Statistics 6+ Flashcards
What are the similarities of non-linear regression and linear regression?
There is a slope/gradient value and there is a constant value which determines the y-intercept. They also both have independent and dependent variables
What are the differences of non-linear regression and linear regression?
Non-linear is characterised by the process of when changes are made to the predictor value (x) they do not result in the same degree of changes to the corresponding y values
What is a slight problem of non-linear regression in terms of its variability?
there are so may forms that it can take
What is the general format of non-linear regression curves and other non-linear curves?
Y = A + bX
What do the A and B values of any regression line mean for non-linear regressions?
B is either a multiplication factor or a power which determines the curvature
A is either a multiplicative or an additive which represents the location of the curve which has been determined by B
What is a multiplicative?
means you multiply any x-value by the same number every time to get the corresponding y-value
What can a multiplicative also be and how?
A divide function by multiplying by a reciprocal
What is an additive?
means you add the same number to any x-value to get the corresponding y-value
What can additive also be and how?
subtraction by adding a negative value
Which important regression statistics are also used/available to understand the regression line in SPSS?
R square
F ratio
T value
What is the general format for a quadratic regression line?
aX^2 + bX + c
What do the a, b and c constants all do in a quadratic function?
a determines whether the curve is u or n shaped and the greater the value the greater the intensity of this shape
b determines the location of the parabola vertex along the parabola
c determines the y-intercept
What is the general format for a cubic regression line?
ax^3 + bx^2 + cx + d
What does the a constant do?
Determines whether the cubic curve goes from down to up then down or the opposite