Definition of validity, reliability, and accuracy for science (1 Viewer)

[Kimyia]

Hey,
I'm confused about the definitions for validity, reliabiltiy and accuracy for science subjects. I know they come up heaps so could someone help me out with what I should or what I need to talk about for each?

 

[deebsm]

Validity is the application of the experiment. Like you wouldnt use a themometer to measure length.
Reliability is like whether the experiment has been repeated or not. Always put repetition if if asks for something like "How could you improve the reliability of this experiment?" Also depending on the experiment it reliability can be improved through a large sample size.
Accuracy is like parallax error and using the same equipment because timers are different and rulers are different...

I hope this helped :S I never really understood the meanings of these either and so I kind of made it up and for now it has proven me fine. Most probably look it up in the textbook?

 

[Aluminesis]

Damn, my teacher gave us a sheet of definitions for these but I chucked all my HSC sheets out a few days back. I never completely understood them myself, but the general consensus from all my classes was:

Validity: the experiment correctly tests the aim - i.e. all factors are kept constant aside from those being investigated. If in biology, for example, you were testing say the level of enzyme activity with respect to changes in temperature, you obviously wouldn't do one experiment at 30 degrees and say ph 4 and another at 40 degrees at pH 8; you'd keep everything except temperature constant.

Reliability: how consistent the results are. When you repeat an experiment and get the same results, it is reliable.

Accuracy: the easiest one, simply how accurate it is. Measuring with a stopwatch that calculates milliseconds is obviously more accurate than one that only calculates up to seconds. A data logger is obviously more accurate than the naked eye.

A funny example my teacher told us was this: a scientist set out to prove that cockroaches listened through their legs. He ripped of a cockroach's legs, set it on a table and clapped loudly over it. It didn't move. He did it again; ripped off the legs, clapped, and it didn't move. "See," the scientist said. "Without its legs, the cockroach can't hear the clap and doesn't move." My teacher said this experiment was reliable because you got the same results every time, but invalid because obviously, the poor roach needs legs to run away.

Again, not completely sure that my definitions are right. Reliability and accuracy I'm fairly confident about; validity was the one I was always iffy with. I suggest checking with a couple of teachers and classmates to be on the safe side. In school assessments, it'd be safest to go by what your teacher tells you.

 

[Uniqueness]

Damn, my teacher gave us a sheet of definitions for these but I chucked all my HSC sheets out a few days back. I never completely understood them myself, but the general consensus from all my classes was:

Validity: the experiment correctly tests the aim - i.e. all factors are kept constant aside from those being investigated. If in biology, for example, you were testing say the level of enzyme activity with respect to changes in temperature, you obviously wouldn't do one experiment at 30 degrees and say ph 4 and another at 40 degrees at pH 8; you'd keep everything except temperature constant.

Reliability: how consistent the results are. When you repeat an experiment and get the same results, it is reliable.

Accuracy: the easiest one, simply how accurate it is. Measuring with a stopwatch that calculates milliseconds is obviously more accurate than one that only calculates up to seconds. A data logger is obviously more accurate than the naked eye.

A funny example my teacher told us was this: a scientist set out to prove that cockroaches listened through their legs. He ripped of a cockroach's legs, set it on a table and clapped loudly over it. It didn't move. He did it again; ripped off the legs, clapped, and it didn't move. "See," the scientist said. "Without its legs, the cockroach can't hear the clap and doesn't move." My teacher said this experiment was reliable because you got the same results every time, but invalid because obviously, the poor roach needs legs to run away.

Again, not completely sure that my definitions are right. Reliability and accuracy I'm fairly confident about; validity was the one I was always iffy with. I suggest checking with a couple of teachers and classmates to be on the safe side. In school assessments, it'd be safest to go by what your teacher tells you.

^ This +1

 

[Bobbo1]

There's a really good explanation of this in Andrew Harvey's notes: under the skills section..But just a quick run down, an example of a person throwing darts

Accurate but unreliable: hits the bulls eye first 5 times and then fails the next 5 attempts alternatively
Accurate & reliable: hits the bulls eye consistently 10/10 times
Inaccurate & reliable: consistently hits the 1s 10/10 times
Inaccurate & unreliable: hits all the low scores (1,2,3,4,5,6...) without every hitting the same target again

(if you get what I mean and understand the above you should be good with these 2)

Validity simply means whether the scientific procedure used is correct

 

[Kimyia]

Yeah, I've tried looking for it in my textbook (in two textbooks actually) but I can't find it anywhere Thanks so much for your answer though

 

[Uniqueness]

View attachment reliability_accuracy_and_validity.doc

This is really good if you have a read too

 

[Bobbo1]

View attachment 23616

This is really good if you have a read too

Yep, this is pretty good but what I found is examples are needed for it to really stick to the mind. I first tried rote learning these definitions and failed miserably but really understanding is the way to go

 

[Kimyia]

View attachment 23616

This is really good if you have a read too

Thank you so so much!! That had exactly what I needed

 

[someth1ng]

Reliability: How good is the result if this experiment was repeated.

Validity: How well does this experiment solve the aim/problem --> how fair the experiment is.

Accuracy: How precise were the experiments?


Example:

Determining the molar heat of combustion of Ethanol.
- Experiment repeated 5 times yielded almost identical results - very reliable.
- Heating 100mL of pure water in a high conductive calorimeter and using a thermometer to measure the change in temperature and use the change of enthalpy formula to solve - valid experiment.
- Solving the problem, it was found that only 700kJ/mol was delivered to the water - inaccurate experiment since the true value is ~1367kJ/mol.

As you can see, it's the inaccuracies that tend to affect school lab experiments as equipment is rather basic.

 

[Alkanes]

View attachment 23616

This is really good if you have a read too

This is a very good source. +1

 

[Uniqueness]

Yep, this is pretty good but what I found is examples are needed for it to really stick to the mind. I first tried rote learning these definitions and failed miserably but really understanding is the way to go

I completely understand the feeling. It depends on how you learn as well. Some people just likes pictures where as others prefer words. I'm more of a both personally. I link words to pictures. So if I see a picture I could identify it with phrases or sentences. This way I could remember more

 

[scorer]

i need to practice data processing skills for hsc chemistry and physics for upcoming exams. does have any practice papers on these

 

[maxmax]

No, accuracy is wether or not your results support the accepted value. what your thinking of is precision. an example of accuracy is if you dropped a car and a book off a building and the book landed first, the experiment is not accurate as the accepted result is that a car will land first