CATPAC

1
CATPAC WordStat

• Anne D. Sito
• Erin Sonenstein
• COM 633 FA 09

2
CATPAC
3
Overview of CATPAC

• Designed to recognize frequently used words in
  text
• Identifies and groups patterns of similar words
• Provides output of clustering algorithms,
  perceptual maps, and interactive clustering

4
Data Preparation Text
5
1. Convert document into .txt file
6
2. Inputting Data
7
3. Select Text File You Want to Analyze
8
4. Select Make Dendrogram
9
5. Initial Output Screen
10
6. Output Data Screen
11
7. Output Dendrogram
12
8. Data Presented in ThoughtView 2D
13
9. Data Presented in ThoughtView 3D
14
10. Thought View 3D (Rotated)
15
Discussion and Limitations

• s
• Found words like you, youll, and to be
  the most used in this text.
• Examines relationships between words based on
  proximity in the text.
• -s
• Words are measured based on frequency, not
  importance.
• Focuses less on what words mean or how they fit
  together based on dictionaries.

16
WordStat http//www.provalisresearch.com/wordsta
t/wordstat.html
17
Overview of WordStat

• Content Analysis Module for SIMSTAT
• Specifically designed to process textual
  information geared for open-ended data which
  includes journal articles, speeches, electronic
  communication, interviews, etc.
• Has existing dictionary library and can also run
  analyses from new dictionaries built by the user
• Can perform statistical analyses (i.e., factor
  analysis, word frequencies, multiple regression,
  etc.)
• KWIC Key Word In Context tables are available
  for any included or not included word or word
  pattern

18
Data Comparing Reviews of the Book on Amazon.com
Between Men and Women
19
1. Create a Text File
20
2. Input Text File to WordStat
21
3. Define Your Variables
22
4. Running the Analysis
23
5. Existing Dictionary Was Not Relevant for
Our Data
24
6. New Dictionary Available Online!
25
7. (Free) New Dictionary Download
26
8. Import New Dictionary Maintain Exclusion List
27
9. Level 1 Analysis
28
10. Level 2 Analysis
29
11. Overall Frequencies
30
12. Gender Differences
31
13. Dendrogram
32
14. Clustering
33
15. 3-D Figure of Output
34
16. Concurrence Matrix
35
17. KWIC by Gender
36
18. Words by each Text Case
37
19. Word Count Category Frequency
38
20. Aggression Example
39
21. Limitations TerrificAnxiety?
40
Discussion Limitations

• Allows multiple independent variables
• Dictionaries may not always be complete
• Words in .txt file must be be spelled correctly
• Could not distinguish between quotes from the
  book and original thoughts
• May not account for different usage of certain
  words, (e.g., combating, terrific)

41
Any Questions? Thank You!