Thursday, January 21, 2010

Understanding Treatments: 1.20.09


Treatments are a kind of variable…each varies along a distance or curve

Continuous data has not ends, no gaps, can point, or be numerical

Other variables can be treatments like our experimental treatment and control treatment which is not continuous. These variable are categorical.


A control treatment doesn’t have changes (is constant). You have a baseline to judge the treatment against. In our example a single snail with:

1. Normal flow

2. Normal temperature

3. As much food as it wants

4. That’s constantly submerged

5. No experimental damage to its shell


You can change one variable and analyze that…one control and one variable.

When there is too much error you can’t tell the difference.

What error metric is applied and how to interpret it.

Part of collaboration in science is learning from our mistakes.


Hypotheses

An educated guess

Prediction

The kind of idea you can test

with a single experiment

A single article

A paper in a journal


Theories

libraries

Many, many hypotheses

tested, rejected, accepted

A scientific framework

Explains a body of previous observations

They can change


Science: truth without certainty

Everything results in more tests

There is still uncertainty that what we do is incorrect

Theories do not provide absolute truth

No comments:

Post a Comment