User Tools

Site Tools


general_design_topics:maintaining_data_integrity

Report a bug, broken link, or incorrect content

Maintaining Data Integrity


prepared by Grant Hamilton

Introduction

This information is derived from the Monitoring Manual for Grassland, Shrubland and Savanna Ecosystems and is intended for use with the monitoring methods described therein. These principles are also applicable to a variety of rangeland monitoring methods and protocols.

The power of monitoring data cannot be over¬stated. As data are applied to land management decisions and research questions, the utility of the data are amplified. A data error in the field can be compounded as analysis and interpretation of the data progresses, and can ultimately affect results and conclusions. Conversely, high quality data will be strengthened by strict adherence to protocols and procedures to minimize sampling error. For this reason, correct and consistent technique among field observers and careful attention of data recorders is critical. A carefully planned sequence of quality assurance and quality control steps will ensure the integrity and accuracy of the data.

Data integrity is the responsibility of every member of a monitoring team, from the principal investigator, to individual field technicians, to the database manager(s). It is an ongoing process that begins before the monitoring project begins and follows the project to its conclusion. Data quality protocols will vary from project to project.

Types of Sampling Error

Sampling error refers to measuring errors that arise due to flawed measuring or recording of data. Faulty equipment can result in inaccurate data collection. Data integrity can be compromised during the recording process by human error (e.g. neglecting to account for declination when taking a compass bearing for a transect that must run from true north to true south). For general information on sample design see Sample Design for Rangeland Assessment and Monitoring.

Quality Assurance

Quality assurance is a proactive process intended to prevent the occurrence of error. This process occurs throughout the collection, recording, tabulation, and analysis processes.

Quality Control

Quality control is a reactive process to detect measurement errors after the data collection process is complete. Unlike quality assurance it does not occur in the field as data is measured and recorded. Quality control will also determine compliance with applicable standards and can be project or protocol specific. After fieldwork is completed and error is discovered, it can be corrected during the data analysis process.

Data Calibration

Before beginning a monitoring project, protocols should be developed or adopted. Determining an acceptable margin of error and data variation expectations for the project is the initial step of data calibration. Expert monitoring personnel should train novices following a standard protocol. A training manual may be beneficial. When training field crews, practice procedures should follow the actual methods that will be in the field.

Suggestions for Promoting Data Integrity

Core Method Indicators Assessed
Continuosuly• Practice proper technique.
• Maintain data organization.
• Document errors.
• Keep ecological context in mind.
• Solicits expert advice if needed.
• Back up your data.
Daily • Review data sheets for completeness. If errors are found, return to the plot to collect the correct data.
• Upload and name photos.
• Identify unknown plant species.
• Back up your data after corrections have been made.
Weekly • Review data for completeness and errors with an ecosystem expert or team leader.
• Identify any remaining unknown plant species.
Back up your data.
Monthly AND upon change to a new ecosystem • Calibrate data gatherers.
• Review data for completeness and errors with an ecosystem esperts or team leader.
• Back up your data.

References

Bureau of Land Management. (1996). Sampling vegetation attributes. Interagency technical reference BLM/RS/ST-96/0002+1730. http://www.nrcs.usda.gov/Internet/FSE_DOCUMENTS/stelprdb1044175.pdf

  • Chapter 3 covers sample design and sampling error.

Hoshmand, R.A. (2006). Design of Experiments for Agriculture and the Natural Sciences, 2nd Ed. Boca Raton, FL: Taylor and Francis

  • Includes a chapter on sources of error and how to reduce the likelihood of error.

USDA-ARS Jornada Experimental Range. (2014). Monitoring Manual for Grassland, Shrubland and Savanna Ecosystems, Volume I: Core Methods. 2nd Ed. http://jornada.nmsu.edu/monit-assess/manuals/monitoring.

Discussion/Comments

You must have an account and be logged in to post or reply to the discussion topics below. Click here to login or register for the site.

general_design_topics/maintaining_data_integrity.txt · Last modified: 2014/02/17 21:18 by jgh