Discussion Overview
The discussion revolves around the relationships between parameters in a simple regression involving three variables: X, Y, and Z. Participants explore the conditions under which the relationship h = b/d holds, particularly in the context of estimated parameters versus actual parameters.
Discussion Character
- Exploratory
- Technical explanation
- Debate/contested
Main Points Raised
- One participant proposes the relationship Y = a + b*X + errorTerm and Z = c + d*X + errorTerm, leading to the question of how h relates to b and d.
- Another participant suggests that if h, b, and d are actual parameter values, then h = b/d can be derived under certain conditions, but this may not hold if they are estimated parameters.
- A third participant questions whether the parameters are known or estimated and asks about the specific distribution or analysis being considered.
- The original poster clarifies that all parameters are estimated and expresses uncertainty about the validity of using h = b/d in this context, acknowledging the dependence on assumptions and estimations.
Areas of Agreement / Disagreement
Participants do not reach a consensus on whether h = b/d can be applied when using estimated parameters, highlighting the complexity and uncertainty involved in the estimation process.
Contextual Notes
Participants note that the validity of the relationship h = b/d may depend on whether all assumptions of the regression are met, which remains unresolved in the discussion.