Business
Statistics
Exam II
Wednesday May 13
8:00 pm
You will need a calculator for this exam. I will provide normal and t tables, as well as a list of useful formulas.
I. Ethics
A. Honesty
1. don't falsify data
2. don't intentionally bias results
3. don't suppress findings
B. Data collection
1. do no harm
2. manage risks to protect subjects
3. informed consent
4. confidentiality
C. Institutional Review Board (IRB)
1. protect human subjects
2. approval required for research
II. Estimation
A. sample statistics are estimators for population parameters
1. imperfect estimates
2. random error
B. sampling errors
1. standard error
a. normally distributed
i) central limit theorem
ii) known
variance
b. calculations
i) mean
ii)
proportion
2. estimate intervals
3. increasing sample size increases
accuracy
C. confidence intervals
1. calculating interval endpoints
a. normal distribution
i) large samples, or known variance
ii) find z
values in normal table
iii)
interval = mean plus/minus z standard errors
b. t-distribution
i) small samples, estimated variance
ii) find t
values (n-1) in t-table
iii)
interval = mean plus/minus t standard errors
2. precision/ accuracy trade-off
D. Small populations
a. possible to sample large fraction
b. standard error is smaller
c. finite population multiplier
III. Hypothesis Testing
A. answering questions about populations based on sample results
1. base hypothesis on an idea
2. uncertainty; use probability
B. Inference
1. specify null and alternative hypotheses
2. specify confidence level
3. calculate critical region
a. standard error
b. normal table or t-table
4. compare sample statistic to critical value
a. reject null hypothesis
b. don't reject null
5. testing
a. two-tailed test
b. one tailed test
C. Comparing groups
1. difference of means
2. difference of proportions
D. Regression
1. typical test is there a
statistically significant relation between Y and X ?
2. null hypothesis b = 0
3. use standard error of the
coefficient to construct critical region
4. reject if b is outside critical
region
5. multiple regression
E. Significance versus Power of a test [if time]
1. Type I error
2. Type II error