New expert forum on field trial statistics – April 30th (online)
Nordic Field Trial Network is launching a new expert forum on statistics in field trials. The aim is to strengthen the quality of Nordic field trials through knowledge sharing on trial design, handling of deviations, and statistical analysis. Members are encouraged to consider colleagues who work directly with planning and analysing trials.
The goal of this network is to strengthen the level of field trials in the Nordics by highlighting and discussing the decisions that affect the statistical conclusion of field trials. That includes
- Field trial design and placement
- Handling of deviations from the original plan
- Statistical analysis of field trial data
We are aiming to meet two times a year, the first time being Thursday, 30 April, 13:00–14:30 CEST via Teams. Participants may come from industry or academia - whether statisticians or study directors who regularly face these decisions.
Agenda for the first meeting:
- Introductions (brief, 10 minutes)
- Presentation (25 minutes): How blocking is used in Nordic field trial system and why (Svend V. Nielsen, Danish Technological Institute).
Abstract:
Blocking is a fundamental technique in field trial design that divides plots into groups where experimental conditions are more homogeneous. By accounting for spatial or environmental variation in the statistical analysis, blocking increases precision and reduces experimental error. This presentation will introduce the commonly used design types: the randomized complete block design, row-column design, incomplete block design (a generalization of the alpha design), and row-column incomplete block design. Often, more than one of these four designs could be applied to a given trial. This talk will propose recommendations for when to use each design type and the reasoning behind these conclusions, drawing on experience from the Nordic field trial system.
- Topic sharing (in breakout rooms if many participants, 45 minutes):
Each participant presents 1-2 issues/opportunities they would like to discuss in the forum. It could be concrete questions such as “would you remove this plot as an outlier in this specific trial?” or more general such as “when to choose an alpha-design rather than a complete block design?”
After each presentation, others can weigh in: how they would handle it, how critical the decision is, and what the best possible solution would look like.
- Evaluation and ideas for next step (10 minutes).
After the first meeting, we will add all the questions to a Teams group, where you can continue the discussion.
Link to: sign-up form