Statistics 2 Flashcards
(21 cards)
What is the equation for the product moment correlation coefficient?
r = sxy/square root of (sxx)(syy)
What are the conditions for a PMCC?
Population must be bivariate normal
Data from a random sample
Bulk of data should form an ellipse on a scatter diagram
Hypothesis test for PMCC
H0:r=0
H1: r>0, r<0, r not equal 0
if r is bigger than or equal to the critical value reject H0
Spearman’s rank equation
rs = 1 - 6sumd^2/ n(n^2 -1)
Spearman’s hypothesis
H0: no association
H1 : positive association, negative association, some association
If rs is greater than the critical value reject H0
What is PMCC?
measures how close to a straight line the points on a scatter graph lie, if r=0 no correlation
How do you work out line of regression?
y = a + bx b = sxy / sxx a = y bar - b x bar
Equation for residuals
observed y value - expected y value
positive - line is too low
negative - line is too high
What is interpolation and extrapolation?
Interpolation - reliable- within data set
Extrapolation - out of data range - not reliable
Poisson distribution equation
x - p (lamda)
P(X =r) = e^-lamda x lamda^r / r!
Conditions for poisson
Random, independent and at a fixed rate of occurrence
mean is roughly equal to the variance
Normal distribution
X - N(mean , standard deviation squared)
z = x-mean / standard deviation
Normal hypothesis test
H0: mean = k
H1: mean >k, mean
x - N(mean, standard deviation squared/n)
z = sample mean - mean / (standard deviation/square root of n)
Estimating variance
n/n-1 [(sum x^2/n) - (sum x/n)^2)]
Chi-squared equations
expected = row total x column total/ overall total X^2 = (o-e)^2 / e v = (row-1) (column-1)
Chi-squared hypothesis
H0: no association
H1: some association
X^2 is greater than the critical value, reject H0
Approximating
Binomial to poisson
X - B(n,p) —————–> X - P(np)
n is large
p is small
Approximating
Poisson to normal
X - P(lamda) ————–> X - N(lamda, lamda)
lamda is bigger than 10
continuity correction
Approximating
Binomial to normal
X - B(n,p) ————> X - N(np, npq)
n is large
p is roughly 1/2
What is a continuity correction?
make the equal to group 0.5 bigger
What is the assumption for chi squared?
Independent