86-88

Causal relationship between two variables must meet three criteria:
1. Empirical correlation exists.
2. Cause precedes effect in time.
3. Relationship is not due to a third variable's influence.
Only relationships satisfying all criteria are considered causal.

Necessary Causes:
- Conditions that must be present for an effect to occur.
- Example: A person must be charged with a criminal offense before they can be convicted, but being charged alone does not guarantee a conviction (e.g., acquittal or being found not guilty).
Sufficient Causes:
- Conditions that can guarantee the effect.
- Example: Pleading guilty to a criminal charge is sufficient to ensure a conviction, but there are other routes (trials) that can also lead to convictions.

Discovering a cause that is both necessary and sufficient is ideal but rare in research.
- E.g., Identifying a single condition that both leads to juvenile delinquency and always results in delinquency is challenging.
Most causal relationships in criminal justice are probabilistic, only partially explaining cause and effect.

Importance of assessing the truth of causal statements.
Validity involves examining threats that could undermine cause-effect relationships, crucial for criminal justice research.

Validity: Approximate truth of an inference.
Validity Threats: Possible sources of false conclusions about causal relationships. Researchers strive to mitigate these threats.

Ability to determine if a change in the suspected cause correlates with a change in the suspected effect.
If there’s no statistical relationship observed (e.g., drug use and crime rates being equal), we cannot infer a causal link.
Small sample sizes can threaten this validity.
- Example: Comparing arrest rates between a small number of drug users and non-users might yield insignificant results.

Concerns whether an observed relationship between two variables is genuinely causal.
Example: Observing that drug users sentenced to probation are rearrested less frequently compared to those sentenced to prison may overlook prior criminal records as a confounding variable causally influencing outcomes.
Internal validity is compromised by nonrandom or systematic error rather than random error.

Explores whether findings from one study are applicable to other contexts or populations.
- Example: Do research findings in Minneapolis regarding mandatory arrest for family violence hold true in other cities like Milwaukee?
Variation in crime issues and justice responses across locales highlights the importance of external validity in criminal justice research.