Download Statistics Flowchart 2

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts
no text concepts found
Transcript
State a quantity or proportion
Probability of Observing a Quantity or Proportion
Quantity
Excel functions
standard deviation of quantities  stdev(data)
area to the left of t-score  t.dist(t-score, degrees of freedom, TRUE)
area to the left of z-score = norm.s.dist(z-score, TRUE)
Proportion
Calculate the score
t-score 
stated quantity  sample mean
standard deviation
Calculate the score
z-score 
stated proportion  sample proportion
standard deviation
Find the area to the left of the score
t
The shaded area is the probability of observing a quantity or proportion less than or
equal to the stated quantity or proportion.
State the hypothesis
Hypothesis Test for a Single Proportion
Comparing a sample proportion to a fixed quantity
H0 : population proportion  k
standard error 
HA proportion > k
HA proportion < k
HA ≠
k 1  k 
N
Calculate the score
z-score 
Confidence Interval
p   critical value 
sample proportion  k
standard error
p 1  p 
N
Find the p-value (standard normal distribution)
p-value = 1 – norm.s.dist(z-score, TRUE)
p-value = 2 x (1 – norm.s.dist(abs(z-score), TRUE))
p-value = norm.s.dist(z-score, TRUE)
Excel functions
area left of the score = norm.s.dist(z-score, TRUE)
z-score = norm.s.inv(area left of the score)
State the conclusion
p-value approach: The probability of the null hypothesis being true is the p-value.
Significance approach: If p-value < significance level, reject the null hypothesis.
State the hypothesis
Hypothesis Test for Comparing Proportions
Comparing two sample proportions
HA >
HA <
HA ≠
H0 : population proportion 1 = population proportion 2
standard error 
p1 1  p1  p2 1  p2 

N1
N2
Calculate the score
z-score 
sample proportion 1  sample proportion 2
standard error
Confidence Interval
p1  p2   critical value 
p1 1  p1  p2 1  p2 

N1
N2
Find the p-value (standard normal distribution)
p-value = 1 – norm.s.dist(z-score, TRUE)
p-value = 2 x (1 – norm.s.dist(abs(z-score), TRUE))
p-value = norm.s.dist(z-score, TRUE)
Excel functions
area left of the score = norm.s.dist(z-score, TRUE)
z-score = norm.s.inv(area left of the score)
State the conclusion
p-value approach: The probability of the null hypothesis being true is the p-value.
Significance approach: If p-value < significance level, reject the null hypothesis.
State the hypothesis
Hypothesis Test for a Single Mean
Comparing a sample mean to a fixed quantity
H0 : population mean = k
HA mean > k
HA mean < k
s2
N
degrees of freedom  N  1
HA ≠
standard error 
Calculate the score
t-score 
sample mean  k
standard error
Confidence Interval
s2
N
x   critical value 
Find the p-value (t distribution)
p-value = 1 – t.dist(t-score, df, TRUE)
p-value = 2 x (1 – t.dist(abs(t-score), df, TRUE))
t
p-value = t.dist(t-score, df, TRUE)
Excel functions
area left of the score = t.dist(t-score, degrees of freedom, TRUE)
t-score = t.inv(area left of the score, degrees of freedom)
State the conclusion
p-value approach: The probability of the null hypothesis being true is the p-value.
Significance approach: If p-value < significance level, reject the null hypothesis.
State the hypothesis
Hypothesis Test for Comparing Means
Comparing two sample means
H0 : population mean 1 = population mean 2
standard error 
HA >
HA <
HA ≠
s12 s22

N1 N2
Calculate the score
degrees of freedom  N1  N2  2
t-score 
sample mean 1  sample mean 2
standard error
Confidence Interval
x1  x2   critical value 
s12 s22

N1 N2
Find the p-value (t distribution)
p-value = 1 – t.dist(t-score, df, TRUE)
p-value = 2 x (1 – t.dist(abs(t-score), df, TRUE))
t
Excel functions
area left of the score = t.dist(t-score, degrees of freedom, TRUE)
t-score = t.inv(area left of the score, degrees of freedom)
area to the nearest tail (two-tailed test)  t.test(data set 1, data set 2, 2, 3)
area to the nearest tail (one-tailed test)  t.test(data set 1, data set 2, 1, 3)
p-value = t.dist(t-score, df, TRUE)
State the conclusion
p-value approach: The probability of the null hypothesis being true is the p-value.
Significance approach: If p-value < significance level, reject the null hypothesis.
State the hypothesis
Hypothesis Test for a Single Variance
Comparing a variance to a fixed quantity
H0 : population variance = k
HA variance > k
HA variance < k
HA ≠
degrees of freedom  N  1
Calculate the score
 2 -score 
 N  1  s2
k
Find the p-value (Chi-squared distribution)
p-value = 1 – chi.dist(x-score, df, TRUE)
Excel functions
area left of the score = chisq.dist(score, degrees of freedom, TRUE)
 2 -score = chisq.inv(area left of the score, degrees of freedom)
p-value = 2 x (1 – chi.dist(x-score, df, TRUE))
or 2 x chi.dist(x-score, df, TRUE), whichever
is less
p-value = chi.dist(x-score, df, TRUE)
State the conclusion
p-value approach: The probability of the null hypothesis being true is the p-value.
Significance approach: If p-value < significance level, reject the null hypothesis.
State the hypothesis
Hypothesis Test for Comparing Variances
Comparing two variances
H0 : population variance 1 = population variance 2
HA larger > smaller
HA larger < smaller
One of the sample variances will be larger than the other.
degrees of freedom larger sample variance = Nlarger variance  1
HA ≠
Calculate the score
degrees of freedom smaller sample variance  Nsmaller variance  1
F-score 
2
slarger
variance
2
ssmaller
variance
Find the p-value (F distribution)
p-value = 1 – f.dist(f-score, df, df, TRUE)
Excel functions
area left of the score = f.dist(f-score, df for larger variance, df for smaller variance, TRUE)
F-score = f.inv(area left of the score, df for larger variance, df for smaller variance)
p-value = 2 x (1 – f.dist(f-score, df, df, TRUE))
p-value = f.dist(f-score, df, df, TRUE)
State the conclusion
p-value approach: The probability of the null hypothesis being true is the p-value.
Significance approach: If p-value < significance level, reject the null hypothesis.
Related documents