ie27_16_hypothesistesting
DESCRIPTION
UP Diliman BS Industrial Engineering IE 27TRANSCRIPT
Please be prepared to have exercises every class meeting
This means bringing your formula sheets, statistics tables and calculators EVERY MEETING
Exercises every class meeting will be offsetting quizzes for this part of IE 27
HypothesisTesting
Parameter Estimation
We are guessing the true value of a
parameter using a test statistic either by
giving an exact value (Point Estimation) or a
range of values (Confidence Intervals)
Hypothesis Testing
Given an initial guess of the true value of a
parameter we will test whether that guess is
correct
HypothesisTesting
The uses of Hypothesis Testing
Whether to accept a statistical hypothesis
To test the statistical significance of a certain value
HypothesisTesting
You are given the Third exam scores of 10 randomly selected IE 27 students. Test if the mean score for the third long exam is
eqaul to the passing rate (μ = 56)
The average exam score (Xbar) is a random variable
The average exam score (Xbar) is also a sample statistic
In hypothesis testing we also call a sample statistic as a TEST STATISTIC as it will be used to test the hypothesis
Our Hypothesis (Ho) is that the mean score of 46 students is equal than 56 (Ho: μ = 56)
HypothesisTesting
You are given the Third exam scores of 10 randomly selected IE 27 students. Test if the mean score for the third long exam is
passing (μ = 56)
In Hypothesis Testing we create two hyoptheses
Null Hypothesis
Ho: μ = 56
Alternative Hypothesis
H1: μ ≠ 56
We are sure that only ONE of these is true
HypothesisTesting
You are given the Third exam scores of 10 randomly selected IE 27 students. Test if the mean score for the third long exam is
passing (μ = 56)
Null Hypothesis
Ho: μ = 56
Alternative Hypothesis
H1: μ ≠ 56
In Hypothesis Testing we either, REJECT THE NULL HYPOTHESIS or FAIL TO REJECT THE NULL HYPOTHESIS at a certain level of
significance
We do not say that we “Accept the Null Hypothesis” because at a certain level of significance it may lead us to reject the null
hypothesis
REJECTION IS SUCH A STRONG CONCLUSION…at alam naman nating masakit mareject
HypothesisTesting
You are given the Third exam scores of 10 randomly selected IE 27 students. Test if the mean score for the third long exam is
passing (μ = 56)
Null Hypothesis
Ho: μ = 56
Alternative Hypothesis
H1: μ ≠ 56
In hypothesis testing, a rejection region is computed and the test statistic is compared
HypothesisTesting
Type I Error (Alpha Error)Rejecting the null hypothesis when the null
hypothesis is actually true
Type II Error (Beta Error)
Accepting the null hypothesis when the null hypothesis is actually fallse
HypothesisTesting
ARTBAF
lpha
eject when
rue
eta
ccept when
alse
DecisionHo is True
Ho is False
Reject HoType I Error no error
Fail to reject Ho
no error Type II Error
HypothesisTesting
Decision Mr./Ms. Right Mr./Ms. Right Now
Pinakawalan mo Type I Error no error
Kayo na Forever no error Type II Error
Decision Student is good Student is bad
Fail a student Type I Error no error
Pass a student no error Type II Error
HypothesisTesting
Type I Error
Suppose this is the distribution of the true population mean μ = 60
Suppose we set that α = 5%, thus the corresponding critical region is as follows
In hypothesis testing, a is set by the analyst
HypothesisTesting
Type II Error
Suppose this is the distribution of the true population mean μ = 63
Suppose the critical values are 57.5 and 62.5, we do not reject Ho if the sample statistic is within those limits
But Ho states that μ = 60
HypothesisTesting
Analyzing Type I & II Error
Unlike Alpha, Beta is dependent on the true mean of the parameter
Since the probability of not rejecting the null hypothesis is not set by the analyst but a function of both the sample size and the critical values, the acceptance of the null hypothesis is a
weak conclusion
HypothesisTesting
Analyzing Type I & II Error
If the critical region becomes smaller, the probability of Type I Error, alpha, reduces
HypothesisTesting
Analyzing Type I & II Error
If the true value of the parameter approaches the value hypothesized in the null hypothesis, the probability of Type II
Error, beta, increases
If the difference between the true mean and hypothesized value increases, the probability of Type II Error, beta, decreases
HypothesisTesting
Analyzing Type I & II Error
If we increase sample size, the sample distribution tends to approach the population distribution
If we increase sample size and retain confidence level, both alpha and beta reduces (more accurate, less error)
HypothesisTesting
Analyzing Type I & II Error
If only one of the null and alternative hypothesis is correct, it means the error associated with them are related
If sample size is constant, an increase in the probability of one type of error results in a decrease in the probability of
the other
HypothesisTesting
Power of a Statistical testCorrectly rejecting the null hypothesis when
it is false
Power = 1 – β
A statistical test with higher power is desired. Increasing sample size or increasing
alpha will result to a higher powered test
Power is the probability to “detect” whether a hypothesized value is really far from the
true value
HypothesisTesting
General Procedure for Hypothesis testing
Determine the Parameter of InterestStep 1:
State the Null HypothesisStep 2:
State the Alternative HypothesisStep 3:
Determine AlphaStep 4:
Determine Test StatisticStep 5:
Determine Rejection RegionStep 6:
Compute Test StatisticStep 7:
ConcludeStep 8: