Related Experiment Video
Updated: Jun 2, 2026
09:16
Methods of Soil Resampling to Monitor Changes in the Chemical Concentrations of Forest Soils
Published on: November 25, 2016
A primer for nonresponse in the US forest inventory and analysis program
Paul L Patterson1, John W Coulston, Francis A Roesch
1U.S. Forest Service, Rocky Mountain Research Station, 2150A Centre Ave, Suite 350, Fort Collins, CO 80526, USA. plpatterson@fs.fed.us
Environmental Monitoring and Assessment
|May 10, 2011
Summary
Nonresponse in forest inventory data, particularly from denied access on private lands, can bias estimates. Strategies to mitigate this bias include improved stratification and alternative estimation techniques, avoiding plot replacement.
Area of Science:
- Forestry
- Ecology
- Statistical modeling
Background:
- The USDA Forest Service's Forest Inventory and Analysis (FIA) program quantifies US forest resources.
- Nonresponse in data collection can introduce bias into population parameter estimates.
Purpose of the Study:
- Quantify nonresponse magnitude and mechanisms in FIA data.
- Evaluate FIA's nonresponse assumptions and recommend bias mitigation strategies.
Main Methods:
- Quantified nonresponse rates across different land ownership groups.
- Qualitatively assessed the tenability of the 'missing at random' assumption.
- Evaluated current and proposed plot replacement and estimation strategies.
Main Results:
- Nonresponse rates varied from 0% to 21%, with denied access on private land being the primary cause.
- The 'missing at random' assumption was found tenable in most cases but not universally.
- Plot replacement was identified as a strategy to avoid.
Conclusions:
- Implement stratification schemes to ensure data are missing at random.
- Explore alternative estimation techniques using weighting and auxiliary data.
- Refrain from replacing nonresponse sample locations to prevent bias.
