1/41
Looks like no tags are added yet.
Test Statistic (statistic- parameter)/stdv OR written as Z= p-hat - p / sqrt((p(1-p)/n))
Calculating p-value
5C Method for Interference
Classify
Identify and classify variables
Type of study
Choose
Procedure: one sample z test for p
alpha level
Hypotheses: Ho & Ha
Parameter (p=__)
p-hat=
Check
Random
10%
Large Counts
Calculate
Sigma(p-hat)
Z= p-hat - p / sqrt((p(1-p)/n)) OR written as (statistic- parameter)/stdv
p-value (left bound for < or =/) (right bound for >)
Conclude
Assuming Ho is true (p=___) there is a [p-value] probability of getting a p-hat of ___ or [greater/less/ more extreme in either direction] purely by chance
Because [p-value] (>/<) we (fail to reject/ reject) Ho and we (do not/do) have convincing statistical evidence that [context Ha ]
Choose for one sample z test for p
Procedure: one sample z test for p
alpha level
Hypotheses: Ho & Ha
Parameter (p=__)
p-hat=
Calculate for one sample z test for p
Sigma(p-hat)
Z= p-hat - p / sqrt((p(1-p)/n)) OR written as (statistic- parameter)/stdv
p-value (left bound for < or =/) (right bound for >)
Conclusion for one sample z test for p
Assuming Ho is true (p=___) there is a [p-value] probability of getting a p-hat of ___ or [greater/less/ more extreme in either direction] purely by chance
Because [p-value] (>/<) we (fail to reject/ reject) Ho and we (do not/do) have convincing statistical evidence that [context Ha ]
A recent surgery reported that 43% of people use their cell phone in bed before falling asleep. Zevi believes the percentage is greater for students at his school. He will collect data and perform a hypothesis test about p= the true proportion of all students at his school with who use their cell phone in bed before falling asleep. Which of the following are the appropriate hypotheses for the test?
Ho: p=0.43
Ha: p> 0.43
Ho: p = 0.919
Ha: p > 0.919
The p-value of the test was 0.134.
If 91.9% of all front-seat passengers under the age of 40 wear their seatbelt, the probability of obtaining a sample proportion as large or larger than the one obtained by the social scientist is 0.134
An election for state governor has two candidates: the current governor and a challenger. A polling company selects a random sample of voters from the state to test the claim that more than half of them will vote for the current governor at a significant level of 0.01. The test yielded a p-value of 0.19. Assuming all conditions for inference were met, which of the following is the correct conclusion?
We do not reject the null hypothesis. There is not convincing evidence that more than half of all voters from the state will vote for the current governor
p1-p2
True difference in proportions (in context)
Ho for two sample z test for p
p1-p2= 0 (p1=p2)
Ha for Two sample z test for p1-p2
p1-p2 < 0 (p1<p2)
p1-p2 > 0 (p1>p2)
p1-p2 =/ 0 (p1=/p2)
Interpret p-value
Assuming Ho is true (p1-p2= 0), there is a [p-value] probability of getting a p-hat1 - p-hat2 of ____ or [more extreme] purely by chance
Because [p-value] (>/<) [alpha value] we (fail to reject/ reject) Ho and we (do not/do) have convincing statistical evidence that [context Ha ]
At the alpha = ___ level, what conclusion can be made?
Z=

Classify
Variables
Type of Study
Choose for two sample z test for p1-p2
two sample z test for p1- p2 at alpha = ___
p1-p2=
Ho=
Ha=
p-hat1=
p-hat_2=
p-hat_c=
Check
Random
10%
Large Counts
Calculate for two sample z test for p1-p2

p= [using desmos (normaldist(0,1) select bounds based on H_a]
Conclude for two sample z test for p1-p2
Assuming Ho is true (p1-p2= 0), there is a [p-value] probability of getting a p-hat1 - p-hat2 of ____ or [more extreme] purely by chance
Because [p-value] (>/<) [alpha value] we (fail to reject/ reject) Ho and we (do not/do) have convincing statistical evidence that [context Ha ]
P-hat_C is used
What value is used as the probability when checking the large counts condition and when calculating z score?
Less Likely as this is a type I error. The probability of rejecting Ho when it is true is equal to the alpha value. Reducing an alpha value will make a type I error less likely.
If the student were to use 𝛼 = 0.05 instead of 𝛼 = 0.10, would this make it more or less likely to reject the null hypothesis when the null hypothesis is true? Explain.
Power
Ha is true and rejecting Ho
P (reject Ho | Ha is true)
Sentence about Power
If Ha is true (value in context), there is a [power] probability of finding convincing evidence to reject Ho (context)
Increase Power
Increase sample size
Increase alpha
Increase distance to Ha
Type I Error
Rejecting Ho when Ho is true
false positive
make a conclusion that Ha is true in context when Ho is true
p( Type I Error)= alpha
Alpha
p( Type I Error)=
Type II Error
Not making a conclusion that Ha is true even though it was
false negative
p( Type II Error)= 1- power
1- power
p( Type II Error)=
What type of error can be made if you reject the Ho
Type I Error
You fail to reject Ho. There would not be convincing statistical evidence that [Context of Ha]
If the hypothesis test yields a p-value greater than the alpha value, what conclusion would be made?
Type II Error
What type of error can be made if you fail to reject the Ho
You reject Ho. There would be convincing statistical evidence that [Context of Ha]
If the hypothesis test yields a p-value less than the alpha vue, what conclusion would be made?
(# of rows-1)(# of columns-1)
Degrees of freedom
(row total * column total)/ grand total
Expected Counts
Sum[ (observed- expected)2/ expected]
Chi Square
Chisqdist(df); p(x >= z score)
How to get p-value for Chi Square tests
2+ samples and 1 variable
When do you perform chi square test for homogeneity
1 sample, 2 variables, looking for an association
When do you perform chi square test for independence
5C Method for Inference for Chi square for Homogeneity
Classify
Variable
Type of Study
Choose
Chi Square Test for Homogeneity
Ho: no difference in distribution of [context]
Ha: there is a difference in distribution of [context]
Check
Random
10%
Large Counts
Expected Values > 5
Calculate
Chi Square: Sum [(Observed- Expected)2/Expected]
p-value
Conclude
Assuming Ho is true ([context]), there is a [p-value] probability of getting a chi square test statistic of [Sum from Chi Square] or greater
Because [p-value] (>/<) [alpha], we (fail to reject/reject) Ho and we (do/do not) have convincing statistical evidence of a difference in the distribution of [context]
If you reject Ho
The largest component of Chi Square Test Statistic is [Largest (observed-expected)2/expected)] because the count of [context] was greater than expected.
Choose for Chi Square
Chi Square Test for Homogeneity
Ho: no difference in distribution of [context]
Ha: there is a difference in distribution of [context]
All expected values have to be greater than 5
Large Counts for Chi Square
>, fail to reject, do have convincing statistical evidence
Which of these is a correct grouping?