Lecture Notes: Multivariate Analysis and Elaboration Model

Key Concepts

• Variate and Multivariate

  • Variate means variables.
  • Multivariate means more than two variables.
  • In research, multiple independent variables are used to explain a dependent variable.

• Elaboration Model

  • Explains the relationship between two variables by controlling for a third variable.
  • Important in building models that include several influencing variables.

Case Study: Morale in the Military

• Hypotheses Tested

  1. Promotions affect soldiers' morale.
  2. Black soldiers trained in the North have higher morale than those trained in the South.
  3. Soldiers with more education resent being drafted more than those with less education.
  • None of these hypotheses were supported by the data.

• Concept of Reference Groups and Relative Deprivation

  • People compare themselves to their reference groups rather than objective standards.
  • Relative deprivation occurs when comparisons are unfavorable.

The Role of the Third Variable

• Example: Education and Acceptance of Being Drafted

  • Initial hypothesis: Education affects acceptance of being drafted.
  • Another variable: Whether friends were drafted or not.
  • More educated soldiers accepted being drafted if their friends were also drafted.

• Intervening vs. Antecedent Variables

  • Intervening variables occur between independent and dependent variables.
  • Antecedent variables are unrelated to the causality between the other two variables.

Applications

• Replication and Specification

  • Replication confirms the original relationship under test conditions.
  • Specification identifies different influences or variables affecting the relationship.

• Spurious Relationships

  • Example: Ice cream sales and drowning rates are both related to summer but not to each other.

Interpretation and Explanation

• Interpretation

  • Test/control variable is discovered to be a mediating factor (intervening).

• Example: Children from Broken Homes and Delinquency

  • Supervision, not family status, is the true mediating variable.

Practical Research Examples

• Social Class and Church Involvement

  • Lower social class linked to more church involvement due to lack of secular gratification.

• Criminology

  • Crime rates significantly higher among men than women, unchanged by societal changes.

Research Methodology Insights

• X Post Facto Hypothesis
  • Hypothesis formed after data is collected.
  • Cannot be disconfirmed; importance of starting with a clear hypothesis.

Conclusion

• Importance of multivariate analysis in understanding complex relationships.
• Need for expertise and thorough literature review in conducting accurate research.
• Recognizing and accounting for intervening variables to clarify relationships.