Data Analysis

by admin Mar 05, 2015

This module analyzes ethical questions encountered in the analysis of environmental observations. The examples this module centers on errors that occur from: interpreting noise as significant (Type 1); interpreting real effects as noise (Type 2); overconfidence projections; and other projection biases. This module draws on guidelines and lessons from the American Statistical Association Ethical Guidelines for Statistical Practice and discusses how they might be applied. This module is applicable in an earth systems science course, and is topically concerned with communication of carbon cycle science. 

Ethical issues covered include:

  1. Selection of data 
    • what signifiers to follow
    • what data to include
    • what regions, areas, scales to account for
  2. Scientific uncertainty issues
    • structural uncertainty
    • overconfidence issues
  3. Claiming objectivity
    • False pretense for claims
    • Communicating the overconfidence
    • Proper formation of the negation of the claim (the null hypothesis)
  4. When is there enough data to claim an analytical conclusion? Data robustness
  5. Responsibility in communication of the data to outside of the scientific community, interacting with media