Regression, correlation and hypothesis testing Flashcards
1
Q
y = ax^n
A
logy = log(ax^n)
= loga+ logx^n
= loga = nlogx
2
Q
y = kb^x
A
log y = log(kb^x)
= logk + logb^x
= logk + xlogb
3
Q
Reason for regression lines
A
to model a linear relationship from an exponential one
4
Q
Product moment correlation coefficient
A
described the linear correlation between two relationships
-1 < r < 1
5
Q
r=1
A
perfect positive linear correlation
6
Q
r= -1
A
perfect negative linear correlation
7
Q
r= 0 (or close to)
A
no linear correlation
8
Q
interpolation
A
calculating inside the data range
9
Q
extrapolation
A
- calculating outside the data range
- not reliable
10
Q
One tailed test
A
H0 : ρ = 0
H1: ρ > or < 0
11
Q
Two tailed test
A
H0 : ρ = 0
H1: ρ ≠0
12
Q
Interpretation of a
A
Constant of proportionality
13
Q
Interpretation of b
A
Proportional rate of which … will change per …
14
Q
Critical value
A
- point on the scale of the test statistic
- beyond which null hypothesis is rejected
15
Q
Hypothesis test
A
- statistical test
- determines whether there is enough evidence
- infer a certain condition is true for entire population
