Fuzzy Set Profiling and Community Analysis Techniques by H9ogoZc

VIEWS: 4 PAGES: 20

									    Fuzzy Set Profiling and
     Community Analysis
         Techniques


      Mark Schafer             Ashley Barras
   Associate Professor         PhD Candidate
   Dept. of Agricultural   Department of Sociology
Economics & Agribusiness           LSU
     LSU AgCenter
       Acknowledgement
• Funded Research
   – Minerals Management Service
   – Cooperative Agreement
      • MO7AC13301
      • Keithly, Fannin, and Schafer 2007-2010
   – Assessing the Impact of OCS Activities on
     Public Infrastructure, Services, and
     Population in Coastal Communities
     Following Hurricanes Katrina and Rita
   – Role: Develop parish profiles
           Talk Outline
1. Fuzzy-set basics
2. Samples of Fuzzy-Set Community
   Profile Elements
3. Fuzzy-set Analysis Walk Through
4. Concluding Thoughts
5. Applications to Impact Analysis
          Fuzzy-set Basics
• A fuzzy-set is a continuous set that has been
  calibrated to indicate degrees of membership
• Based in Set-Theory. (Ragin 2000)
   – Cases may be full or partial members of sets
   – Calibration involves finding meaningful differences
      between cases
• Fuzzy sets are binary and metric at same time
   – Combine categorical and metric assessment in a
      single instrument
                             Fuzzy-Set Plot for 64 :
                              Louisiana Parishes
                                                              Outcome

                             1.20


                             1.00
Fuzzy-set Membership Score




                             0.80


                             0.60


                             0.40


                             0.20


                             0.00
                                    1   4   7   10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64
                                                            Districts (by set membership)
  Community Profiling Using
    Fuzzy-set Concepts
• Parish Profiles development
   – Use Secondary data
   – Develop profile elements
   – Group “like” parishes into one of four
     classifications:
      • Full member of set
      • Partial member, mostly in
      • Partial member, mostly out
      • Not member
               Examples
•   Education Profile Elements:
    1. Adequate School District Performance
    2. Instructional Focus
•   Labor Market Profile Elements:
    1. Low Unemployment
    2. Competitive Per Capita Personal Income
    3. Large Business Presence
Adequate Performance Map


             District Performance
                   Not in the set
                   Partial member, more out than in
                   Partial member, more in than out
                   Full member of set
Instructional Focus Map



             Instruction Expenditure
                   Not in the set
                   Partial member, more out than in
                   Partial member, more in than out
                   Fully in the set
Competitive Per Capita
      Income


              Per Capita Incom e
                   Not in the set
                   Partial member, more out than in
                   Partial member, more in than out
                   Fully in the set
Low Unemployment



         Unemployment
             Not in the set
             Partial member, more out than in
             Partial member, more in than out
             Fully in the set
Large Business Presence


              Large Business Presence
                   Not in the set
                   Partial member, more out than in
                   Partial member, more in than out
                   Fully in the set
   Community Analysis
Techniques using Fuzzy-sets
• Fuzzy-set Qualitative Comparative Analysis
  (FsQCA)
   – Fuzzy-set theory
   – Combinatorial logic
   – Boolean minimization
• Analyzes combinations of case characteristics
   – Whether any single condition or
   – Combination of condtions
   – Is necessary or sufficient
• To produce an outcome
  Steps in Fuzzy-set Analysis
1. Define Outcome
2. Define Conditions
3. Generate Truth Table
   • Linking conditions to outcomes
4. Generate Solution Table
   • Minimization of pathways or
      configurations
   • Necessary and sufficient conditions
   • Identify whether outcome can be
      reached by one or multiple combinations
      pathways
Table 4: Truth Table for Fuzzy-Set Analysis of Poor School
   District Performance in Louisiana
                 Conditions                         Outcome
           Low
                   Comp.      Large           Poor
Instruct   Un-                        N                  Consist
                  Income      Firm            Perf.
           Emp
   1        0        0         0       3        1         .899
   0        0        0         0      12        1         .850
   0        1        0         1       3        0         .754
   0        1        0         0       3        0         .724
   0        1        1         0       7        0         .719
   0        1        0         0      15        0         .680
   1        1        1         0       3        0         .638
   1        1        1         1       5        0         .618
   1        1        0         1       3        0         .605
   0        1        1         1       7        0         .423

   **                                 61
       Solution Table
Table 5: Solution Table for Fuzzy-Set Analysis of Poor
         School District Performance in Louisiana
Causal Combination                Coverage     Consistency
1.                                .344         .847
Not low Unemployment*
Not Competitive Per Capita
Income*
No Large Business Presence
N                                 61
Rows                              9 (of 16 possible)
Frequency threshold               3
Consistency threshold             0.85
    Necessary and Sufficient
          Conditions
• Three Necessary Conditions:
   – High Unemployment
   – Inadequate Per Capita Income
   – No large business presence
• Sufficient Conditions:
   – No single condition is sufficient
   – One set of conditions (pathway) to poor
     district performance
• Note: Instructional focus was neither
  necessary nor sufficient
   – Can be seen in truth table
Table 4: Truth Table for Fuzzy-Set Analysis of Poor School
   District Performance in Louisiana
                 Conditions                         Outcome
           Low
                   Comp.      Large           Poor
Instruct   Un-                        N                  Consist
                  Income      Firm            Perf.
           Emp
   1        0        0         0       3        1         .899
   0        0        0         0      12        1         .850
   0        1        0         1       3        0         .754
   0        1        0         0       3        0         .724
   0        1        1         0       7        0         .719
   0        1        0         0      15        0         .680
   1        1        1         0       3        0         .638
   1        1        1         1       5        0         .618
   1        1        0         1       3        0         .605
   0        1        1         1       7        0         .423

   **                                 61
     Concluding Thoughts
Fuzzy-set analysis is
•   Is concept focused
•   Promotes analysis of complex relationships
•   Can incorporate both quantitative and qualitative data
•   Promotes focus on relevant diversity across
    communities or cases
•   Can be used to explore complexity
•   Can be used to sharpen focus of standard
    econometric techniques
           Impact Analysis
• Sharpen focus by grouping parishes or communities
• Assess necessity and sufficiency of conditions for
  outcomes
• Fuzzy-set analysis can incorporate time dimension
   – Quick recovery period after 2005 hurricanes
   – Increasing (or decreasing) reliance on OCS related
     activity for employment
• May help make use of “limited” data
   -Depends on capacity to link available data to
      conceptual focus

								
To top