Scientific articles as media products: Linguistic changes in research articles from the field of cross-cultural psychology (1970-2014) indicate a trend towards academic capitalism
1Leibniz Insitut für Wissensmedien Tübingen, Germany; 2Otto-von-Guericke-Universität Magdeburg; 3University of Helsinki
Media products are often defined in the context of media management as artifacts, for example, in form of texts, videos, or audio recordings, which are created within a given context of production to cater the needs of a certain target audience. Such media products compete with other products for publicity and market shares (e.g., Gläser, 2014). It is somewhat unusual to think of scientific publications as media products, because - in an ‘ideal (positivistic) world’ - knowledge interests and the ‘scientific method’ alone should guide the whole research process. Still, the production context of scientific publications has changed dramatically over the last decades. In particular, the number of scientific publications has increased exponentially as a consequence of researchers and academic institutions having to compete increasingly for scarce resources such as external funding, students, and tenured positions (e.g., Sarewitz, 2016). Münch (2014) calls this a trend towards ‘academic capitalism’. We will analyze in our paper if and how these contextual changes have led to linguistic changes in scientific articles.
Several recent empirical studies have investigated developments in academic language over the last four decades: An analysis of a large number of PubMed abstracts (Vinkers, Tijdink, & Otte, 2015) showed a dramatic rise in words with positive connotations such as 'novel', 'robust', 'innovative', and 'unprecedented'. This seems to indicate increasing self-marketing tendencies. In the field of psychology, Pritschet, Powell, and Horne (2016) found an increase in the reporting of ‘marginally significant’ statistical findings. This may indicate a tendency to present even rather weak empirical evidence as significant - as long as it supports a researcher's hypothesis and is helpful in getting research results published. Michael Billig (2013) raised concerns about an increase of technical jargon in the social sciences over the last decades. The abundance of scientific publications seems to facilitate the emergence of scientific sub-communities, which set themselves apart from each other mostly through a specific vocabulary. Furthermore, 'big' and 'technical' words are according to Billig sometimes used to mask theoretical and methodological inconsistencies and weaknesses. All these findings can be interpreted as a consequence of increasing publication pressure and the need to become ‘visible’ as a researcher or ‘vanish’.
In our study, we analyzed the aforementioned trends based on a sample of 1680 research articles from the Journal of Cross-Cultural Psychology (JCCP) and 657 research articles from Cross-Cultural Research (CCR), all published between 1970 and 2014. Using the LIWC and the AntConc software with a variety of custom and self-created dictionaries, we found a consistent increase in positive framing, an increase in reports of ‘marginally significant’ statistical findings, as well as indicators for an increase in technical jargon and 'big words'. These findings indicate that self-marketing strategies seem to be also on the rise in cross-cultural psychology, although the increase in positive framing seems to be less pronounced than the one Vinkers and colleagues (2015) had found for PubMed abstracts over the same time period.
An experimental paradigm of measuring open-earedness
1Technische Universität Braunschweig, Germany; 2SHR Hochschule der populären Künste, Berlin, Germany
In music psychology open-earedness is defined as an open behavior towards different kinds of music. Following the hypotheses by Hargreaves (1982) many studies were conducted in the last 15 years in Germany (Auhagen et al., 2014). The current state of research includes two general paradigms: a) the valence-oriented paradigm and b) the decision-making paradigm. The first measures open-earedness by the situational valence for specific experimenter-selected music, which is often subjectively categorized in conventional and unconventional music (e.g. Gembris & Schellberg, 2007; Kopiez & Lehmann, 2008; Louven, 2014). This paradigm shows theoretical deficits and therefore problems of operationalizing of open-eardness (von Georgi & Frieler, 2014; Bötsch & Rothmann, 2016). Until now a validation of valence as a predictor for open earedness is still missing.
To overcome the existing problems, von Georgi & Frieler (2014) developed an experimental design to measure open-earedness independently of valence and music categories, which we call the decision-making paradigm. This paradigm measured open-earedness by the variation in decisions for different categories of music. Therefore open-earedness correlates with Openness to Experience and Experience Seeking. This decision-making paradigm could be replicated by a sample of adolescence (Bötsch & Rothmann, 2016) so far.
Aim of this study is the replication and validation of the decision-making. Furthermore we are interested in the development over lifetime by combining the samples of existing studies (von Georgi & Frieler, 2014, Bötsch & Rothmann 2016 and the following samples). Another aim is to identify other psychological constructs which may predict open-earedness.
The sample consist of 242 subjects, 51.3 percent are females, with an average age of 30.87 (Md=25; SD=17.94; min=12; max=83). All subjects were asked to choose a piece of music considering their own individual preferences and interests out of a pool of music-categories (genres). Every category had sub-categories, in which each subject find different pieces of music. After each listening they conducted the SAM (Lang, 1980). After 15 times of decision-making and listening the subjects conducted the NEO-FFI (Borkenau & Ostendorf, 1993) and the Sensations Seeking ScaleV (Beauducel, Strobel & Brocke, 2003) as well as some question about demographics, music preferences etc. For analyses the Entropy (Shannon, 1948) and the Oewd, another coefficient of frequency, were computed as described by von Georgi & Frieler (2014). Hypotheses and explorative questions were tested by correlation and regression analyses. The level of significance was set at α≤.05.
The results of the new samples are consistent with the hypotheses. We found significant correlations between Entropy and Oewd. Both coefficients correlate with Openness to Experience. No correlation with age could be found. The variation in general music preferences as an indicator of multiple music preferences seems to be a predictor of open-earedness.
The decisions-making paradigm offers an experimental design for measuring open-earedness independently of valence and different music categories and therefore the possibility of validation. Because of the stimulus-independency, this design could be used a broader range of an openness to aesthetic experiences. Probably open-earedness is part of a meta-construct “openness to aesthetic experiences”.
Extraction of low-level video features for observation and content analysis using OpenCV
University of Würzburg, Germany
Applying methods of content analysis and observation can yield results far beyond questionnaire studies, but can also prove to be tedious and time consuming. While complex coding tasks are--at least to date--still largely a domain of human coders, very basic stimulus analyses can be conducted objectively and more economically by using software tools. We present methods and applications for two such basic analyses of video material, frame differencing and optical flow, as well as case studies for example applications. Following the open science idea, these tasks are entirely realized using free, open software.
Geise, Rössler and Kruschinski (2016) give a detailed overview of recent developments in automated image content analysis, and also offer a classification scheme for the possible steps of analysis. In their terms, the present approach could be classified as image processing for extraction of low-level features.
In the present case, we use the tool box OpenCV (Bradski, 2000) to analyze and objectively describe videos and video games as media-psychological stimulus material, with a special focus on the attention-related stimulus property of movement. The OpenCV toolbox provides a broad assortment of algorithms for image analysis, video analysis, and computer vision. The respective tools can be accessed with help of simple programs written in e.g. Java or Python. Two applications of the OpenCV tools are used here, namely calculation of frame differences in videos, and calculation of optical flow in videos. Three case studies will illustrate the respective implementation and how the methods integrate into the scientific workflow.
Frame differencing yields a measure for depictions of movement in videos by comparing luminance changes in subsequent video frames. Videos are pre-processed frame by frame, so finally each frame is represented as a matrix containing the luminance value of each pixel. The differences between all of the frames’ summed up luminance values then form a time series of frame differences. In case study 1, frame differencing was used to economically identify transitions between shots in a TV show, as well as the degree of movement within shots. Other applications include an economical measure of e.g. degree of participant movement in an experimental setting, calculated from readily available video recordings.
While the first measure aimed at calculating the total amount of visual information change in videos, the second measure aims at calculating the directional change of the visual field, the so called optical flow. OpenCV offers several computational methods for calculation of optical flow. In this case, the method for dense optical flow proposed by Farneback (2003) was used to quantify the two-dimensional projection of movement through a three-dimensional space: In cinema, movement of the camera is reproduced as a movie stimulus (study 2), and in virtual environments of video games, movement through the game environment is reproduced on the player’s screen (study 3).
Further details will be provided on methodological considerations with a focus on the needs of media-psychological theorizing and study design. Online supplementary material will be available along with the presentation.
We are slowly getting there: Towards Continuous Assessment of Presence in Virtual Environments
1Institute for Media Research, Chemnitz University of Technology, Germany; 2Institute Human-Computer-Media, Julius-Maximilian-University Wuerzburg, Germany
The phenomenon of non-mediation is an important factor in virtual experiences. Users seem to ignore their physical surroundings and become present in the virtual environment (VE). Assessing presence in VEs with continuous measures is an ongoing effort both to complement the data provided by self-report measures and to identify specific events during virtual experiences that disrupt presence (“breaks in presence”; BIPs, Slater & Steed, 2002). Early studies employed self-report measures for BIPs, but have also identified specific psychophysiological reactions bearing the potential to develop a continuous measure—the nature of these reactions, however, has been unclear. We argue that BIPs can be considered as orienting responses (ORs; Sokolov, 1963) as users interact with VEs based on dedicated mental models and direct all their attention to virtual stimuli. When a new virtual stimulus is incompatible with these mental models, an OR is triggered, directing user attention at the source of the stimulus to determine its cause. ORs have been linked to psychophysiological reactions (Bradley, 2009) similar to the ones reported for BIPs.
We designed a series of three studies to investigate this account to BIPs. Two studies are already complete, while a third study is currently being conducted. In all studies, users played Skyrim (Bethesda, 2011) modified to include stimuli that can be considered as BIPs not belonging to the VE. We measured psychophysiological responses using ECG and GSR. In study 1 (N = 31) participants played for M = 82.08 minutes (SD = 34.23) and encountered ten different types of BIP stimuli (e.g. error message, game crash) determined empirically through a prestudy. In line with previous literature on ORs, we found significant responses in ECG and GSR for all BIPs with a clearly defined stimulus onset. Additionally, we found a habituation effect of repeated exposure to the error message consistent with habituation of orienting. Analysis of blinking activity showed prolonged intervals during BIPs consistent with the information acquisition function of ORs. In study 2 (N = 69) participants played for M = 22.02 minutes (SD = 7.1). We employed a modified secondary-task reaction time (STRT) procedure in addition to psychophysiological measurements to compare ORs toward a game-related stimulus with ORs towards a BIP-stimulus. By pressing a footswitch, participants had to react to a crying baby, embedded into the game’s narrative, or a similarly pitched interference pattern of two sinewaves. Although we expected faster reaction times for game-related stimuli, we found similar reaction times for both types of stimuli. However, we could replicate the psychophysiological patterns from study 1. Both sounds equally elicited ORs with game-related stimuli leading to stronger responses and habituation.
The data from the first two studies indicate that BIPs can be considered a type of OR, suggesting a promising avenue towards continuous measurement of presence. Following a similar design, Study 3 employs a STRT task based on more intricate considerations of the processes in human working memory (Oberauer, 2009). Additionally, pupil dilation is being assessed to investigate cognitive load associated with responses to game-related stimuli and BIPs.