The Essence of Good Laboratory Technique

"I'll believe it when I see it."   -   G. K. Wertheim

The most important requirement of decent experimental science is honesty. Observe what actually happens. An unexpected result is usually a consequence of carelessness and not an indication of a breakthrough, but it still must not be ignored. Be inspired to check your methods, but do not simply erase the disturbing data.

Make your measurements as precise and as accurate as possible, and keep track of not just your data but also of your precision. You can use a stopwatch to measure how long it takes a lump of clay to fall 10 meters. The stopwatch may report time to hundredths of a second, but it's unlikely your measurements will be that precise. If you measure the time for the same height 3 or 4 times, you'll find your measurements are scattered about some average value. You might then say that the time required to fall 10 meters is 1.3 s ± 0.2 s, so that your precision for any one measurement is only about 0.2 seconds. (Mind you, if you take the average of 10 measurements, the precision of that average value is less than 0.2 s, but we won't go into how to calculate that precision in this course.)

When you're making those measurements, it's possible that you tend to start the watch a bit late, but stop it at just about the right moment. The times you measure, therefore, will systematically be lower than the actual time required for that lump of clay to fall 10 meters. Good experimentalists strive to identify and to eliminate these systematic errors.

You must design your experiments correctly. It's best to vary one thing at a time. For example, drop the same object from a variety of heights. Then drop a different object from that same selection of heights. Don't drop a piece of clay from 10 meters and then a crumpled piece of paper from 6 meters and then a maple seed from 5 meters. It's difficult to learn anything from that. By varying only one parameter at a time, you are collecting data that are likely to tell you something useful when you graph them.