Is there an ethics of algorithms?

We argue that some algorithms are value-laden, and that two or more persons who accept different value-judgments may have a rational reason to design such algorithms differently. We exemplify our claim by discussing a set of algorithms used in medical image analysis: In these algorithms it is often necessary to set certain thresholds for whether e.g. a cell should count as diseased or not, and the chosen threshold will partly depend on the software designer’s preference between avoiding false positives and false negatives. This preference ultimately depends on a number of value-judgments. In the last section of the paper we discuss some general principles for dealing with ethical issues in algorithm-design.     

Student Ratings and Professor Self-Ratings of College Teaching: Effects of Gender and Divisional Affiliation

Twenty female and 23 male professors at a liberal arts college participated along with their 803 undergraduate students in a questionnaire study of the effects of professor gender, student gender, and divisional affiliation on student ratings of professors and professor self-ratings. Students rated their professors on 26 questions tapping five teaching factors as well as overall teaching effectiveness. Professors rated themselves on the same questions as well as on nine exploratory ones. On student ratings, there were main effects for both professor gender (female professors were rated higher than male professors on the two interpersonal factors) and division (natural science courses were rated lowest on most factors). These patterns were qualified by significant interactions between professor gender and division. Although professor self-ratings varied by division, there were few significant correlations between professor self-ratings and students' ratings. Implications for future research are discussed.Appreciation goes to Laura Capotosto and Julie Phelan for their research assistance.Suzanne Montgomery is now at Widener Law School, Philadelphia, PA, USA.     

French by osmosis

The typical nursery school in the province of Quebec, Canada, is conducted in French, the official language of schools and business. Even so, it is not unusual to find English- as well as French-speaking students at many of these schools. Both languages are spoken in the province.     

Making the Critical Appraisal for Summaries of Evidence (CASE) for evidence-based medicine (EBM): critical appraisal of summaries of evidence

Objectives:

Standards for evaluating evidence-based medicine (EBM) point-of-care (POC) summaries of research are lacking. The authors developed a “Critical Appraisal for Summaries of Evidence” (CASE) worksheet to help assess the evidence in these tools. The authors then evaluated the reliability of the worksheet.

Methods:

The CASE worksheet was developed with 10 questions covering specificity, authorship, reviewers, methods, grading, clarity, citations, currency, bias, and relevancy. Two reviewers independently assessed a random selection of 384 EBM POC summaries using the worksheet. The responses of the raters were then compared using a kappa score.

Results:

The kappa statistic demonstrated an overall moderate agreement (κ = 0.44) between the reviewers using the CASE worksheet for the 384 summaries. The 3 categories of evaluation questions in which the reviewers disagreed most often were citations (κ =  0), bias (κ = 0.11), and currency (κ = −0.18).

Conclusions:

The CASE worksheet provided an effective checklist for critically analyzing a treatment summary. While the reviewers agreed on worksheet responses for most questions, variation occurred in how the raters navigated the tool and interpreted some of the questions. Further validation of the form by other groups of users should be investigated.

Highlights

• Few critical appraisal tools have been evaluated with inter-rater reliability testing.
• The ways that users of evidence-based medicine (EBM) point-of-care (POC) tools interpret how to appraise an evidence summary—particularly when defining the grading of evidence, currency, and bias—may vary even when a standard evaluation sheet is used.
• The Critical Appraisal for Summaries of Evidence (CASE) worksheet had a moderate level of inter-rater reliability, similar to previous evaluative studies of critical appraisals tools.

Implications

• Medical librarians can develop tools useful for librarians, students, and clinicians to guide them in appraising clinical evidence summaries.
• The CASE worksheet can be a valuable tool to consider the quality of individual evidence summaries and to see patterns of overall quality in EBM POC tools.

     

Investigating the Role of Problem Severity and Channel Publicness on the Sophistication of Support-Seeking Messages on Social Network Sites: A Replication Study

Oh and LaRose (2016) contended that problem severity and channel publicness shape support seekers’ goals, time spent composing messages, and the sophistication of those messages. The current study seeks to replicate and extend Oh and LaRose’s study to produce a more accurate understanding of the factors that influence the sophistication of support-seeking messages online. Support seekers’ goals varied according to the severity of a stressor and the publicness of a channel, and the amount of time people spent composing a message influenced its sophistication. We extend the original study by examining associations that were not initially tested. Time spent composing messages mediated the influence of problem severity on the quality of people’s messages, and this effect was moderated by channel publicness.     

Getting ‘What Works’ working: building blocks for the integration of experimental and improvement science

ABSTRACT

As a systemic approach to improving educational practice through research, ‘What Works’ has come under repeated challenge from alternative approaches, most recently that of improvement science. While ‘What Works’ remains a dominant paradigm for centralized knowledge-building efforts, there is need to understand why this alternative has gained support, and what it can contribute. I set out how the core elements of experimental and improvement science can be combined into a strategy to raise educational achievement with the support of evidence from randomized experiments. Central to this combined effort is a focus on identifying and testing mechanisms for improving teaching and learning, as applications of principles from the learning sciences. This article builds on current efforts to strengthen approaches to evidence-based practice and policy in a range of international contexts. It provides a foundation for those who aim to avoid another paradigm war and to accelerate international discussions on the design of systemic education research infrastructure and funding.     

Use of Standardized Tests in Teacher Evaluation for Career Ladder Systems

In most current teacher evaluation systems, tests are typically used for all teachers or for none. Peterson describes how three Utah school districts employed tests, specifically National Teachers Examinations, as one line of optional evidence for use along with a selection of other evidence in the promotion of eligible teachers. The author indicates that the optional use of test scores both answers many objections to the testing of veteran teachers and allows the test results of teachers choosing to submit them to inform promotion decisions.