Science Inquiry: Physical Sciences

Error Analysis

Having taken numerical measures of experimental variables, the experimenter has the opportunity to conduct an analysis of errors. This should be an integral portion of the experimenter’s report. For this course, it is not worthwhile to devote the necessary time to conduct highly sophisticated error analysis, but rather to develop and practice giving error measures which are sufficiently sufficient to allow the experimenter the opportunity to experience the benefits of error analysis.

The purpose of error analysis is to validate the results and application of the results of the experimental process. Proof is very difficult and any presentations of conclusions from an experimental procedure and analysis require an analysis of errors.

Two general types of errors will be considered for analysis. These will be known as internal error and external error. Internal errors are one arising from your experimental process and predict the likely replication of your experimental results. If you repeated trials, would you have the same results? External errors measure the agreement of your experimental observations with those of other experimenters. External errors measure agreement with the rest of the world (often standard values may be considered.

Errors may be expressed or measured in two manners, relative or absolute. The relative errors express the size of the error as a percentage of the measurements. Absolute errors state the actual size of the error. For purposes of comparison and providing simple expressions of error, absolute measures of error are effective, but not as well accepted as the relative error expression.

Returning to the use and application of errors, let us examine the significance of internal and external error. An experiment with low external error would show the experimenter’s work to be in agreement with the findings of others. When and experimenter’ values are close to standard or accepted values, the experimenter does not prove the standards to be correct, but supports the acceptance of such values. By contrast, when the results of an experiment are quite different from standard values, accepted values, or the work of a wide group of experimenters, the experimenter is in the position of questioning the correctness of what has been previously accepted. External error analysis correlates the experimenter’s work to the work of numerous others. The internal error must be first considered to determine whether the work of the experimenter has any credibility. If the internal error is high, the work is questioned. This means the work is in danger of being disregarded because the inconsistent results for the experimental observations.

This all impacts your presentation of your experimental process. To sustain any conclusion, one needs to address the question: If I did it again, would I get the same results? This is the internal error. The internal error should be less than 10% and preferably less than 5%. When the internal error is over 20%, any conclusion reached by the experimenter could be questioned or even totally ignored. The second question would be: Is that what others saw? The external error identifies where ones’ experimental values correlate to the work of others. As with internal error, the external error has influences upon application or interpretation of the experimental work. If the external error is low, the work agrees with others. If the external error is high, the data is demanding the work of others be rejected. When it conflicts with standard values, it would suggest they are wrong. Obviously, if the internal error is high, any challenge of accepted standards or theories suggested by the external error is irrelevant.

Within the lab report, the error section comes between the analysis and the conclusion. Error analysis cannot be properly conducted until the analysis process has produced a unified result. When only one trial is conducted, what measure could one have for internal error? The error section likewise is presented prior to the conclusion. If the error analysis fails to demonstrate a compelling reason for the reader or the world to respond to the conclusion, the reader would be excused from even reading the conclusion. A complete, compelling error analysis is a necessary portion of the presentation of the experimental work without which the entire work fails.

In Scientific Inquiry: Physical Sciences, lab reports are used as an instrument to practice and model scientific inquiry. From the student’s viewpoint, they are part of the course grade. From the instructor’s viewpoint, they are an integral part of scientific inquiry. Scientific inquiry has been presented as a combination of experimental and theoretical processes. These seek to identify causal relationships with predictive capability. Engineering and technology should be able to utilize these relationships to improve the quality of life for us and allow us to better fulfill God’s will for our lives. Error analysis is seen by your instructor as an integral portion of that process.