INFO 370 - L09 - October 27, 2004
Notes: Prins, Egaas, Yaptinchay, Fortier


Subject Heading Librarian: Nancy Huling

Kraut Reading
    > Deductive
    > Published in 1998

Population (with social ties)
    > Internet Usage: Parents of child in a H.S. journalism class
    > Internet Usage: Adults was board member of community group
    > No Internet Experience
    > Invited 93 families-- 90 accepted (256 people)-- 73 completed (169 people)
    > Non-random sample

Type of Study: 
    > Casual study
    > Qualitative
    > Quantitative  
    > Non-participant
    > Obtrusive
    > Causual Study

Longitudinal Study
    > March 95, March 96, and then both March 97 (two observations)
    > They really don't differentiate between T1 and T2, however the best was T3

Depedent Variables:
    > Psychological Well-Being
        * Loneliness
        * Stress
        * Depression
    >Social Involvement
        * Social Support
        * Size of Local Network
        * Size of Distant Network
        * Family Communication
        
Independent Variables
    > Usage (ratio)
        * Intra-correlation
            - Hours
            - # of e-mails sent/received
            - # sites of accessed
        
Control
    > Factors
        * Age    
        * Race
        * Gender
        * Income
        * Social Extroversion
    > Hypothetical Factors
        * Occupation
        * Region/Area
        * Education
        
        * Network Externality
    > Why control these factors?

Overall Findings 
    > Negative impact
    > Small but statistically significant

// BEGIN SLIDES \\

Causation
    > Causation is not correlation!
    
Causal Explanation
    > Idiographic - Pertains to description of the single unique facts and processes
        * Explorartory Research
    > Nomothetic - Pertains to the study of general (scientific) laws and principals
        * uh forgot what he just said
    > X->Y
    > Certeris Paribus (All things being equal)
        - Experimental design
        - Non-experimental design

Association
    > Changrs in X coinside with changes in Y (the first condition)
        ^       . ╱.
        │   .  .╱  .
        │  .  ╱.         ASCII R00LS!!
        │ . ╱ .
        │.╱ .   .
        └────────┤
        ex. positive correlation    

Time Order
    > Did X occur before Y?

Nospuriousness
    X -> Y
    
    or
        X
    A--<
        Y

X causes Y or A can cause X and Y

Casual Mechanism & Context
    > Process
        * Must be able to explain the mechanism in which this cause and effect relationship occrus
        * "That's how it works"
        
    > Conditions under which the process occurs

Unit of Analysis (elements on the top are constant in their children)
                 [ Database ]
            ┏━━━━━┻━━━━━━━━━━┓
        [ Interface ]             [ Interface ]
   ┏━━━━━━┻━┓              ┏━━━━┻━┳━━━━━━━┓
[ User ]     [ User ]        [ User ]    [ User ]    [ User ]

Sources of Error
    > Ecological Fallacy
        - Using group level data to describe individual level processes
    > Reductionist Fallacy
        - Using individual level data to describe group level of processes
        - a.k.a. bad generalization
# END #