# Statistical Significance for Percentage Change

#### bwells94

##### New Member
Hello,

I'm wondering if there is a way to test for statistical significance if the only data I am given is a sample size and a percentage change from 0.

For example: Based on data from 79 games. There is a 17.9% increase in total runs scored. Is there an equation to use to tell if this is significant or is there not enough data given?

#### hlsmith

##### Less is more. Stay pure. Stay poor.
I would think a measure of dispersion may be needed. But it is still a little unclear. So this is for a bunch of teams averaged or a single team across games? Also, increase versus what?

#### bwells94

##### New Member
I would think a measure of dispersion may be needed. But it is still a little unclear. So this is for a bunch of teams averaged or a single team across games? Also, increase versus what?

It's the runs scored at the stadium based on weather factors. It would be a bunch of teams but the stadium itself is the constant. The % increase would be based off the avg # of runs scored at the stadium which is not given to me.

#### bwells94

##### New Member
I'm trying to figure out if the 15.3% increase in runs is statistically significant based off a sample size of 79 games. This is all the information that I am given.

#### Attachments

• 61.2 KB Views: 2

#### hlsmith

##### Less is more. Stay pure. Stay poor.
You are presuming I know something you don't about these data. What the heck does the pic represent? Is that image trying to say Twins have 15% more runs than Cleveland on comparable days? What information is available on the dropdown?

Given my interpretation is correct, I still don't know if that estimate is just a crude count of twin hits / crude count of Cleveland hits or mean hits in twins games / mean hits in cleveland games, etc.?

It would also assume their competition was exchangeable, my comparable and venue didn't matter, right?

Not trying to be curt, just not much to work from.

#### bwells94

##### New Member
You are presuming I know something you don't about these data. What the heck does the pic represent? Is that image trying to say Twins have 15% more runs than Cleveland on comparable days? What information is available on the dropdown?

Given my interpretation is correct, I still don't know if that estimate is just a crude count of twin hits / crude count of Cleveland hits or mean hits in twins games / mean hits in cleveland games, etc.?

It would also assume their competition was exchangeable, my comparable and venue didn't matter, right?

Not trying to be curt, just not much to work from.
Sorry for the vague information.

All of the information is weather based. So there have been 79 games at the Cleveland Indians ball park that match the weather seen on the left side of the screenshot. In those 79 games, there have been 15% more runs than the average. Unfortunately I do not have the average number of runs scored, I am only given that there is 15% more.

#### hlsmith

##### Less is more. Stay pure. Stay poor.
I think regardless, you are out of luck. I say this since we don't know if the 79 games were used in the overall average and more importantly, what is the dispersion about the means. If the number of hits varies quite a bit from game to game you may not be able to rule out chance.

You can play around with numbers if you have some and say that if they have about 35 hits a game and a std of 8, what would be 15% more plus confidence intervals, etc., say 5 additional hits. Also, this might be a situation where there should be a dispersion parameter on the 15% too. It is hard to tell how the numbers were calculated.