Anyone good at statistic want to help me out?

[urbanclassic]

---

Members don't see this ad.

Okay say a top score in a class exam was 98, and someone has a 93. The standard deviation w/in that class was a 5.5. Could you find the percentage of people who were within the top score and less one standard deviation( 98-92.5) ? Isn't there a set percentage for standard deviations?

Thanks.

 

[theshellyb]

Okay say a top score in a class exam was 98, and someone has a 93. The standard deviation w/in that class was a 5.5. Could you find the percentage of people who were within the top score and less one standard deviation( 98-92.5) ? Isn't there a set percentage for standard deviations?

Thanks.

Standard deviation is relative to the mean. So you can go from the mean up, but not the top score down.

 

[urbanclassic]

Standard deviation is relative to the mean. So you can go from the mean up, but not the top score down.

Haha, okay, thank you! I always hated stats 😛

 

[DrSnips]

With the mean and SD you can tell what percentage of the class fell between any 2 scores. Provide the mean.

 

[loveoforganic2]

Okay say a top score in a class exam was 98, and someone has a 93. The standard deviation w/in that class was a 5.5. Could you find the percentage of people who were within the top score and less one standard deviation( 98-92.5) ? Isn't there a set percentage for standard deviations?

Thanks.

Only if you had the mean and only if the distribution was normal

 

[music2doc]

Okay say a top score in a class exam was 98, and someone has a 93. The standard deviation w/in that class was a 5.5. Could you find the percentage of people who were within the top score and less one standard deviation( 98-92.5) ? Isn't there a set percentage for standard deviations?

Thanks.

Yes, if your distribution were normal (it's not -- I can pretty much guarantee that). However, you can usually approximate it still. You almost certainly have a negative skew (with an unknown amount of skew). I would probably use your median instead of the mean since there is skew but if you only have the mean, you could still use it. Just calculate Z-score, (SCORE - MEAN_SCORE)/STD_DEV and then Google Z-score table to match the Z-score to the appropriate percentile. That will give you a reasonable approximation of where you are in the class. For reference, top 25% is generally eligible for AOA (but not all 25% will get it).

 

[urbanclassic]

Yes, if your distribution were normal (it's not -- I can pretty much guarantee that). However, you can usually approximate it still. You almost certainly have a negative skew (with an unknown amount of skew). I would probably use your median instead of the mean since there is skew but if you only have the mean, you could still use it. Just calculate Z-score, (SCORE - MEAN_SCORE)/STD_DEV and then Google Z-score table to match the Z-score to the appropriate percentile. That will give you a reasonable approximation of where you are in the class. For reference, top 25% is generally eligible for AOA (but not all 25% will get it).

Thanks 🙂 I was trying to see where the score lies against the class to see if I honored or not. Honor is top 10-15% class... I probably didn't as there were about 25% who got b/w 90-98 and I am at the lower end of that.

 

[DrSnips]

Thanks 🙂 I was trying to see where the score lies against the class to see if I honored or not. Honor is top 10-15% class... I probably didn't as there were about 25% who got b/w 90-98 and I am at the lower end of that.

Based on what you just said, assuming normal distribution, the mean is about 86. The cutoff to be in the top 15% would be a 92. The cutoff to be in the top 10% is 93.3.