspects of the Development of the All-Russian Institute for Scientific and Technical Information

Although the USSR and the United States had a comparable number of employees in the field of scientific information, there was a substantial difference in funding, which resulted in a lack of information support for Soviet scientists. While 90% of the publications in the United States were available almost immediately after their release, in the USSR they were delayed by 1.5 to 2 years. Since the publication of the Abstracts Journal by VINITI (AJ), its content was consistently expanding and reached its peak in 1990 (1.5 million documents per year). The next quarter of a century was characterized by a decline in both the composition of the AJ and the time required for its document coverage. The steps for restructuring VINITI activities are discussed. The focus is made on improving the limited coverage of the Russian-language part of the global flow of scientific information by Western information systems.     

How Persuasive Are Health Advertisements in First-Person Shooter Games? Exploring Knowledge-Activation and Thought-Disruption Mechanisms

This study examined whether persuasive health messages embedded in shooter games have broad or targeted effects on players’ willingness to engage in risk behaviors (N = 145). Participants presented with in-game health messages discouraging alcohol-impaired driving of motor vehicles showed reduced willingness to drink alcohol and to operate both motor and nonmotor vehicles, compared to those in a no-message gaming control condition. There were no spillover effects on willingness to smoke cigarettes or marijuana, thus implying targeted persuasive effects. In addition, players experiencing high instead of low cognitive load showed decreased postexperiment willingness to drink and operate nonmotor vehicles, thus suggesting that playing a game under high cognitive load can influence players’ attitudes. The findings replicate previous research and further expand on knowledge-activation and thought-disruption mechanisms underlying the persuasiveness of health messages.     

Using word embeddings in Twitter election classification

Word embeddings and convolutional neural networks (CNN) have attracted extensive attention in various classification tasks for Twitter, e.g. sentiment classification. However, the effect of the configuration used to generate the word embeddings on the classification performance has not been studied in the existing literature. In this paper, using a Twitter election classification task that aims to detect election-related tweets, we investigate the impact of the background dataset used to train the embedding models, as well as the parameters of the word embedding training process, namely the context window size, the dimensionality and the number of negative samples, on the attained classification performance. By comparing the classification results of word embedding models that have been trained using different background corpora (e.g. Wikipedia articles and Twitter microposts), we show that the background data should align with the Twitter classification dataset both in data type and time period to achieve significantly better performance compared to baselines such as SVM with TF-IDF. Moreover, by evaluating the results of word embedding models trained using various context window sizes and dimensionalities, we find that large context window and dimension sizes are preferable to improve the performance. However, the number of negative samples parameter does not significantly affect the performance of the CNN classifiers. Our experimental results also show that choosing the correct word embedding model for use with CNN leads to statistically significant improvements over various baselines such as random, SVM with TF-IDF and SVM with word embeddings. Finally, for out-of-vocabulary (OOV) words that are not available in the learned word embedding models, we show that a simple OOV strategy to randomly initialise the OOV words without any prior knowledge is sufficient to attain a good classification performance among the current OOV strategies (e.g. a random initialisation using statistics of the pre-trained word embedding models).