Summary Lectures Applied Methods and Statistics

Summary: Applied Methods and Statistics

Lecture 1: Basic Elements of Path Analysis


Path Analysis
 Extended version of regression -> combines regressions in larger model
 There are many different hypotheses in one model




-> mediation model

 Can the correlations between a group of variables be explained by a causal model?
 It is more a causality than a regression

Factor Analysis




 Can the correlations between a group of variables be explained by one underlying
construct (or multiple)?

Structural equation modelling
 Factor analysis and path analysis come together

,  Can the correlations between a group of variables be explained by underlying
constructs and causal effects between those?

Goal of path model
The goal of a path model is to capture theories such as Hank Green’s in a formal form, the
“path diagram”, to examine whether the assumed theory corresponds to observed
correlations in reality.

The Basic Elements of a Path Model
 Variables (=squares in the model)
 Relations between variables
 Covariation (correlation)
 Causal effects
 Types of relations/effects
 Direct
 Indirect
 Unknown
 Spurious
 Reciprocal
 Conditional effects

Variables

 Are characteristics of research units (such as people, couples, schools…) that you are
interested in
 Research unit = entity you are studying (also time points)
 There must be variation in the characteristics across research units
 If there is no variation in the variables it is not a variable but a constant
 E.g. for this course’s students is “student/no student” a constant

Common mistakes:

 Confusion of the values of the variable with the variable itself
 “rich” and “poor” are two values of the same variable “income” -> do not include
them in path model separately
 Confusion of a process or theory with a variable
 E.g. attribution theory, math-related career choice process

,Variables:

 Neglecting your care/not taking medication -> self-care
 Stress
 Emotional states
 Lifestyle/behavioral factors
 Biological factors
 Isolation -> social interactions

Relations

 A statement in which:
 Two variables occur
 Higher/lower values on one variable are associated with higher/lower values on the
other variable
 Two types of statements
1. Covariation statement
2. Causal statement
 Causal relation:
 A causal relation means that change in one variable leads to change in the other
 If you change the independent, then also the dependent will change
“Paying attention leads to knowledge” -> if you switch these variables meaning
would change: “Knowledge leads to paying attention”
 Signal words: produce, cause, influence, result in, is more likely to

Spurious relations

(Causal statement 1) “Chocolate causes happiness”

(Causal statement 2) “Chocolate causes longevity”

Therefore also:

(Covariation statement) “Happy people live longer”

General:

(Statement 1) “Variable x causes 𝑦 ”

(Statement 2) “Variable x causes 𝑦 ”

 (Covariation statement) “𝑦 is related to 𝑦 ”

, More ice cream sold more people were attacked by sharks

 Cannot assume that, it is merely related but not a cause -> it is a covariation
 Spurious relation -> actually caused by hot weather




Correlation and causation are NOT the same



Direct and indirect effects

“People characterized by their optimism, happiness, love and positive feelings often live
significantly longer”




(C1) “Feeling blue leads to neglecting self-care”

(C2) “Neglecting self-care leads to bad health”




Valence of emotion:

 Direct effect on self-care
 Indirect effect on health (via mediator self-care)

“Negative emotions are related to health because of the behavioural factor ‘self-care’”