-- How do we know if a particular thing (cause) has produced some other
thing (the effect)
-- X then Y: drug abuse then delinquency?
-- Does water freeze because it decides to?
-- Do individuals choose a life of delinquency, or is it because their options
are limited?
-- Our free will is actually very limited and our behavior is highly determined
by non-choice forces.
-- When we attempt to establish cause-and-effect relationships, we use a
model of human behavior that assumes we actually have very little freedom
of choice -- the deterministic model.
1. Not all human actions, thoughts, and feelings are determined
2. Causal patterns are not simple, but instead complex
3. We are not all controlled by the same factors
-- Certain factors make delinquency more or less likely to occur
-- it is assumed that behavior is shaped and influenced by outside forces
factors
- An indiographic explanation is very detailed
- Delinquency is caused by: pre-natal care, early childhood nurturing,
SES status, educational opportunities, peer groups, mental health, etc.
- A Nomothetic model of explanation focuses on the most important elements
that cause delinquency.
- Pre-natal care and childhood nurturing may be predisposing factors,
it appears peer group influence is the largest factor
- What we want is the greatest amount of explanation with the smallest
number of causal variables
- This is also known as parsimony: an economy of explanation
is based on Occam's razor
- William of Occam (1837) was a scientific philosopher that
developed the idea that the simplest explanation is the best, or
the simplest of competing theories be preferred to the more
complex
- Also, it was postulated that explanations of unknown
phenomenon by sought first in terms of know quantitities
highly likely, or very unlikely.
- We want to have the greatest explanatory power using the least number of
of considerations - or independent variables.
- If we achieve this condition, the explanation is parsimonious (simple)
that also has high explanatory power
We are attempting to determine all the causes that explain something
Criteria for Causality
-- Maxwell (1996) states the main criteria for accurate explanation are:
1. credibility or believability, is it logical, does it make sense
2. the dismissal of alternative explanations, the ability to dismiss all other
possibilities
1. Cause must precede the effect: X must come before Y
- drug use causes delinquency || not delinquency then drug abuse
2. Two variables must be empirically correlated with each other
- When X changes, Y must change
- If poverty increases, delinquency should increase, likewise, if
poverty decreases, delinquency should decrease
3. The observed relationship between the two variables cannot be explained
by the presence or influence of a third variable that causes both of them
- drug abuse and delinquency is caused by peer influence
- both X and Y are caused by Z
1. the cause precedes the effects in time
2. there is an empirical correlation between them
3. the relationship is not caused by from the effects of some third variable
Necessary and Sufficient Causes
-- A necessary cause is a condition that must be present for the effect to
follow
- for a curfew arrest to occur, you must have a law in effect, it is a
necessary condition to cause a violation
- the mere existence of a curfew law will not cause an arrest
-- A sufficient cause is a condition that guarantees the effect
- a curfew task force performing a curfew crackdown arrests violators
M --- > G
E --> M
E -->G
-- But itself is not necessarily sufficient
1. had to be present for delinquency to develop
2. always results in delinquency
-- a detailed explanation is micromediational
- We don't need to know the bio-physiology of brain chemistry to know
that excessive alcohol consumption produces diminished neuro-motor
response
-- A molar explanation does not build on a detailed account concerning so
many ML of a liquid consisting of a certain percentage of alcohol that
depends of body mass in KG in order to predict physical impairment
-- We rely on molar statements expressed in probabilities.
- These molar statements are sufficient to identify and describe patterns and
tendencies rather than infallible laws
- Is the water causing AIDS?
- 90% of cases come from areas with fluoridated water
- A statistical correlation with no causal link
- Is class attendance related to grades?
- Students with lowest attendance have poorest grades
- Students who sit in the front have higher grades
- What about interest in the subject?
- Good subject interest be the cause of both high attendance and
sitting up front?
- It is easy to make illogical interpretations and commit faulty reasoning
- The protection that we have is in the use of scientific procedure and
reasoning
Validity and Causal Inference
- For a measure, or indicator to be valid, it must accurately represent the concept under study.
intended to do
what it purports to measure
- Validity is a matter of degree, objective and perfect indicators are difficult to
produce
measuring what you purport to measure?
Concept = social disorganization
Indicators = abandoned houses
poverty
high crime
public disorder
trash and litter
broken windows
poor school performance
high unemployment
- 4 general categories
> Statistical conclusion validity
> Internal validity
> Construct validity
> External validity
- The ability to detect, statistically, the change in the cause that is associated
with the change in effect
- Does the cause and effect covary? If X goes up, what happens to Y?
- Poor measures will not detect a change even though there could be a
direct relationship
- Death penalty and deterrence?
- Can we demonstrate a statistical relationship between the two?
- Illegal behaviors are difficult to measure accurately
- If the conclusion is based on a small number of cases, it may very well be inaccurate, the relationship could be by chance alone
- Hurricanes, global warming, reduction in homicides
- There may be a claim of covariation, when in fact non exists
- Claims may be based on false assumptions - lottery.
- In actuality, the process is random
- The association between two variables is not due to another variable
- This occurs from systematic error or nonrandom error
- Example:
Alcohol breath analyzer ----> alcohol level
Internal validity problems: not calibrated properly or
operated improperly
- The outcome measure is reporting high levels of alcohol levels when
in reality there is alcohol present because of a third variable: calibration
error
- How well does what we observe and measure match the real-world things?
- When are we really measuring juvenile delinquency?
- Will the research findings about cause and effect apply equally to other
areas?
- If the juvenile curfew reduced serious juvenile victimization in New Orleans will similar results occur in Orlando?
- Can the findings from one study be reproduced in another under different
conditions?
- If we can demonstrate it in the lab, will it be the same outside of the lab in
the real world?
- If kids dramatically change their behavior in an intensive boot camp setting, what happens when they go back home?
- Group into two categories:
1. Bias
2. Generalizability
- Internal validity = systematic bias
- Statistical conclusion validity = nonsystematic bias (random)
- Generalizability problems result when constructs do not accurately measure
the variables in the relationship, or when the cause/effect relationship in
one setting operates differently in another
- Does drug use cause crime
- temporal problem: which comes first?
- 50% are delinquent before using drugs
- Is it general deviancy that causes both drug use and delinquency?
- Internal validity problem : third variable
- Construct problems because there are so many different patterns of drug
use and delinquency
- A complex phenomenon with no simple single explanation: not a simple
cause and effect relationship
Linking Measurement and Association
1. Theory construction
- Abstraction that describes some aspect of the natural or social world
2. Create theoretical hypotheses
- Best guess concerning the relationships among the component concepts
3. Operationalization of concepts
- Specify empirical indicators that represent the theoretical concepts
that are precisely defined
- This results in an empirical hypothesis
4. Collection of empirical data
5. Empirical testing of hypotheses
- It is almost impossible to produce unambiguous operationalizations of
concepts
- A juvenile may be defined as the chronological age of 18, but we all have
known some that are 20 that behave like juveniles
- A delinquency is not is or is not: it is a continuum.
- Empirical associations are not always perfect
- Not all juvenile delinquents are arrested
- Science is a continual circular process that pursues understanding.
- We theorize, hypothesize, measure, test, and modify the theory only to
repeat the process to move closer to a real-world explanation
The Interchangeability of Indexes
X1 - poverty
X2 - cultural norms
X3 - ethnicity
X4 - educational level
X5 - personality traits
If we find X1, X2, and X5 are the most influential indicators, then we
have improved our understanding of the association among the various
indicators of delinquency.
Then, if that is the case, X1, X2, and X5 become interchangeable indicators
of delinquency
The X1, X2, and X5 measures and not only interchangeable, but have a
greater association with delinquency that X3 and X4