Dear researchers,
I plan to investigate possible causalities of certain psychological constructs and variables like Engagement, Goal attainment, Well-being...
Based on correlation analyses and theoretical considerations, I created a hypothesis system of causal paths (eg. "mindfulness" -> "Autonomy of Goals; "Autonomy of goals" -> "Engagement in the goal" and many more). So that I have now a empirically and theoretically derived hypothesis system of 9 variables. For each of the variables I have a valid and reliable quantitative measurement.
My question is: Is it reasonable to measure all variables just at one point of time to calculate possible causlities and prove my hypothesis system?
Or do I need longitunal data for the variables (e.g. "Change of autonomy of Goals" over a 3 months period) to conduct path analysis?
It would be awesome if you could share your knowledge with me.
Kind regards
Benedict
I get your point. But I read about studies which argue that cross-sectional designs may also be able to support causal effects if there is enough background information about the possible direction of the causation.
E.g. if I measure a person's Intention to behave socially and also measure the person's social behavior at one point in time. Then I could only measure correlations according to your argument. But if I support the causation direction from intentions -> behavior with many studies from literature that show that the main predictor of behavior are a person's intentions wouldn't it be possible to measure the causal effects based on my measured correlations in combination with the background Information from other studies that show that intentions predict behavior to a high degree.
I hope my argument is understandable.
Bene
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling