Power function - finding if the slope is significant or not?

#1
Hi,

I'm using power functions to look at a relationship between time and age and I need to be able to test whether or not the slope of this relationship is significant or not. I am using SYSTAT, but cannot find out how to run this and produce a p-value.

Is this not possible to test at all? Any alternatives can think of to test whether slope of a power function is significant or not??

Any help would be great
Cheers
 
#2
Here's an example of one output from running non-linear regression for function ( y = z * x^k )

All I need to do is find out whether this slope is significantly different from 0. Am unsure of how to go about testing this

Sum of Squares and Mean Squares
Source:________/...SS............./df.... /Mean Squares
Regression______/ 42.162539 __/2 ___/21.081269
Residual________/ 0.385111 __/5 ___/0.077022
Total___________/ 42.54765 ___/7
Mean corrected___/0.649023 ___/6

Parameter____Estimate____ASE____Parameter/ASE____ Wald 95% Confidence Interval
_______________________________________________Lower_______Upper
K__________18.475309___20.882025___0.884747____-35.203645____72.154264
Z__________-0.274805___0.154224___-1.781853____-0.67125____0.121641
 

JesperHP

TS Contributor
#3
if a power function is \( y = a x^b\) then what is the slope?

Loglinearize perhaps: \( log y = log a + b log x\) test significance in linear model ... otherwise use non-linear regression ...