Validity (1 Viewer)

[NexusRich]

What are some ways of increasing validity ? let's say that I am conducting a prac of projectile motion in a classroom setting and aiming to find the angle that will produce a certain distance.

 

[Eagle Mum]

What are some ways of increasing validity ? let's say that I am conducting a prac of projectile motion in a classroom setting and aiming to find the angle that will produce a certain distance.

Eliminate or minimise errors and/interferences such as wind draughts, any initial velocity at the time of launch (for example if you use a ‘run up’ to launch, horizontal velocity may vary at your origin) and keep the force(s) at launch (eg. from coiled tension) constant for each different angle that you test. It’s very likely to be impossible to eliminate these, so it would then be appropriate to discuss how these errors, variations and interferences affected your results.

Measurement error also affects validity - for example, your methods (estimating from visual observation where the object actually lands, use of a tape measure) might make measurements valid to the nearest metre or centimetre but invalid to the nearest millimetre.

 

[CM_Tutor]

The term "validity" is, in my opinion, used somewhat carelessly by students and teachers. You need to appreciate the concept but also to align with your teacher's views for assessments prior to the HSC itself.

The term is interpreted differently in different fields and by different researchers / textbook writers / etc. Some use expressions like "increasing validity" (as the OP has used above); others doubt that this phrase. I can make changes in an experiment to increase reliability, precision, and accuracy because those clearly come in degrees... but is the same true of validity? "Reasonably accurate," "high precision," "poor reliability", all of these phrases have a clear meaning. The notion of gradations in validity is more problematic. After all, an experimental method can be valid despite being unreliable or inaccurate. A method can be invalid. But somewhat valid?

I take the view that validity is a property that an experiment / method either has or does not have. If the method is invalid, nothing I can do can change that fact. If it is valid, I can make changes to improve accuracy, reliability, precision, but not validity as that is already established. This comes back, then, to the question of what validity actually is. My perspective is that an experimental / scientific method is valid if (1) whatever is sought to be measured actually exists and (2) changes in whatever is being investigated causitively lead to changes in the measurements that will be made.

As an example, suppose I wanted to test the hypothesis that caterpillars become longer as they age. I need a method of measuring length, such as using a ruler which has a pre-defined and marked scale that I can use to assign a value to caterpillar length. As the caterpillar grows, I can use my method to find new values. The changes in values that I record are caused by changes in the caterpillar. The caterpillar exists and has a property that we might call length, and so the method is valid. Now, instead of a ruler, I could place my finger next to the caterpillar and, by eye, estimate what fraction of the length of my finger is represented by the caterpillar. This would also be a valid method, though obviously less accurate, reliable, and precise. However, if I chose to "measure" length by comparison to the leaf that the caterpillar is eating each day, that would not be valid as the leaves would change over time (and be eaten) so changes in the caterpillar are not causitively producing the changes in measurement - I could find the caterpillar has shrunk as it is standing on a larger leaf on the following day, for example.

let's say that I am conducting a prac of projectile motion in a classroom setting and aiming to find the angle that will produce a certain distance.

In my view, a valid method would only require that you have a method for measuring the angle of projection, are controlling other variables, and have a means to either measure distance or have a fixed point / target so that it is the change in angle of projection that is causing the change in distance. There are plenty of factors that I might consider for accuracy, precision, reliability, but the method / design is either valid or its not.

There are some methods that are valid but have poor accuracy. If a method is invalid, either the results will appear random / unintelligible or the results have accuracy and reliability in which case they are likely a valid measurement of something. Consider the Michelson-Morley experiment, which was designed to measure the properties of the aether that permeates all of space by looking for differences in the measured speed of light in perpendicular directions. The experiment is invalid for the simple reason that the aether does not exist. However, the methods for determining the speed of light were valid and the results demonstrated that the aether could not have the properties that were attributed to it. Experiments in the century or more since have demonstrated that any such x- and y- direction difference, , as a fraction of the speed of light, , are . The Michelson-Morley experiment is now viewed as evidentiary support for Einstein's theory special relativity.