Bias, Accuracy, Observation and Measurement

Note: The physical concept of what consitutes a "measurement" is not as easy as you think. More on this later in this course.

In essence, one can think of experimental and/or observational bias in 4 different ways:

• direct bias: choosing what to measure or see based on prior prejudice or belief

• Incomplete observations: not being able to observe enough of the phenomena to have a sample of observations that has statistical integrity (Florida 2000)

• Uncalibrated observations: means that others can not quantitatively produce your results because your original measuring device had a zero point error.

• Imprecise observations: your measuring device does not have enough precision to reliably measure the phenomena.

Observation: what can go wrong?

• the observation is imprecise due to limitations of the senses (atmospheric distortion, poor vision, etc)
• the observation is imprecise due to limitations of the apparatus (thermometer, telescope, etc)
• the observation is imprecise due to cultural constraints
• the set of observations is incomplete / incrementally improves (Herodotus' problem) and the observer does no know that.
• Can one ever attain complete or 'perfect' knowledge?

The interpretation of observations: biased due to

• cultural assumptions about the natural order (Amos)
• 'scientific 'assumptions about the natural order (Aristotle, Ptolemy on cosmos; Okham's Razor,)
• weighing of conflicting observations allows for bias, 'the intuitive' to carry more authority ( immobility of the earth; earth at center)

What can be done about it?

• simply record the facts...as in Hippocratic corpus and Thucydides.
• The accumulation of accurate observation is useful, tho we may not have full understanding. White urine is associated with health.
• record the repeated experiments dealing with independently verifiable events...what Galen suggests, but would that have help Herodotus
• pursuit of ever more knowledge, more precise and verifiable experiments?