Week 14

1
Week 14
2
Tonight

• Housekeeping
• Review Ch. 15

3
Housekeeping

• May 1 (next week) is last class
• Homework this week
• A little more SPSS
• Read two articles about Quant methods
• Read last article using methods weve learned

4
Final projects due in two weeks

• In class, youll be presenting your project just
  like I present every week.
• Entire final projects will be posted to the web
  board.
• In class, youll give a short presentation (5-10
  slides).

5
15 minutes per presentation

• There are 6 of you
• The final exam is 2.5 hours
• Lets say 15 minutes per person
• Break after three folks
• And well be done by 8 p.m.
• Thoughts?

6
Brumberger piece

• Included it because she uses the correlations
  youve read about in both texts
• Won Article of the Year when it was published
• Came out of her dissertation

7
Talk more about correlations and factor analysis

• First, let me say that I dont think anyone can
  really understand factor analysis until theyve
  run one
• However, doing the one in the SPSS textbook is
  pretty close, so I hope this is making some kind
  of sense

8
Correlation leading to factor analysis

• Lets review thisyou read about it in SPSS
• Factor analysis starts with correlations
• We reviewed those last week

9
Correlations

• Correlations are simply measures of how much
  variables change together
• Can be positive (both go up) or negative (one
  goes up, one goes down)

10
Factor Analysis

• Uses correlations to see if the relationships
  have a larger underlying relationship, called a
  factor
• When you run an FA, each variable gets a factor
  loading just a fancy word for a correlation on
  the factor
• Looks exactly the same as a correlation

11
Factor Analysis and Variance

• Cool thing about factor analysis
• It provides you with how much each factor
  accounts for the variance
• Fancy way of saying predicts how much of the
  change

12
Brumberger

• So, for example, Brumbergers three factors in
  typefacesfriendliness, elegance, and directness
• Accounted for 96.3 of the variance
• If you knew the three scores for a typeface, you
  can predict the outcomes for all of those scales

13
Wave example from last week

• What factor analysis does is look for waves under
  lots of driftwood, not just two pieces
• The idea is to try to find 3 or 4 waves that
  helps you explain the behavior of all the
  driftwood

14
Like last weeks communication piece

• They used three communication competence scales
  and checked for a relationship with interview
  success
• Found none
• But if they had, they could use those scales to
  predict interview success

15
Another example

• If you were trying to predict student success
• Measure successful students
• Extra curriculars
• Grades
• SATs
• Etc.

16
Conduct an FA

• Perhaps youd find two underlying factors
• Intelligence
• Social experience

17
Brumberger

• Mentions how she cant use age as a variable
  because most of her students were the same age
• And had too many majors to evaluate
• Weve been reading about minimums for statsthis
  is a nice example

18
(No Transcript)