MetaQuestion: A web application for expert knowledge elicitation addressing plant health and applied plant ecology

1. Expert knowledge elicitation provides information to characterize ecological systems and management options. Linking expert knowledge elicitation with a curated question catalog supports a community of practice for ongoing improvement of question quality. 2. The MetaQuestion web app we introduce here draws on the PlantQuest catalog of questions addressing applied plant ecology and management options, making the catalog available in a flexible form for organizers of expert knowledge elicitation. Organizers can select among questions in the catalog, modify them as needed, and generate an instrument customized to their elicitation project. MetaQuestion makes available PlantQuest questions specialized for the study of invasive species such as pathogens and arthropod pests, such as geographic analyses of prevalence and network analysis of the movement of plant materials. 3. Experts answer questions in the customized instrument and their responses are compiled. For settings where internet access may be sporadic, there are options to download the instrument for experts’work and then upload responses later. MetaQuestion provides the resulting dataset in a CSV file for analysis in users’choice of software 4. Development of the PlantQuest catalog and the MetaQuestion app is ongoing, incorporating lessons learned from applications of the app. The MetaQuestion app could also be adapted to address questions from other subject areas.

💡 Research Summary

The paper presents MetaQuestion, a web‑based platform designed to streamline expert knowledge elicitation in applied plant ecology and plant health. At its core lies the PlantQuest catalog, a curated repository of hundreds of research‑grade questions that have been systematically annotated with metadata such as topic (invasive species, pathogens, arthropod pests), analysis type (geographic prevalence, network movement, risk assessment), and response format (multiple‑choice, Likert scale, free text). By exposing this catalog through an intuitive interface, MetaQuestion enables project organizers to assemble, edit, and deploy a bespoke questionnaire in minutes rather than weeks of manual drafting.

The system architecture follows a modern three‑tier model: a React front‑end for questionnaire design and data entry, a Node.js/Express back‑end for API services, and a PostgreSQL database that stores both the question metadata and the collected responses. A key innovation is the offline mode, implemented with service workers and IndexedDB, which allows experts to download a self‑contained questionnaire, record answers on a tablet or laptop without internet, and later synchronize the data when connectivity is restored. This capability is crucial for field work in remote agricultural regions or low‑resource settings where network access is intermittent.

Once elicited, responses are stored with a standardized schema (question ID, text, respondent ID, timestamp, selected option) and can be exported as a single CSV file. The CSV format is deliberately simple so that users can immediately import the data into their preferred analytical environment—R, Python, SAS, or GIS software—without additional transformation. Moreover, the platform can generate an automated preprocessing pipeline that handles missing values, encodes categorical variables, and, when needed, converts geographic coordinates for spatial analysis.

MetaQuestion currently emphasizes plant health applications. Sample questions include “How has the prevalence of pathogen X changed across region Y over the past five years?” and “Which nodes in the plant‑material movement network act as hubs for pest dissemination?” These items are directly linked to downstream analytical methods such as spatial time‑series modeling, social‑network analysis, and risk mapping, allowing expert judgments to be incorporated into quantitative decision‑support models. By doing so, the platform bridges the gap between qualitative expert insight and the quantitative tools used by policymakers and land managers.

The authors report that beta testing with plant pathologists, entomologists, and ecological modelers has yielded valuable feedback. Iterations have refined the user interface, expanded the question catalog, and added basic visualizations of response distributions. Security is addressed through HTTPS encryption and JWT‑based authentication, ensuring that sensitive expert data remain protected.

Future development plans include multilingual support (English, Spanish, French), role‑based access controls for larger collaborative projects, and an automated reporting module that synthesizes key statistics, generates maps, and proposes management recommendations. The authors also envision extending the PlantQuest catalog beyond plant health to domains such as soil science, water resources, and climate‑adaptation strategies, thereby positioning MetaQuestion as a generic expert‑elicitation platform for environmental decision‑making.

In conclusion, MetaQuestion offers a practical, scalable solution for capturing high‑quality expert knowledge in contexts where field conditions are challenging and rapid, evidence‑based decisions are needed. Its combination of a curated question library, flexible questionnaire customization, offline data collection, and seamless export to standard analysis tools makes it a valuable addition to the toolbox of researchers, extension agents, and policy makers working on complex plant‑health challenges.