Transcript Review of Econ424

Review of Econ424
Fall 2007
Format of Final Exam
– open book
– understand the concepts
– use them in real examples
– Dec. 14, 8am-12pm, Plant Sciences 1129
– Vote
• Option 1(2) Option 2(3) Option 3 (4)
• Decision: Option 3 for the Excel Part, which means Do not
turn in Excel file, the hard copy will ask more detailed
questions to incorporate step-to-step calculation
Course evaluation
• Course Evaluation
www.courseevalum.umd.edu
• Teaching theater evaluation
www.oit.umd.edu/tt/st_fdbck.htm
Concepts to grasp (1)
• Population / sample
• Population
– Cdf (prob(var<x))
– Pdf (first derivative of cdf)
– population mean, population std. dev.
• Sample
– Histogram, quartiles, percentiles, sample mean, sample std. dev.
• Population  sample
– Central limit theorem xbar~N(µ, σ/sqrt(n))
• Sample  Population
– Xbar is a proxy of µ with noise
Concepts to grasp (2)
• Inference
– Type I error, Type II error
– Confidence level α
– Confidence interval
– Hypothesis testing
•
•
•
•
H0
H1
Accept/reject?
One-tail, two-tail test
Technical stuff
• Excel – midterm review
• SAS – notes, old exams
Summary of Excel (1)
• Basic excel
– open, save and close files
– cut, paste and paste special
– change format for cell, row or columns
– sort data by one or two variables
– chart wizard
– freeze panes
– drag cells
– use excel functions
Summary of Excel (2)
• Data description
– mean, median, trimmed mean
– standard deviation, variance
– quartiles
– mode, skewness, kurtosis
– histogram (absolute frequency)
– relative frequency polygon
Summary of Excel (3)
• Probability theory
– PDF, CDF
– mean and standard deviation
– bernoulli, binomial
– uniform, normal
– how to simulate them in Excel?
– Central limit theorem
– how to see central limit theorem in excel?
Summary of Excel (4)
• Estimation and Hypothesis testing
–
–
–
–
–
–
–
–
–
use sample mean to estimate population mean
confidence interval
type I error and type II error
null hypothesis (H0) and alternative hypothesis (H1)
one-tail vs. two-tail
t-statistics, critical value, p-value
one-sample test
two-sample test (independent)
two-sample test (matched pair)
Summary of Excel (5)
• Linear regression
– model
• one variable on the right hand side
• more than one variables on the right hand side
• create and use binary variables
– fit of the model
•
•
•
•
R square
F test
scatter plot
correlation coefficient
– coefficient estimates
• point estimate
• hypothesis testing
• omitted variable bias
Summary of SAS (1)
• Why do we need Excel and SAS?
– What is the advantage of SAS?
– What is the advantage of Excel?
• .sas, .log, .lst
– How to edit, save, and run .sas in your machine? What
commands need change?
– How to generate and read .log in your machine?
– How to generate and read .lst in your machine?
– How to define library? What does “work” library mean?
– How to find and use datasets in your library?
• Data newdata; set mydata; …; run;
• Proc … data=mydata; ..; run;
Summary of SAS (2)
• How to generate summary statistics in SAS?
–
–
–
–
–
–
Proc means (for the full sample, or by groups?)
Proc univariate
Proc means with output written in a data file
Proc freq
Proc chart
Proc plot
• How to conduct mean comparison?
– Two groups
– More than two groups
Summary of SAS (3)
• How to run and read regressions in SAS?
–
–
–
–
Proc reg
Proc glm
Regressions with fixed effects?
Compare different regressions?
Final words
• Warning #1:
– “Now I can use fancy and sophisticated
statistics everywhere!”
– Excel and SAS are tools that may be useful for
your research question. Their usage should be
driven by your research question, not the other
way around.
Final words
• Warning #2:
– “Now I am a master of statistics!”
– Materials taught in this class are at most a
starting point for future learning and application
of statistics.
– Be aware of the limitations of basic statistics.
For example, a typical OLS regression requires
a set of strong assumptions. Every time when
you apply an OLS regression, think hard why
you choose to run the regression in this way.
Final words
• Is economic statistics an art or a science?
– there might be multiple interpretations for a
simple statistics. Be aware of how the numbers
are created and what assumptions have been
made between the pure numbers and their
economic meanings.
– Some answers are definitely wrong, especially
those that jump to the conclusion!