Can Survey Design Reduce Anchoring Bias In Recall Data Evidence From Malawi

Recall biases in retrospective survey data are widely considered to be pervasive and have important implications for effective agricultural research. In this paper, we leverage the survey design literature and test three strategies to attenuate mental anchoring in retrospective data collection: question order effects, retrieval cues, and aggregate (community) anchoring. We embed a survey design experiment in a longitudinal survey of smallholder farmers in Malawi and focus on anchoring bias in maize production and happiness exploiting differences between recalled and concurrent responses. We find that asking for retrospective data before concurrent data reduces recall bias by approximately 34% for maize production, a meaningful improvement with no increase in survey data collection costs. Retrieval cues are less successful in reducing the bias for maize reports and involve more data collection time, while community anchors can exacerbate the bias. Reversing the order of questions and retrieval cues do not help to ease the bias for happiness reports.

Coffee is a growth market. Current estimates indicate that global coffee production (in volume) has increased by more than 60% since the 1990s. Coffee is produced by around 25 million farmers, which are mainly smallholders in developing and least developed countries, and over 70% of the coffee produced is exported, resulting in about 20 billion US dollars annual foreign exchange earnings (ICO, 2020). COVID-19 represented a severe joint supply and demand shock to the global coffee sector, particularly during the first months after the start of the pandemic. As noted by Hernandez et al. (2020), the coffee industry experienced important disruptions downstream the value chain, including the functioning of key export infrastructure and international shipping, which combined with local currency devaluations and volatile coffee prices, which resulted in significant challenges for coffee growers, farm workers, and traders.

Using data from several countries, including Cote d'Ivoire, India, Pakistan, Taiwan, and Thailand, this book analyzes household survey data from developing countries and illustrates how such data can be used to cast light on a range of short-term and long-term policy issues.

In Cooking Data Crystal Biruk offers an ethnographic account of research into the demographics of HIV and AIDS in Malawi to rethink the production of quantitative health data. While research practices are often understood within a clean/dirty binary, Biruk shows that data are never clean; rather, they are always “cooked” during their production and inevitably entangled with the lives of those who produce them. Examining how the relationships among fieldworkers, supervisors, respondents, and foreign demographers shape data, Biruk examines the ways in which units of information—such as survey questions and numbers written onto questionnaires by fieldworkers—acquire value as statistics that go on to shape national AIDS policy. Her approach illustrates how on-the-ground dynamics and research cultures mediate the production of global health statistics in ways that impact local economies and formulations of power and expertise.

Since 2000, IOM has been producing world migration reports. The World Migration Report 2020, the tenth in the world migration report series, has been produced to contribute to increased understanding of migration throughout the world. This new edition presents key data and information on migration as well as thematic chapters on highly topical migration issues, and is structured to focus on two key contributions for readers: Part I: key information on migration and migrants (including migration-related statistics); and Part II: balanced, evidence-based analysis of complex and emerging migration issues.

The interwoven futures of humanity and our planet are under threat. Urgent action, taken together, is needed to change course and reimagine our futures.

This engaging and clearly written textbook/reference provides a must-have introduction to the rapidly emerging interdisciplinary field of data science. It focuses on the principles fundamental to becoming a good data scientist and the key skills needed to build systems for collecting, analyzing, and interpreting data. The Data Science Design Manual is a source of practical insights that highlights what really matters in analyzing data, and provides an intuitive understanding of how these core concepts can be used. The book does not emphasize any particular programming language or suite of data-analysis tools, focusing instead on high-level discussion of important design principles. This easy-to-read text ideally serves the needs of undergraduate and early graduate students embarking on an “Introduction to Data Science” course. It reveals how this discipline sits at the intersection of statistics, computer science, and machine learning, with a distinct heft and character of its own. Practitioners in these and related fields will find this book perfect for self-study as well. Additional learning tools: Contains “War Stories,” offering perspectives on how data science applies in the real world Includes “Homework Problems,” providing a wide range of exercises and projects for self-study Provides a complete set of lecture slides and online video lectures at www.data-manual.com Provides “Take-Home Lessons,” emphasizing the big-picture concepts to learn from each chapter Recommends exciting “Kaggle Challenges” from the online platform Kaggle Highlights “False Starts,” revealing the subtle reasons why certain approaches fail Offers examples taken from the data science television show “The Quant Shop” (www.quant-shop.com)