Help with regression lines? Help Needed!

A

AbhiKap

Junior Member
Joined
Aug 23, 2012
Messages
60
Hey guys, AbhiKap here

I am in 9th grade and am working on a project about olympic events. Basically, it asks us to predict if women's performance in the olympics are going to surpass men's performance and if so, when?

First we had to collect data and make a scatter plot graph, and then add the regression lines. But now we have to "solve the system of equations using graphing" We then have to check our work by using another method like substitution

I tried to get the two equations I have for the two regression lines and set them equal to each other to try to find X but I got 0. Also, the two lines are parallel.

Please help! Thanks!
 
Last edited:
AbhiKap said:
Hey guys, AbhiKap here

I am in 9th grade and am working on a project about olympic events. Basically, it asks us to predict if women's performance in the olympics are going to surpass men's performance and if so, when?

First we had to collect data and make a scatter plot graph, and then add the regression lines. But now we have to "solve the system of equations using graphing" We then have to check our work by using another method like substitution

I tried to get the two equations I have for the two regression lines and set them equal to each other to try to find X but I got 0. Also, the two lines are parallel.

Please help! Thanks!
Click to expand...

What are those equations? Please post those!!
 
Subhotosh Khan said:
What are those equations? Please post those!!
Click to expand...

Sure, the equations are kind of long, but here you go.

1. y= 0.006091001089833x + -9.791654989879405

[FONT=arial, sans, sans-serif]2. y= 0.00645337069905x + -10.84579012922221 [/FONT]

[FONT=arial, sans, sans-serif]As I said, there are pretty much parallel [/FONT]


​
 
AbhiKap said:
1. y= 0.006091001089833x + -9.791654989879405

2. y= 0.00645337069905x + -10.84579012922221
As I said, there are pretty much parallel
Click to expand...
Yes, but they're not exactly parallel.

AbhiKap said:
I tried to get the two equations I have for the two regression lines and set them equal to each other to try to find X but I got 0.
Click to expand...
How did you "get" this result? Please show your work. Thank you! ;)
 
stapel said:
Yes, but they're not exactly parallel.


How did you "get" this result? Please show your work. Thank you! ;)
Click to expand...

Well actually I used a TI 84 Calculator. I simply put what y equals in the first equation into the y for the second one. I got 0. Any Help?
 
AbhiKap said:
Well actually I used a TI 84 Calculator. I simply put what y equals in the first equation into the y for the second one. I got 0. Any Help?
Click to expand...
Yes. Don't use a calculator which is limited by round-off error. Do the algebra yourself! ;)
 
stapel said:
Yes. Don't use a calculator which is limited by round-off error. Do the algebra yourself! ;)
Click to expand...

OK, I shall try that. But I want to know: is what im doing correct? And would elimination be a better method?

Ill post the answer I got as soon as Im done but I don't know if ill be able to :(

OK I got 2.99....does that seem to make any sense? because it doesnt to me
 
Last edited:
AbhiKap said:
Well actually I used a TI 84 Calculator. I simply put what y equals in the first equation into the y for the second one. I got 0. Any Help?
Click to expand...

If you have access to a spreadsheet program like MS-Excel, use it. I got reasonable answers using that software.
 
AbhiKap said:
OK, I shall try that. But I want to know: is what im doing correct? And would elimination be a better method?

Ill post the answer I got as soon as Im done but I don't know if ill be able to :(

OK I got 2.99....does that seem to make any sense? because it doesnt to me
Click to expand...

You got 2.99 for what?
 
Subhotosh Khan said:
If you have access to a spreadsheet program like MS-Excel, use it. I got reasonable answers using that software.
Click to expand...

I do not have Excel but I can use Google Spreadsheet...how can I use that for help here?

EDIT: Using an online tool, I managed to get the answer 2909 for X and 7.92 for Y. Does this mean that on the year 2909, women will beat men by 7.92 distance? Im confused
 
Last edited:
AbhiKap said:
I do not have Excel but I can use Google Spreadsheet...how can I use that for help here?

EDIT: Using an online tool, I managed to get the answer 2909 for X and 7.92 for Y. Does this mean that on the year 2909, women will beat men by 7.92 distance? Im confused
Click to expand...
The meaning of your result is the prediction that in the year 2909, men and women will have the same performance number, which will be 7.92.
 
AbhiKap said:
I do not have Excel but I can use Google Spreadsheet...how can I use that for help here?

EDIT: Using an online tool, I managed to get the answer 2909 for X and 7.92 for Y. Does this mean that on the year 2909, women will beat men by 7.92 distance? Im confused
Click to expand...

When found the regression lines:

What was the definition of variable 'x'?

What was the definition of variable 'y'?
 
Subhotosh Khan said:
When found the regression lines:

What was the definition of variable 'x'?

What was the definition of variable 'y'?
Click to expand...

I'm not sure but I believe the X was the years. Hence, the Y should be the distance. Is there a way to know what X and Y are for sure?
 
AbhiKap said:
I do not have Excel but I can use Google Spreadsheet...how can I use that for help here?

EDIT: Using an online tool, I managed to get the answer 2909 for X and 7.92 for Y. Does this mean that on the year 2909, women will beat men by 7.92 distance? Im confused
Click to expand...

I'm going to make a couple assumptions here. First your regression lines relate performance in some event to the year of that event. You have separate regression lines for men and women. The x values are the year, and the y values are the performance.

If these assumptions are correct then the fact that the two regression lines intersect at (2909, 7.93) means that, yes the women will catch up to the men in the year 2909 and they will do so at the performance level of 7.93.
 
Last edited:
Romsek said:
I'm going to make a couple assumptions here. First your regression lines relate performance in some event to the year of that event. You have separate regression lines for men and women. The x values are the year, and the y values are the performance.

If these assumptions are correct then what the fact that the two regression lines intersect at (2909, 7.93) means that, yes the women will catch up to the men in the year 2909 and they will do so at the performance level of 7.93.
Click to expand...

Yes, I do believe the assumptions of yours are correct. I don't know if this is against the forum rules, but is it possible for someone to use a calculator/paper/ or anything else to solve the system of equations of the two regression lines. This is because I would like to check if my answers are correct.

Also, the answer look wierd to me. The event is high jump, and how is it possible for someone to jump up 7.92 metres? even though this is in year 2909..we'll still be humans right?
 
The answer looks weird because records so NOT typically increase linearly. They tend, rather, to be "logistic" functions which level off as they near an upper limit.
 
HallsofIvy said:
The answer looks weird because records so NOT typically increase linearly. They tend, rather, to be "logistic" functions which level off as they near an upper limit.
Click to expand...

OK, but is the answer I got correct? I need to make sure it is correct :)
 
AbhiKap said:
Why does no one answer :(
Click to expand...
No one answers because we believe that we have already given you the answers. I shall try to summarize.

Romesk confirmed that your answers are arithmetically correct if you calculated the equations of your regression lines correctly.

Halls of Ivy explained, however, that the calculated regression lines are almost certainly not realistic. First, due to virtually inevitable errors in measurement, regression lines are almost always approximations even when the underlying relationship is actually linear. Second, as indicated by a lot of research, the relationship that you are studying is not linear at all. That means, that even if there are no measurement errors, this is still an approximation that is reliable over only a limited domain.

I have no clue what type of answer your school will approve. They may want you to stick to the math and abandon common sense: in which case, the answer is that men and women will be equally super-heroic around the start of the next millennium and be able to leap over houses in a single bound. Or they may want an answer saying that probably (because this is an approximation) it will not happen during this century. Or they may want an answer that gives the mathematical result and explains why that result is probably not meaningful.

You have in fact received a wealth of answers, but it is your job to figure out what those answers mean.
 
AbhiKap said:
Why does no one answer :(
Click to expand...

It is very disappointing for me to see that you have calculated regression lines without "understanding" the variables. Did you calculate those lines yourself - or you got those from some body else?

I did not answer anymore because I suspect you are trying to find a shortcut to your assignments ......
 
You must log in or register to reply here.
Top