Comparing automated text classification methods

Online social media drive the growth of unstructured text data. Many marketing applications require structuring this data at scales non-accessible to human coding, e.g., to detect communication shifts in sentiment or other researcher-defined content categories. Several methods have been proposed to automatically classify unstructured text. This paper compares the performance of ten such approaches (five lexicon-based, five machine learning algorithms) across 41 social media datasets covering major social media platforms, various sample sizes, and languages. So far, marketing research relies predominantly on support vector machines (SVM) and Linguistic Inquiry and Word Count (LIWC). Across all tasks we study, either random forest (RF) or naive Bayes (NB) performs best in terms of correctly uncovering human intuition. In particular, RF exhibits consistently high performance for three-class sentiment, NB for small samples sizes. SVM never outperform the remaining methods. All lexicon-based approaches, LIWC in particular, perform poorly compared with machine learning. In some applications, accuracies only slightly exceed chance. Since additional considerations of text classification choice are also in favor of NB and RF, our results suggest that marketing research can benefit from considering these alternatives.     

Campaign dynamics of cognitive accessibility of political judgments: measuring the impact of campaigns and campaign events using response latencies in two German rolling cross section studies

Do campaigns matter? Based on two rolling cross-section computer-assisted telephone surveys conducted in the run-up to the 2009 and 2013 German Federal Election, we test whether we can detect campaign effects on the accessibility of voters’ judgments: how do response latencies of political judgments evolve over the course of campaigns? The study uses response latencies, i.e. the standardized time it takes respondents to answer a survey question, as a proxy measurement of cognitive accessibility of political judgments. If campaigns do help voters to make up their minds, we should be able to observe changes at the implicit level of response latencies. Do people answer questions about their voting behavior and political attitudes faster as Election Day comes closer? Our results suggest that attitudes towards candidates and voting intentions become more cognitively accessible during campaigns whereas the accessibility of party identification is conditional on the contextual features of campaigns. In addition we find specific short-term effects of TV debates.     

Country-level Accounting Enforcement and IPO Underpricing

Using a sample of up to 2,503 initial public offerings (IPOs) in 32 countries from 2011–2017, we predict and find that higher levels of country-level accounting enforcement are associated with lower levels of IPO underpricing. IPOs in countries with a relatively low accounting enforcement score (second quintile) exhibit a mean underpricing of 19%, whereas the mean underpricing amounts to just 9% in countries with a relatively high score (fourth quintile). The results remain qualitatively the same when we employ a multi-level model or a difference-in-difference design. In countries that substantially strengthened their accounting enforcement in the 2003–2009 period, the level of IPO underpricing decreased significantly. We show that accounting enforcement matters for the cost of going public.     

DOES SPEED MATTER? THE ROLE OF HIGH-FREQUENCY TRADING FOR ORDER BOOK RESILIENCY

We analyze limit order book resiliency following liquidity shocks initiated by large market orders. Based on a unique data set, we investigate whether high-frequency traders are involved in replenishing the order book. Therefore, we relate the net liquidity provision of high-frequency traders, algorithmic traders, and human traders around these market impact events to order book resiliency. Although all groups of traders react, our results show that only high-frequency traders reduce the spread within the first seconds after the market impact event. Order book depth replenishment, however, takes significantly longer and is mainly accomplished by human traders’ liquidity provision.     

Managing service employees: literature review,expert opinions,and research directions

This article reviews the service management and marketing literature on managing people with a particular emphasis on managerial relevance. This review explores the market and financial results of managing people effectively, emphasizing that it is probably harder to duplicate high-performing human assets than any other corporate resource. The challenges inherent in boundary-spanning frontline jobs are discussed, including role conflict and emotional labor. Next, recommended human resources (HR) strategies and practices related to recruitment, training, empowerment, service delivery teams, and employee motivation are reviewed. The literature review concludes with a section on service culture, climate, and leadership. Each section is complemented with further research suggestions that emerged from interviews with eight academic and practitioner experts. The last section outlines six themes for new research opportunities with high potential managerial relevance; they relate to (1) the financial impact of HR practices and strategies, (2) motivating service employees, (3) training, (4) emotional labor, (5) dealing with rude customers, and (6) the impact of technology on managing service employees.     

Soziale Lage — Was sind die drängendsten Probleme?

Democracy seems to be at risk. People elect populists and autocrats, who offer simple solutions to social problems. Often, these problems are not even real ones but are only perceived as such. Economists have measures to determine which problems are fact-based, but they cannot draw a clear distinction between academic analysis and normative demand. If politics pursues a correction of a market outcome, any normative criteria other than efficiency have to be outlined explicitly. Nevertheless, some authors consider inequality one of the central contemporary social and economic challenges. Marcel Fratzscher emphasises that the solution is not found in more redistribution via taxes and transfers, but rather through greater equality of opportunity and social and educational mobility. Bert Rürup thinks that to stabilise the acceptance of the compulsory pension scheme, the dominance of the equivalence principle should be reduced. Further, the existence of different federal subsidy schemes should be replaced by a fixed federal contribution rate in alignment with the pension expenditures. In an international comparison, the fiscal burden for households with small and mid-level earned incomes is exceptionally high in Germany. In order to reduce this burden, Rürup suggests a contribution-free allowance for the social insurance contributions.     

Deindustrialisierung: Eine neue „britische Krankheit“?

In vielen EU-L?ndern ist ein Rückgang der Besch?ftigung im gewerblichen Bereich zu beobachten. Besonders ausgepr?gt ist die „Deindustrialisierung“ jedoch in Gro?britannien—bei gleichzeitiger Ausweitung des Dienstleistungsbereichs. Welches sind die Ursachen dieser Entwicklung? Ist dieser Prozess als neue „britische Krankheit“ zu bezeichnen oder Ausdruck einer relativ dynamischen und flexiblen Wirtschaft? Ist in Deutschland eine ?hnliche Entwicklung zu erwarten? Die Autoren vertreten hier ihre pers?nliche Meinung und nicht notwendigerweise die der Deutschen Bundesbank.     

Siemens: Maßhalten beim Moralisieren

Ohne Zusammenfassung Prof. Dr. Jochen Zimmermann ist Professor für Betriebswirtschaftslehre an der Universit?t Bremen