1. ## Comparing two data

Hi,
Could anyone suggest a way to compare the following two set of data
data 1 :
Mean average=54.7 gram
standard deviation = 1.1 gram
x=50

Data 2:
Mean average=50.3 gram
standard deviation = 3.7 gram
x=50

I have tried the normal distribution on the first set of data only to end up getting z score of -4.27 ,I can change this to positive but the z score table end with 3-4 maximum limit
Is there any other method to compare these two set of data?

2. ## Re: Comparing two data

Originally Posted by nova12
Hi,
Could anyone suggest a way to compare the following two set of data
data 1 :
Mean average=54.7 gram
standard deviation = 1.1 gram
x=50

Data 2:
Mean average=50.3 gram
standard deviation = 3.7 gram
x=50

I have tried the normal distribution on the first set of data only to end up getting z score of -4.27 ,I can change this to positive but the z score table end with 3-4 maximum limit
Is there any other method to compare these two set of data?
Not exactly sure what you are trying to do here.

Are you trying to do a hypothesis test that these two sets of data came from the same underlying distribution or not?

Are you trying to decide which of the two underlying distributions x is most likely to have come from?

3. ## Re: Comparing two data

You should use the pooled standard deviation instead of just using 1.1 if you do that you should get a more reasonable z value.

4. ## Re: Comparing two data

Originally Posted by romsek
Not exactly sure what you are trying to do here.

Are you trying to do a hypothesis test that these two sets of data came from the same underlying distribution or not?

Are you trying to decide which of the two underlying distributions x is most likely to have come from?
Yeah am trying to figure out you second suggestion.

5. ## Re: Comparing two data

Originally Posted by nova12
Yeah am trying to figure out you second suggestion.
Since the two sets of data are from different underlying distributions I can't agree with Shakarri.

If all you want to do is decide between the two distributions then simply select the one with the largest z-score.

$\left(\dfrac {50-54.7}{1.1}\right)=-4.27$
$\left(\dfrac{50-50.3}{3.7}\right)= -0.27$

$-0.27 > -4.27$ and thus it is more likely that your sample came from the second distribution.

The corresponding probabilities are

$Pr[-4.27] = 9.77x10^{-6}$

$Pr[-0.27] = 0.39$

6. ## Re: Comparing two data

Originally Posted by romsek
Since the two sets of data are from different underlying distributions I can't agree with Shakarri.

If all you want to do is decide between the two distributions then simply select the one with the largest z-score.
Oh you are right, I read it as 50 being the sample size for each distribution and I thought he was seeing if the two sets of data were from the same distribution