Observation and inference

Observations

An observation is information you collect directly with your senses (sight, hearing, touch, smell) or with an instrument. In the lab you never use the fifth sense, taste, because anything on the bench could be contaminated.

Observations come in two types:

Quantitative observations need units. "The plant grew 3" tells nobody anything: 3 leaves? 3 cm? 3 mm a day? Units also let you compare: 30 mm and 5 cm are hard to compare at a glance until you write them both in the same unit (3.0 cm and 5.0 cm).

Instruments don't just make observations more exact; they let us observe things our senses can't detect at all. A microscope shows cells, a thermometer detects a 0.5 °C change your hand would miss, and a Geiger counter detects radiation you can't sense in any way.

Inferences

An inference is an explanation or conclusion you draw from observations, using what you already know. It goes beyond what you directly saw.

Notice that several inferences can fit the same observation. That's normal. To decide between them you collect more observations: look for a broken window, check whether the sprinklers ran. The more reliable your observations, and the more of them you have, the more confident your inference can be.

Some people are experts at this. Aboriginal trackers such as Yuendumu man Teddy Egan, who helped Northern Territory police find missing people and escaped prisoners, read the land with decades of knowledge: a bent grass stem, a scuff on a rock, which way the sand has slumped in a footprint. Each is a small observation; together they support a confident inference about who passed, when and which way they went.

Inferences can change

Because inferences go beyond the evidence, new evidence can overturn them. Observations, if they were made carefully, stay true. This is one of the most important habits in science: hold your inferences loosely and be ready to change them.