1. ## Reasons why

Hey,

I'm after a bit of advice.

I have some data and its exponentially increasing by a growth factor of 1.618;

Now im trying to think of two reasons as to why an exponential model would be more suitable than a linear model for my data?

Any one help?

Thanks

Hey,

I'm after a bit of advice.

I have some data and its exponentially increasing by a growth factor of 1.618;

Now im trying to think of two reasons as to why an exponential model would be more suitable than a linear model for my data?

Any one help?

Thanks

I don't understand your question. You say it is "exponentially increasing". Why do you say that? Whatever your answer, that's a reason isn't it?

If this is a homework problem and you are just "given" some data and asked to "give two reasons", obviously, the reasons will depend on the specific data.

Hey,

I'm after a bit of advice.

I have some data and its exponentially increasing by a growth factor of 1.618;

Now im trying to think of two reasons as to why an exponential model would be more suitable than a linear model for my data?

Any one help?

Thanks

Well if it is growing by a constant factor each time it won't fit a linear model, since the amount it's increasing by keeps changing. In exponential models it increases by a constant factor, as in your data, and in linear it increases by constant amount, which clearly wouldn't fit here.

4. ok so we have then the data:

Time/wks - $1, 2, 3, 4, 5, 6, 7$ (List 1)
Toy ownership - $3, 5, 8, 13, 21, 33, 55$ (List 2)

I've used exponential regression when plotting the data, worked out the growth factor to be 1.618

Now why would i use exponential model over linear?

ok so we have then the data:

Time/wks - $1, 2, 3, 4, 5, 6, 7$ (List 1)
Toy ownership - $3, 5, 8, 13, 21, 33, 55$ (List 2)

I've used exponential regression when plotting the data, worked out the growth factor to be 1.618

Now why would i use exponential model over linear?
Well...as you say, it's 'exponentially increasing'. Isn't that reason enough? A linear model wouldn't work.

6. so why is a exponential model more suitable than a linear model?

so why is a exponential model more suitable than a linear model?
Plot the data, overlay the plot with the result of linear regression, and exponential. Is one better than the other?

Alternativly compute the sum of square residuals for both fits to the data is one smaller than the other? With the residuals you could if you wish construct some sort of statistical test for the adequacy of the two models.

CB

8. I think the answer is easier than i'm making out, i just cant seem to put down in words what they are after?

Is it because an linear model doe not fit the data point accurately enough, and therefore an exponential model with regression creates a better picture of whats happening?

9. Yes and because using a liner model for exponential data would be plain stupid.