How do we know what we know? We use science to investigate things. As a scientist, doc has been trained to collect information by processing data. The measurement of data is not as straight-forward as one might think. There are all sorts of checks and balances built into the system, so that we know that the data that I collect can be compared to the data that you collect. It is important to compare apples with apples and not oranges. Proof happens when different correctly collected data sets are in agreement about the objective being measured.
Let's talk about information. We are not talking about bits and bytes on a computer – we are talking about how we study a field of endeavor – an approach to the science of reality. We take basic concepts from each field and apply that reason to create a new line of thinking. Then we have to demonstrate that the new idea holds by showing measurement that supports the logic. The rigor of testing a hypothesis is such that the theory is constantly questioned, broken into smaller bits and questioned again.
What is required to collect data? The first step is to understand your objective. You are setting up a protocol for people to follow, so that when they measure the same thing, they get a similar result. The details of the protocol should be followed robotically, because the conditions of the measurement are defined by that method. The observer reports the conditions, describes the specifics of the process and keeps a notebook that details each measurement effort.
When you follow a specific protocol, it explains how to calibrate your recording device. A meter is only as good as the settings on the meter. If you take a measurement of a known sample, you can set the meter to give you the proper reading of that solution (chemists tend to talk about solutions). For instance, a pH meter is calibrated with two solutions, pH=4 and pH=10, to give accurate measurement within the most common pH range.
Sometimes, you have to change protocol in the field on the fly. In this case, you still take all the measurements, but you write down the specific reasons for variance from the protocol and the actions taken in collecting the new measurement. When you change the protocol, your results might not compare to results using the original procedure, but the information can still be accurate. In fact, it is these on the fly changes that often impact the theory of the science – you observe something different and follow the question – what happened?, then answer the question why?.
We design the experiment based on what you need to know. If you are doing routine measurement, the protocol is the guide, especially when you have to measure the same thing the same way on a recurring basis. A one time measurement may not have a protocol to follow. Before measuring, you should be familiar with how to use your instrumentation, the devices that you use to physically measure.
There are two basic things to do to destroy the data you collected and make your time spent worthless. The first is to change data on the fly. Once you have recorded a measurement in the field, it stands as the measurement. If you repeat the measurement and get a different number, then write it down and repeat again. Do not erase the first data point and replace it with the second measurement. This biases the data directly. A single line strike can be drawn thru a data point that you feel was measured inaccurately, but that measurement still exists and is recorded. Second, do not to interpret data on the fly, so that your expectation does not bias the data. There is plenty of future time to look back at the measurement process and repeat it if discrepancies make themselves apparent. Working on the interpretation in the field changes the thought process and biases the testing.
The internal check methods on data are called quality assurance and quality control (QA/QC). This is a mathematics based process that takes extra measurements, (blanks, duplicates and splits) to test the limits of the instumentation used in the process. This is different from calibration, but calibration is defintely a component of quality control. What you are doing is optomizing your equipment and process, by repeating measurements and recording the accuracy and precision of these repetitions. Quality control measurements are built into the protocols of analytical chemistry laboratories – those labs designed to measure quality and quantity of chemistry components in all things.
If you want good information, you need to design a good method for collecting and collating your data. The data are the set of bits of information that collectively define the parameters of your measurement. The experiments or games can be repeated through multiple iterations until the answer converges, or can be run through limited repeats with statistical application on the results. Probability and statistics is an absolute essential for proper science, but not necessary for an individual to do good science. Good science is just good observation with proper explanation. Statistical significance is important to the refinement process, in supporting your hypothesis.
Typically, you want to have a lot of data points collected using different measurement techniques. Water quality monitoring is a function of watershed councils; WQM defines the areas of importance when measuring water. Conductivity, turbidity, pH and temperature are examples of the typical measures of water, from streams, ponds and taps. Meta-data is also collected, things like time of day, air temperature and weather. Photo-points are extremely useful, especially for long term data collation and all phones now have cameras. Special tests are also available on demand.
You want to answer your own questions from as many different perspectives as you can.You support your theory with the collected data, using charts and graphs and other techniques of display. A spreadsheet is extremely important for laying out what you know in a way that you can see things. I typically build a mind map and a flow chart and keep on the straight and narrow process of digging deep and not going wide. This is how we can make correllations between things that affect out measurements and our points of view.
Finally, we want to track out depth of comprehension and ask if the data supports the hypothesis. We attempt to devise new tests to support predictions made by the theory and the more multiple confirmations, the merrier. I tend to track costs accurately, things like time spent, mileage, spending on food, entertainment and shelter – it depends on the scope and scale of your queries, and who is paying for the work. Good science provides a foundation for what we know. Feel free to ask questions and develop a theory – we can assist everyone in finding reality with accurate science measurements.
Namaste' ... doc
hiveblocks