Basics of Analysis

1
(No Transcript)
2

• Topic 15
• Basics of Analysis

3
Analysis software

• Spreadsheet
• SPSS
• SAS
• Others

4
Questions

• Put this on a piece of paper
• Day of the week, Todays date
• Your name
• Your signature
• Topic 15

5
Types

• Three basic areas of analysis
• Also look at distribution
• Also how to deal with

6
Deciding which type

• Keep in mind
• 
  (descriptive and relational)
• and
• (real and
  categorical)

7
Descriptions of the sample

• Was the study representative?
• 51 female?
• Fits educational distribution?
• Lots of refusals?
• Lots of no answers?
• Returned mail from one geographic area?

8
Response rate

• Original telephone sample size 600
• Completed interviews 350
• Contacted residence, refused 60
• Businesses and other non-residences 45
• Disconnected numbers 55
• No answer (assumed residences, includes
• answering machine, voice mail, call block) 90
•  
• Legitimate numbers (completerefused
• no answer) 500
•  
• Response rate (350/500) 70

9
Other rates

• Check them out in the book
• Cooperation rate
• Refusal rate
• Contact rate

10
Descriptive analysis

• variable at a time
• Useful for
• giving an overall feel for the population, such
  as what percent own homes
• answering specific descriptive research questions

11
Descriptive example
12
Another example

• Descriptive is simple, often
• More than 500 employees returned the
  reader-response card from the commemorative
  edition, and the feedback was exceptionally
  positive
•         99 said they read most or all of the
  issue
• 98 said they are very or somewhat likely
• to keep this edition

13
Relational analysis

• Compare
• E.g., gender and company image
• Tables can be more complicated because you have
  various combinations of measures each requiring
  different analyses
• two categorical variables use
• one categorical and one real use
• two real variables use

14
Crosstabulation example

• A campaign is evaluated by relating attitudes
  with time period in the campaign

15
Means by subgroups example

• The relation between scale score for image of the
  company comparing image (higher numbers are more
  positive) and gender

16
Correlation

• r can be
• (up to 1)
• (down to -1)
• correlation--as the values of
  one variable so do
  the values of another.
• E.g., as salary goes up so does the image score
  (.32 in the example that follows)

17
Correlation (cont.)

• correlation--as the values of
  one variable increase the values of another
• E.g., as number of years of employment increases
  the image score goes down (-.56 in the example
  that follows)
• correlation--as the values of one
  variable increase the values of another
• E.g., as the number of colleague friends goes up
  image stays the same (.03 in the example that
  follows)

18
Correlation example

• Advantage of correlation is that you can tell
  (e.g., -.56 is
  strongest .03 is weakest)

19
Questions

• Do not use variables I have used in the examples
• 1.
• 2.

20
Distribution

• opposed to
  the overall measure like average
• Same average can have
  distributions
• Newsletter use distribution

21
Distribution (cont.)

• Interpretation varies

22
Verbatims
23
Verbatims (cont.)

• Use care interpreting, e.g., billing blood
  suckers can elicit a gut reaction
• Can alert you to problems for further study,
  e.g., long wait at ER mentioned by one person

24
Verbatims (cont.)

• What if you want categories?
• Calculate

25
Question

• Which of the following is a positive comment,
  which is negative?

26
Now for somethingcompletely different

• QUESTION Critique the following studies
• SpokesPerson-Review Q
• Drug Policy Foundation Q

27
Next time

• Significance
• Explained variance

28
(No Transcript)