Statistics for Decision Making

1
Statistics for Decision Making
QM 2113 -- Spring 2003

• Wrapping Up the Semester (b)

Instructor John Seydel, Ph.D.
2
Agenda

• Address questions regarding SPSS
• Description
• Categorical data
• Numeric data
• Charts scatterplot, bar chart, histogram
• Regression
• Crosstabs
• Hypothesis testing and confidence intervals
• Discuss review exam
• Revisit final exam from last year
• Discuss broader view of regression analysis

3
SPSS Questions

• May have another in-class exercise session, so
  you do need to be comfortable with SPSS
• SPSS approaches
• Corresponding Excel procedures
• SPSS toolbox
• Data preparation descriptions, selecting cases,
  recoding variables
• Description
• Categorical data (Analyze Descriptive
  Statistics Frequencies)
• Numeric data (Analyze Descriptive Statistics
  Descriptives)
• Charts
• Include scatterplot, bar chart, histogram
• Should be able to edit results
• Regression (Analyze Regression Linear)
• Crosstabs (Analyze Descriptive Statistics
  Crosstabs)
• Hypothesis testing and confidence intervals

4
Exam Solutions

• First, a quick review
• Overview of inference
• Summary of important measures
• Hypothesis testing process
• Review exam
• Comments
• Look at some, but not all, problems
• Final exam (from last year)
• Look at
• Problem 1(a,c)
• Problem 3
• Problem 5 discuss what normally distributed
  means
• SPSS approach as appropriate

5
Important Regression Applications

• First, consider
• The general model
• The hypotheses
• General model (extension of bivariate model)
• Population or process
• y b0 b1x1 b2x2 b3x3 . . . e
• Estimated by
• y b0 b1x1 b2x2 b3x3 . . .
• Null hypothesis is no relationship(s) exist(s)
• Model ? H0 All bi 0
• Each factor (xi) ? H0 bi 0
• Significance values (p-values) are given by
  regression output
• Extensions of bivariate regression
• Multiple regression
• Time series (simply a bivariate model treating
  time as a factor)

6
Lets Look at Some Sports Data

• How about baseball?
• Get data from www.sportingnews.com
• Use Excels data import feature
• On the Web
• Go to home page
• Select STL (or other team) from MLB drop-down box
  (Teams)
• Click on Stats link
• Copy URL from address box
• Import the data into Excel . . .
• Repeat for other teams in the NL Central division

7
Working with Imported Data in Excel

• Overview
• Determine URL for where data table is located
• Import the data
• Reformat as needed
• Import procedure
• Click on Data Import External Data New Web
  Query
• Paste URL into Address box (Ctrl-v)
• Click on Go button and then Import (dont select
  any tables)
• Click on OK and wait while data are imported
• Reformatting
• All HTML tables on entire page are imported
• Need to delete rows and columns that arent
  needed
• Remaining data will likely need to be reformatted
  for better display

8
Some Other Nifty Excel Imports

• Enter any recognized stock symbol
• Note the smart tag
• Can click on the smart tag and import stock info
• Alternatively, use existing web query
• MSN stock quotes
• MSN currency rates
• Custom query (easy to create)
• Process
• Data Import External Data Import Data
• (Need to provide stock symbols)
• Still need to do some formatting
• Can set so updating is dynamic

9
Final Exam Info

• Plan on meeting during the final exam time
• Thursday _at_ 1230-230
• BU 205A (computers are working better)
• In-class exercise
• Primarily working with Excel rather than SPSS
• Miscellaneous analyses
• Importing data
• Additional features of PivotTables
• Time series analysis
• Other . . . ?

10
Appendix
11
Inference Schematic View
Population or Process
Sample
Statistic
Parameter

• Inferences

12
Sampling Distributions (Reminder)
13
Hypothesis Testing Process

• Identify the parameter of concern (m, s, p, b1)
• Define hypotheses
• Null (H0)
• What well believe until proven otherwise
• We state this first if were seeing if
  somethings changed
• Alternate (HA)
• Opposite of H0
• If were trying to prove something, we state it
  as HA and start with this, not the null
• Then state your willingness to make the wrong
  conclusion (a)
• Draw a sketch of the sampling distribution
• Determine the decision rule (DR)
• Gather data and compare results to the DR