A DEFINITION OF A COMMON MANAGEMENT FRAMEWORK
Note: I thank John Porter, Karen Baker, and John Anderson for their input, and specially Don Henshaw for half of the concepts stated herewith are his ideas.
The centerpiece of a common management framework would be a data access policy:
- Data users must agree to a set of rules about secondary use of public data
- Protection is provided to the data provider through a database citation, disclaimer, and encouraging ethical behavior.
- It also describes an agreement about what data will be put online and when this will be accomplished (i.e., within two years of collection).
- Justification for data sets not put online would be provided.
Information is made available on an in-house computer file system to facilitate research and analysis to the investigators. The information should include:
- A catalog of all data sets - including those for new studies just underway, and the data along with its meta-data.
- A description of all the projects with a list of data sets (with abstracts) that each generate
- In-house accessibility to their data must be provided to owners while ensuring integrity of the data:
- Data sets owners and secondary users would have exclusive access to their data sets
- Mechanisms that provide access to the file system would be made available to investigators that reside off-site
- Computer programs residing on the file system would provide the necessary tools to extract, manipulate, and transfer the data sets to formats suitable for data analysis
- Read-only access to the data sets conserves the flow of the procedures established by the investigator and the data management staff thus assuring data integrity
- Security of the system is assured the implementation of a backup system designed to recover in case of disaster
For every data set, the role of the data manager with respect to every data set should be well defined, as well as the role of pi's and the field's crew. The following questions should be addressed:
- Decide at what stage in the process will the data manager be involved.
- Decide the timing to get the data to the data manager
- Establish the format or software in which data will be passed to the data manager
- Decide who will do data processing, data cleanup, and quality assurance.
- Establish protocols on how field crew personnel interact with pi's and the data manager
- Decide who is responsible to document the data sets.
- Prepare documentation standards.
The content of the common management framework should be formally established:
- Decide if the data catalog and other meta-data will be stored in data base tables.
- The site should decide where is going to archive data
- Methods to ensure security of the data be should be established
The relationship between the common information management framework and the research conceptual framework should be defined and documented in the metadata:
- A cluster of two or more data sets could be defined by grouping sets of data bases with a broader common research context.
- When a cluster is defined, it can be described by answering the question of why the set of data bases are collected and why are they usefull to the originator (the owner and creator of the data sets) and what are the scientific questions that they want to answer by gathering these data.
- A title and the description of this cluster could be included in the metadata for each of the data sets involved. The description can point (link) to another metadata file containing the description.
- The list of these cluster and the standalone data bases provide a front end interface for the users visiting the Data Web Page.
The person(s) that will maintain the web pages must be designated.
The way in which this data will interact with tabular data should be established and implemented.