Reliabilty and Accuracy Related?? Help! (1 Viewer)

[Mr_Kap]

Is reliability affected by accuracy? Like if you had 5 results and averaged them all, however one of the results was an outlier, you would end up with an inaccurate result. Would this outlier now affect the reliability of the result, as if performed another 5 times but no outlier occurs, when you average the results you will get a different average to the one before.

 

[Mr_Kap]

bump!! Anyone?

 

[ChasingMedForLife]

Accuracy refers to the equipment you use. So rather than using a 100ml beaker to measure 9ml of water, you use 10ml measuring cylinder.
Reliability refers to receiving similar results across repeats. i.e repeat experiments 5 times. If similar results are obtained then the experiment is reliable.

 

[Mr_Kap]

Accuracy refers to the equipment you use. So rather than using a 100ml beaker to measure 9ml of water, you use 10ml measuring cylinder.
Reliability refers to receiving similar results across repeats. i.e repeat experiments 5 times. If similar results are obtained then the experiment is reliable.

Thx, but it doesn't help me, i already know what they mean.

Can anyone answer the situation i had above?

 

[Mr_Kap]

bump

 

[teridax]

Thx, but it doesn't help me, i already know what they mean.

Can anyone answer the situation i had above?

I don't think reliability and accuracy overlap if that's what you're asking - they're completely distinct scientific terms.

 

[Mr_Kap]

I don't think reliability and accuracy overlap if that's what you're asking - they're completely distinct scientific terms.

But here the result is innacurate because an outlier was not omitted from the average.

In turn, when the experiment is performed again and 5 more results are taken, the outlier is most likely not going to happen again, so the average will be different to before. If the averages are different, then its not reliable results.

So does that mean in this case because the outlier was not ommited in the first experiment results, it lead to innacurate and unreliable results? Hence, it is the lack of accuracy in the first experiment which lead to the unreliablity of the results?

 

[Librah]

I think you need to re-look the definitions of reliability and accuracy lol

You can take 1000 trials and if they give consistent results it's reliable, but the results could be entirely inaccurate.

Also you don't take outliers in your averages, you disregard them, since you assume something experimental went wrong.

 

[Mr_Kap]

I think you need to re-look the definitions of reliability and accuracy lol

You can take 1000 trials and if they give consistent results it's reliable, but the results could be entirely inaccurate.

Also you don't take outliers in your averages, you disregard them, since you assume something experimental went wrong.

I know what validity, accuracy and reliability mean (i think). This is off the top of my head.
Validity - If the method is 'correct' and meets the aim, involving controlling variables (so you are testing what you are meant to be testing), minimising all systematic errors and random errors (by taking multiple measurements to get rid of outliers, as you can't have valid results without results being reiliable).
Accuracy - How 'correct' /exact you're results are. Can be improved by using better equipment and avoiding human errors like parallax and by having a valid experiment. Am i right in saying you can't have accurate results without a valid experiment?
Reliability - Refers to the consistency of the results. Usually reliability can be improved by repeating the experiment or making more measurements in the experiment.

Anyway, That's not what i'm trying to say. Maybe i'm not making myself clear enough.

There was a question where they DID take the outlier in the average of 5 results, and asked to assess the validity and reliability of the average result.

What I am saying to do with the reliability part, is that the average result is not accurate because it did not exclude the outlier, meaning that it is not reliable because when repeated again the outlier may not happen again, or may not be included in the average (like what is supposed to happen). This means that a different average result is obtained from the first meaning the first was not reliable.

Is this correct or not?

 

[Librah]

I know what validity, accuracy and reliability mean (i think). This is off the top of my head.
Validity - If the method is 'correct' and meets the aim, involving controlling variables (so you are testing what you are meant to be testing), minimising all systematic errors and random errors (by taking multiple measurements to get rid of outliers, as you can't have valid results without results being reiliable).
Accuracy - How 'correct' /exact you're results are. Can be improved by using better equipment and avoiding human errors like parallax and by having a valid experiment. Am i right in saying you can't have accurate results without a valid experiment?
Reliability - Refers to the consistency of the results. Usually reliability can be improved by repeating the experiment or making more measurements in the experiment.

Anyway, That's not what i'm trying to say. Maybe i'm not making myself clear enough.

There was a question where they DID take the outlier in the average of 5 results, and asked to assess the validity and reliability of the average result.

What I am saying to do with the reliability part, is that the average result is not accurate because it did not exclude the outlier, meaning that it is not reliable because when repeated again the outlier may not happen again, or may not be included in the average (like what is supposed to happen). This means that a different average result is obtained from the first meaning the first was not reliable.

Is this correct or not?

Outliers address validity, not reliability. If the question was to assess the validity, you would say it's not valid and justify etc. If the other 4 results were concordant, you say that it would be reliable, only if you excluded the outlier. But you're just stating the obvious, that it's unreliable to include an outlier.

Also why does including an outlier make an experiment inaccurate? The outlier could give you the value your looking for, but if it's not consistent with your other experimental results, you can't include it in your final results.

Part I bolded is wrong.

 

[Mr_Kap]

Outliers address validity, not reliability. If the question was to assess the validity, you would say it's not valid and justify etc. If the other 4 results were concordant, you say that it would be reliable, only if you excluded the outlier. But you're just stating the obvious, that it's unreliable to include an outlier.

Also why does including an outlier make an experiment inaccurate? The outlier could give you the value your looking for, but if it's not consistent with your other experimental results, you can't include it in your final results.

Part I bolded is wrong.

And the Part bolded is correct as can be seen from HSC online and other websites. This is from HSC online, "A test can be reliable but invalid. That is, a test can give reliable, consistent results, but not measure what it is supposed to. A test cannot be valid however, if it is not reliable."

And the question said to assess the validity, accuracy and reliability of the AVERAGE RESULT, not 5 individual results. So what I'm trying to say (in regard to the reliability of the AVERAGE RESULT), is that the average result is not reliable because the outlier was included in the average, and if the experiment is performed again, then that outlier may not occur, or even if it does occur it won;t be included in the average if done properly next time. This will lead to a different average result to the one that included the outier, meaning that average result is unreliable.

Is this correct or wrong? And why.

 

[Librah]

And the Part bolded is correct as can be seen from HSC online and other websites. This is from HSC online, "A test can be reliable but invalid. That is, a test can give reliable, consistent results, but not measure what it is supposed to. A test cannot be valid however, if it is not reliable."

And the question said to assess the validity, accuracy and reliability of the AVERAGE RESULT, not 5 individual results. So what I'm trying to say (in regard to the reliability of the AVERAGE RESULT), is that the average result is not reliable because the outlier was included in the average, and if the experiment is performed again, then that outlier may not occur, or even if it does occur it won;t be included in the average if done properly next time. This will lead to a different average result to the one that included the outier, meaning that average result is unreliable.

Is this correct or wrong? And why.

"A test cannot be valid however, if it is not reliable." I suppose that's true if your considering past experimental data as part of your experiment.

Your misinterpreting the question,

I don't know how to answer what your asking because it's ill-informed, your not taking 5 results, averaging them and then treating them discretely and taking more averages. Think about this, why would you take 5, 5, 5 results, average them individually, instead of just taking 15 and then averaging them? Questions aren't that badly written.

It's asking you to assess "the average result," as in, the one with 5 results.

Just for clarity, post the original question word for word.

 

[DepressedPenguino]

Is reliability affected by accuracy? Like if you had 5 results and averaged them all, however one of the results was an outlier, you would end up with an inaccurate result. Would this outlier now affect the reliability of the result, as if performed another 5 times but no outlier occurs, when you average the results you will get a different average to the one before.

I dont understand what u mean by this part : "as if performed another 5 times but no outlier occurs, when you average the results you will get a different average to the one before."

But to answer the top part: "Is reliability affected by accuracy? Like if you had 5 results and averaged them all, however one of the results was an outlier, you would end up with an inaccurate result. Would this outlier now affect the reliability of the result", you would not have an accurate conclusion or whatever if you take the average of the outlier. For this reason, we don't take into account the outlier when averaging and we repeat our experiment and take the averages of all the repetitions to ensure greater reliability (aka minimising outliers)

 

[rand_althor]

I dont understand what u mean by this part : "as if performed another 5 times but no outlier occurs, when you average the results you will get a different average to the one before."

I think he means a scenario like this, where Result 5 of Trial 2 is an outlier which results in Trial 2 having a different average to the other trials:

 

[DepressedPenguino]

I think he means a scenario like this, where Result 5 of Trial 2 is an outlier which results in Trial 2 having a different average to the other trials:

Oh xD

 

[Mr_Kap]

I think he means a scenario like this, where Result 5 of Trial 2 is an outlier which results in Trial 2 having a different average to the other trials:

yeh, and then the question would be asking to assess the reliability, validity, and accuracy of the average result in trial 2.

 

[Librah]

lmao is that actually the question? Could you show us the original?

 

[Mr_Kap]

lmao is that actually the question? Could you show us the original?

No its not the actual question lol. I can try find the actual question though. It was in a chem trial paper I took.