The difference between accuracy/reliability and validity (1 Viewer)

[sneeble]

I've found my way through half a dozen or so past papers, and a key theme rearing its head is the difference between accuracy and validity with respect to experimental procedures.

Accuracy refers to consistent results. The best way to check it is to repeat the experiment, or compare results with another group in the class. If my dictionary and brain are serving me correctly, accuracy = reliability (an accurate experimental procedure = a reliable one, and vice versa).

Validity refers to the ability of the procedure adopted to actually test the hypothesis you want it to. Do the results reflect what you want them to? Identifying and controlling variable factors is a way of checking your validity.

Anyone want to comment or correct me? It'd be nice to know if I've got it wrong. My teacher isn't great with words - explain/discuss/evaluate/assess/outline all mean the same thing to him - hence my asking here.

 

[Xayma]

Reliability is not always accuracte.

Reliability is that you will get the same result each time. Eg if you have 10mL unknown solution and put .001mol of OH^- and you get a precipitate, (looking for Pb²⁺) then your results will probably be reliable however, not accurate since you would get multiple precipitates.

 

[randhi]

These are my basic definitions of the three
- Validity is when, if the method is correct, state improvements for that experiment.
- Reliability - when the experiment is repeated the same results are obtained
- Acurracy - the results of the experiment are compared with reported values to show they are similar

 

[bhavo]

i got a top notch understanding about acc/rel/val from conq chem website, and where he made notes about one of the past HSC papers (maybe 2001??)

he gave a very good analogy. suppose u hafta hit a bullseye. all your shots may be bunched together, but far from the bullseye, i.e reliable but not accurate. u could have shots which are close to the centre, but scattered around it, i.e accurate but not relaibe.

he explained that reliability is the process of obtaining consistent results in your experiment. this come about by controlling all variables EVERY time, and repeating. accuracy was getting results that are near published data, i.e experiments that were conducted in controlled environments and in standard conditions etc.

the sum of accuracy and reliability, i.e when u have a relaible procedure, and an accurate one, u can call that experiment valid.

hope it helps!

 

[nit]

The problem with these terms is that they're not standard terms used by chemists - the more common terms are accuracy and precision as opposed to reliability - ie precision refers to how close the values in your data set are to one another