Utterance Clustering Using Stereo Audio Channels.

Utterance clustering is one of the actively researched topics in audio signal processing and machine learning. This study aims to improve the performance of utterance clustering by processing multichannel (stereo) audio signals. Processed audio signals were generated by combining left- and right-channel audio signals in a few different ways and then by extracting the embedded features (also called d-vectors) from those processed audio signals. This study applied the Gaussian mixture model for supervised utterance clustering. In the training phase, a parameter-sharing Gaussian mixture model was obtained to train the model for each speaker. In the testing phase, the speaker with the maximum likelihood was selected as the detected speaker. Results of experiments with real audio recordings of multiperson discussion sessions showed that the proposed method that used multichannel audio signals achieved significantly better performance than a conventional method with mono-audio signals in more complicated conditions.

To Whistleblow or Not to Whistleblow: Affective and Cognitive Differences in Reporting Peers and Advisors.

Traditional whistleblowing theories have purported that whistleblowers engage in a rational process in determining whether or not to blow the whistle on misconduct. However, stressors inherent to whistleblowing often impede rational thinking and act as a barrier to effective whistleblowing. The negative impact of these stressors on whistleblowing may be made worse depending on who engages in the misconduct: a peer or advisor. In the present study, participants are presented with an ethical scenario where either a peer or advisor engages in misconduct, and positive and the negative consequences of whistleblowing are either directed to the wrongdoer, department, or university. Participant responses to case questions were evaluated for whistleblowing intentions, moral intensity, metacognitive reasoning strategies, and positive and negative, active and passive emotions. Findings indicate that participants were less likely to report the observed misconduct of an advisor compared to a peer. Furthermore, the findings also suggest that when an advisor is the source of misconduct, greater negative affect results. Post-hoc analyses were also conducted examining the differences between those who did and did not intend to blow the whistle under the circumstances of either having to report an advisor or peer. The implications of these findings for understanding the complexities involved in whistleblowing are discussed.

Effective Practices in the Delivery of Research Ethics Education: A Qualitative Review of Instructional Methods.

In order to delineate best practices for courses on research ethics, the goal of the present effort was to identify themes related to instructional methods reflected in effective research ethics and responsible conduct of research (RCR) courses. By utilizing a qualitative review, four themes relevant to instructional methods were identified in effective research ethics courses: active participation, case-based activities, a combination of individual and group approaches, and a small number of instructional methods. Three instructional method themes associated with less effective courses were also identified: passive learning, a group-based approach, and a large number of instructional methods. Key characteristics of each theme, along with example courses relative to each theme, are described. Additionally, implications regarding these instructional method themes and recommendations for best practices in research ethics courses are discussed.

A Meta-analytic Comparison of Face-to-Face and Online Delivery in Ethics Instruction: The Case for a Hybrid Approach.

Despite the growing body of literature on training in the responsible conduct of research, few studies have examined the effectiveness of delivery formats used in ethics courses (i.e., face-to-face, online, hybrid). The present effort sought to address this gap in the literature through a meta-analytic review of 66 empirical studies, representing 106 ethics courses and 10,069 participants. The frequency and effectiveness of 67 instructional and process-based content areas were also assessed for each delivery format. Process-based contents were best delivered face-to-face, whereas contents delivered online were most effective when restricted to compliance-based instructional contents. Overall, hybrid courses were found to be most effective, suggesting that ethics courses are best delivered using a blend of formats and content areas. Implications and recommendations for future development of ethics education courses in the sciences are discussed.

Curricular Approaches in Research Ethics Education: Reflecting on More and Less Effective Practices in Instructional Content.

Over the past decade, the effectiveness of ethics education programs has increased with regard to trainee outcomes, such as knowledge, awareness, and ethical decision making. However, despite the overall improvement in training effectiveness, considerable variability still exists across programs. One potential source of variability arises from the substantial range in instructional training content utilized across ethics training courses. The goal of the present effort was to clarify which approaches in ethics education result in positive training outcomes through the identification of instructional content themes. Through a qualitative review of ethics training courses, we identified key themes in instructional content curriculum associated with effective courses: domain-general, domain-specific, standard compliance, professionalism, and process-based. In addition, we identified key themes associated with less effective courses: mixed-specificity, narrow coverage, and idealized ethics. Descriptions and key characteristics of each theme along with example courses are provided. Implications of the content themes for ethics education are discussed.

Review of Instructional Approaches in Ethics Education.

Increased investment in ethics education has prompted a variety of instructional objectives and frameworks. Yet, no systematic procedure to classify these varying instructional approaches has been attempted. In the present study, a quantitative clustering procedure was conducted to derive a typology of instruction in ethics education. In total, 330 ethics training programs were included in the cluster analysis. The training programs were appraised with respect to four instructional categories including instructional content, processes, delivery methods, and activities. Eight instructional approaches were identified through this clustering procedure, and these instructional approaches showed different levels of effectiveness. Instructional effectiveness was assessed based on one of nine commonly used ethics criteria. With respect to specific training types, Professional Decision Processes Training (d = 0.50) and Field-Specific Compliance Training (d = 0.46) appear to be viable approaches to ethics training based on Cohen's d effect size estimates. By contrast, two commonly used approaches, General Discussion Training (d = 0.31) and Norm Adherence Training (d = 0.37), were found to be considerably less effective. The implications for instruction in ethics training are discussed.

Are Ethics Training Programs Improving? A Meta-Analytic Review of Past and Present Ethics Instruction in the Sciences.

Given the growing public concern and attention placed on cases of research misconduct, government agencies and research institutions have increased their efforts to develop and improve ethics education programs for scientists. The present study sought to assess the impact of these increased efforts by sampling empirical studies published since the year 2000. Studies published prior to 2000 examined in other meta-analytic work were also included to provide a baseline for assessing gains in ethics training effectiveness over time. In total,this quantitative review consisted of 66 empirical studies, 106 ethics courses, 150 effect sizes, and 10,069 training participants. Overall, the findings indicated that ethics instruction resulted in sizable benefits to participants and has improved considerably within the last decade. A number of specific findings also emerged regarding moderators of instructional effectiveness. Recommendations are discussed for improving the development, delivery, and evaluation of ethics instruction in the sciences.

Cross-Field Comparison of Ethics Education: Golden Rules and Particulars.

Research misconduct negatively impacts the scientific community and society in general. Providing training in the responsible conduct of research (RCR) to researchers is one viable approach to minimizing research misconduct. Although recent evidence suggests ethics training can indeed be effective, little empirical work has examined the similarities and differences across fields. In the present study, we analyzed 62 empirical studies in engineering, biomedical science, social science, and mixed fields. The findings suggest certain instructional principles, or "golden rules," apply generally to all fields. These golden rules include maintaining a field-specific or field-general approach and emphasizing processes in training. The findings also suggest that content areas contributing to instructional effectiveness vary as a function of field. Generally, it appears that all fields may benefit from taking a multi-pronged approach to ethics education wherein the salient field issues are covered. Implications for RCR education are discussed.

Qualitative Evaluation Methods in Ethics Education: A Systematic Review and Analysis of Best Practices.

Although qualitative research offers some unique advantages over quantitative research, qualitative methods are rarely employed in the evaluation of ethics education programs and are often criticized for a lack of rigor. This systematic review investigated the use of qualitative methods in studies of ethics education. Following a review of the literature in which 24 studies were identified, each study was coded based on 16 best practices characteristics in qualitative research. General thematic analysis and grounded theory were found to be the dominant approaches used. Researchers are effectively executing a number of best practices, such as using direct data sources, structured data collection instruments, non-leading questioning, and expert raters. However, other best practices were rarely present in the courses reviewed, such as collecting data using multiple sources, methods, raters, and timepoints, evaluating reliability, and employing triangulation analyses to assess convergence. Recommendations are presented for improving future qualitative research studies in ethics education.

Validation and Use of a Predictive Modeling Tool: Employing Scientific Findings to Improve Responsible Conduct of Research Education.

Although recent evidence suggests ethics education can be effective, the nature of specific training programs, and their effectiveness, varies considerably. Building on a recent path modeling effort, the present study developed and validated a predictive modeling tool for responsible conduct of research education. The predictive modeling tool allows users to enter ratings in relation to a given ethics training program and receive instantaneous evaluative information for course refinement. Validation work suggests the tool's predicted outcomes correlate strongly (r = 0.46) with objective course outcomes. Implications for training program development and refinement are discussed.

How Do We Know What Works? A Review and Critique of Current Practices in Ethics Training Evaluation.

Training is a costly investment. As such, it is of great interest to know the extent to which that investment is yielding a positive return. Recent meta-analytic efforts have observed that ethics training programs are, indeed, having a positive effect, leading to the conclusion that the programs are working. However, they have also uncovered considerable variability in the effectiveness of ethics training programs, which leads to the purpose of the present study-to review current practices in ethics training evaluation. Through this review of 243 studies, consisting of 380 ethics trainings, we identified major themes in evaluation practices and training design. The tradeoffs associated with these evaluation procedures and designs are discussed, along with directions for future research.

A Comparison of the Effects of Ethics Training on International and US Students.

As scientific and engineering efforts become increasingly global in nature, the need to understand differences in perceptions of research ethics issues across countries and cultures is imperative. However, investigations into the connection between nationality and ethical decision-making in the sciences have largely generated mixed results. In Study 1 of this paper, a measure of biases and compensatory strategies that could influence ethical decisions was administered. Results from this study indicated that graduate students from the United States and international graduate students studying in the US are prone to different biases. Based on these findings, recommendations are made for developing ethics education interventions to target these decision-making biases. In Study 2, we employed an ethics training intervention based on ethical sensemaking and used a well-established measure of ethical decision-making that more fully captures the content of ethical judgment. Similar to Study 1, the results obtained in this study suggest differences do exist between graduate students from the US and international graduate students in ethical decision-making prior to taking the research ethics training. However, similar effects were observed for both groups following the completion of the ethics training intervention.

Professional Decision-Making in Research (PDR): The Validity of a New Measure.

In this paper, we report on the development and validity of the Professional Decision-Making in Research (PDR) measure, a vignette-based test that examines decision-making strategies used by investigators when confronted with challenging situations in the context of empirical research. The PDR was administered online with a battery of validity measures to a group of NIH-funded researchers and research trainees who were diverse in terms of age, years of experience, types of research, and race. The PDR demonstrated adequate reliability (alpha = .84) and parallel form correlation (r = .70). As hypothesized, the PDR was significantly negatively correlated with narcissism, cynicism, moral disengagement, and compliance disengagement; it was not correlated with socially desirable responding. In regression analysis, the strongest predictors of higher PDR scores were low compliance disengagement, speaking English as a native language, conducting clinical research with human subjects, and low levels of narcissism. Given that the PDR was written at an eighth grade reading level to be suitable for use with English as a second language participants and that only one-fourth of items focused on clinical research, further research into the possible roles of culture and research ethics training across specialties is warranted. This initial validity study demonstrates the potential usefulness of the PDR as an educational outcome assessment measure and a research instrument for studies on professionalism and integrity in research.

Playing, sitting out, and observing the game: an investigation of faculty members' perspectives on political behavior in ethical decision making.

Ethical dilemmas are inherently ambiguous, complex, and ill-defined. Additionally, these dilemmas involve multiple stakeholders. These characteristics may induce political behavior as a resolution tactic. Thus, the goal of the present effort was to investigate perspectives on politics among researchers in an ethical decision-making context. A qualitative analysis of interviews with university faculty members revealed that faculty members' perspectives on political behavior in an ethical decision-making context fall into a number of categories, including positive, negative, and realistic views of political activity. The implications of these varying perspectives on ethical decision making are discussed.

Researcher perceptions of ethical guidelines and codes of conduct.

Ethical codes of conduct exist in almost every profession. Field-specific codes of conduct have been around for decades, each articulating specific ethical and professional guidelines. However, there has been little empirical research on researchers' perceptions of these codes of conduct. In the present study, we interviewed faculty members in six research disciplines and identified five themes bearing on the circumstances under which they use ethical guidelines and the underlying reasons for not adhering to such guidelines. We then identify problems with the manner in which codes of conduct in academia are constructed and offer solutions for overcoming these problems.

Researcher Perspectives on Conflicts of Interest: A Qualitative Analysis of Views from Academia.

The increasing interconnectedness of academic research and external industry has left research vulnerable to conflicts of interest. These conflicts have the potential to undermine the integrity of scientific research as well as to threaten public trust in scientific findings. The present effort sought to identify themes in the perspectives of faculty researchers regarding conflicts of interest. Think-aloud interview responses were qualitatively analyzed in an effort to provide insights with regard to appropriate ways to address the threat of conflicts of interest in research. Themes in participant responses included disclosure of conflicts of interest, self-removal from situations where conflict exists, accommodation of conflict, denial of the existence of conflict, and recognition of complexity of situations involving conflicts of interest. Moral disengagement operations are suggested to explain the appearance of each identified theme. In addition, suggestions for best practices regarding addressing conflicts of interest given these themes in faculty perspectives are provided.

Differences in Biases and Compensatory Strategies Across Discipline, Rank, and Gender Among University Academics.

The study of ethical behavior and ethical decision making is of increasing importance in many fields, and there is a growing literature addressing the issue. However, research examining differences in ethical decision making across fields and levels of experience is limited. In the present study, biases that undermine ethical decision making and compensatory strategies that may aid ethical decision making were identified in a series of interviews with 63 faculty members across six academic fields (e.g., biological sciences, health sciences, social sciences) and three levels of rank (assistant professor, associate professor, and full professor) as well as across gender. The degree to which certain biases and compensatory strategies were used in justifications for responses to ethical situations was compared across fields, level of experience, and gender. Major differences were found across fields for several biases and compensatory strategies, including biases and compensatory strategies related to use of professional field principles and field-specific guidelines. Furthermore, full professors tend to differ greatly from assistant and associate professors on a number of constructs, and there were differences in the consistency with which biases and compensatory strategies were displayed within these various groups. Implications of these findings for ethics training and future research are discussed.

The Influence of Compensatory Strategies on Ethical Decision Making.

Ethical decision making is of concern to researchers across all fields. However, researchers typically focus on the biases that may act to undermine ethical decision making. Taking a new approach, this study focused on identifying the most common compensatory strategies that counteract those biases. These strategies were identified using a series of interviews with university researchers in a variety of areas, including biological, physical, social, and health as well as scholarship and the performing arts. Interview transcripts were assessed with two scoring procedures, an expert rating system and computer-assisted qualitative analysis. Although the expert rating system identified Understanding Guidelines, Recognition of Insufficient Information, and Recognizing Boundaries as the most frequently used compensatory strategies across fields, other strategies, Striving for Transparency, Value/Norm Assessment, and Following Appropriate Role Models, were identified as most common by the computer-assisted qualitative analyses. Potential reasons for these findings and implications for training and practice are identified and discussed.

A Qualitative Analysis of Power Differentials in Ethical Situations in Academia.

Power and organizational hierarchies are ubiquitous to social institutions that form the foundation of modern society. Power differentials may act to constrain or enhance people's ability to make good ethical decisions. However, little scholarly work has examined perceptions of this important topic. The present effort seeks to address this issue by interviewing academics about hypothetical ethical problems that involve power differences among those involved. Academics discussed what they would do in these scenarios, often drawing on their own experiences. Using a think-aloud protocol, participants were prompted to discuss their reasoning and thinking behind their ethical decisions. These interview data were content analyzed using a semantic analysis program that identified a number of distinct ways that academics think about power differences and abuses in ethical situations. Implications of these findings are discussed.