• What does the graph-maker want you to feel about level, trend, and/or volatility?
• Is the data censored on the left or on the right?
• Are there missing data points in between?
• Does the range on the vertical axis make sense?
• Is the zero set at 0?
• Are you being shown comparison data for no apparent reason?
• Is the aspect ratio strange?
• Does the graph use labels and fonts to suggest causal relationships or to promote emotional responses?

" Every chart, every graph, represents a set of decisions about how to represent the information. "

• Which “problem” can be accommodated in survey data collection?
  • Incorrect sampling frame
  • Population specification error
  • Sampling error
  • Measurement error
  • Social desirability bias

• Typical logical fallacies
  • Omitted variables
  • Survivorship bias
  • Reverse causality
• Addressing empirical challenges
  • Regression Discontinuity Design
  • Difference-in-Differences

" The fact that two trends seem to fluctuate in tandem does not prove that they are meaningfully related to one another. "

• Marketing strategies capitalizing on the statistical appeal
  • Fabricated data
  • Biased polling and meaningless reference groups
  • Cherry-picked data
  • Misleading visualization

" Knowing and recognizing our cognitive biases will help prevent us from being fooled by misleading ads. "

" Individuals tend to brace information that supports their wishes and reject information that contradicts them. "