Michael Scharkow

The Relative Trustworthiness of Inferential Tests of the Indirect Effect in Statistical Mediation Analysis Does Method Really Matter

A content analysis of 2 years of Psychological Science articles reveals inconsistencies in how researchers make inferences about indirect effects when conducting a statistical mediation analysis. In this study, we examined the frequency with which popularly used tests disagree, whether the method an investigator uses makes a difference in the conclusion he or she will reach, and whether there is a most trustworthy test that can be recommended to balance practical and performance considerations. We found that tests agree much more frequently than they disagree, but disagreements are more common when an indirect effect exists than when it does not. We recommend the bias-corrected bootstrap confidence interval as the most trustworthy test if power is of utmost concern, although it can be slightly liberal in some circumstances. Investigators concerned about Type I errors should choose the Monte Carlo confidence interval or the distribution-of-the-product approach, which rarely disagree. The percentile bootstrap confidence interval is a good compromise test.

The Value of Big Data in Digital Media Research

This article discusses methodological aspects of Big Data analyses with regard to their applicability and usefulness in digital media research. Based on a review of a diverse selection of literature on online methodology, consequences of using Big Data at different stages of the research process are examined. We argue that researchers need to consider whether the analysis of huge quantities of data is theoretically justified, given that it may be limited in validity and scope, and that small-scale analyses of communication content or user behavior can provide equally meaningful inferences when using proper sampling, measurement, and analytical procedures.

Press CRTT to measure aggressive behavior: the unstandardized use of the competitive reaction time task in aggression research.

The competitive reaction time task (CRTT) is the measure of aggressive behavior most commonly used in laboratory research. However, the test has been criticized for issues in standardization because there are many different test procedures and at least 13 variants to calculate a score for aggressive behavior. We compared the different published analyses of the CRTT using data from 3 different studies to scrutinize whether it would yield the same results. The comparisons revealed large differences in significance levels and effect sizes between analysis procedures, suggesting that the unstandardized use and analysis of the CRTT have substantial impacts on the results obtained, as well as their interpretations. Based on the outcome of our comparisons, we provide suggestions on how to address some of the issues associated with the CRTT, as well as a guideline for researchers studying aggressive behavior in the laboratory.

Militaristic Attitudes and the Use of Digital Games

From their beginnings until today, digital games have been a substantial part of what has been labeled the “military–entertainment complex” deeply imbued with militaristic messages and imagery. Within cultivation research, this enhanced exposure to war and militarism is supposed to be associated with the adoption of military norms and thinking. Concepts on narrative persuasion specify this relationship between certain narratives and the adoption of story-inherent beliefs. Based on these theoretical concepts, the present study tries to investigate the relationship between aspects of gaming and militaristic attitudes. We carried out a representative survey of 4,500 gamers with an added control group of 500 nongamers. Militaristic attitudes were measured using a newly developed multidimensional militarism scale. Structural equation modeling did not reveal any relationship between gaming and militaristic attitudes. Moreover, neither the gaming type (multiplayer vs. single) nor the gaming frequency or a preference for shooter games was significantly related to militarism.

Recycled media: An automated evaluation of news outlets in the twenty-first century

Whilst new forms of presentation and user participation make a “way new journalism” possible, researchers around the world critically claim that news websites do not live up to their potential. To offer an explanation for the slow adaption of hyperlinks, multimedia, and interactive options, we examine if there is a path dependence between news websites’ online sophistication and their structural background. Based on a large-scale automated content analysis, we analyzed 65,000 news articles from nine countries. The results reveal that institutional factors (media system, nation, media type) are more relevant for predicting websites’ online sophistication than content-related factors. This suggests that, apart from a few remarkable exceptions, there is a path dependence on a news website’s media background.

Mutual friends’ social support and self-disclosure in face-to-face and instant messenger communication

ABSTRACT In the present study, we investigated long-term effects of self-disclosure on social support in face-to-face and instant messenger (IM) communication between mutual friends. Using a representative sample of 583 German IM users, we explored whether self-disclosure and positive experiences with regard to social support would dynamically interact in the form of a reinforcing spiral across three measurement occasions. If mutual friends self-disclose today, will they receive more social support 6 months later? In turn, will this affect their willingness to self-disclose another 6 months later? We further analyzed spill-over effects from face-to-face to IM communication and vice versa. We found that self-disclosure predicted social support and vice versa in IM communication, but not in face-to-face communication. In light of these results, the impact of IM communication on how individuals maneuver friendships through the interplay between self-disclosure and social support are discussed.

Qualitative Textanalyse. Blaupause und Potenziale (teil-)automatisierter Verfahren

Dieser Beitrag diskutiert den Einsatz textstatistischer und automatisierter inhaltsanalytischer Verfahren im Rahmen der qualitativen Datenauswertung. Im Beitrag werden zunachst Analyseschritte und -ziele qualitativer Forschung identifiziert. Hierbei wurden mehr und weniger standardisierte und sowohl deduktive als auch induktive Verfahren berucksichtigt. Darauf aufbauend wird gezeigt, wo und unter welchen Bedingungen textstatistische und (teilweise) automatisierte Verfahren zur Auswertung qualitativer Daten eingesetzt werden konnen und welche Potenziale und Einschrankungen mit ihrem Einsatz verbunden sind. Hierbei liegt der Fokus auf der Auswertung von Interviewtranskripten.

Correcting measurement error in content analysis

ABSTRACT Conducting and reporting reliability tests has become a standard practice in content analytical research. However, the consequences of measurement error in coding data are rarely discussed or taken into consideration in subsequent analyses. In this article, we demonstrate how misclassification in content analysis leads to biased estimates and introduce matrix back-calculation as a simple remedy. Using Monte Carlo simulation, we investigate how different ways of collecting information about the misclassification process influence the effectiveness of error correction under varying conditions. The results show that error correction with an adequate set-up can often substantially reduce bias. We conclude with an illustrative example, extensions of the procedure, and some recommendations.

Messung der Publikumsagenda mittels Nutzungsstatistiken von Suchmaschinenanfragen

Im Jahr 1970 – zwei Jahre bevor die „Chapel Hill“-Studie (vgl. McCombs & Shaw 1972) veroffentlicht wurde – lag in Deutschland die Reichweite des Horfunks bei 67 Prozent, die der Tageszeitung bei 70 Prozent und die des Fernsehens bei 72 Prozent (van Eimeren & Ridder 2005, 495). In der Ara des Massenpublikums dominierten das Fernsehen, das Radio und die Tageszeitung die Medienagenda und bestimmten im Zusammenspiel mit interpersonaler Kommunikation die Publikumsagenda. Inzwischen zeichnet sich das Ende der Ara des Massenpublikums ab. Immer mehr Burger bewegen sich in einer „multiaxialen Informationsumwelt“ (vgl. ausfuhrlich Delli Carpini & Williams 2001, 173), in der die klassischen Massenmedien nach und nach an Bedeutung verlieren. Laut der ARD/ZDF-Langzeitstudie Massenkommunikation 2010 nutzt in Deutschland die Generation der digital natives (14- bis 29-Jahrige) das Internet bereits an einem durchschnittlichen Tag genau so lange wie das Radio und das Fernsehen (vgl. Ridder & Engel 2010, 527). 41 Prozent der digital natives lesen mittels Internet aktuelle Nachrichten; in der Gesamtbevolkerung betragt dieser Anteil nur 22 Prozent (vgl. Ridder & Engel 2010, 531).