Hello,
I wanted to regress several items on migration on to attitudes (X) towards the welfare state (Y). The data is cumulative from 1980 to 2016 but the two sets (X, Y) were measured at different time points. F.e. the X items are available in 1996, 2006 and 2016 and the Y in 1990, 2000, 2002 and 2012.
Does someone know about a possibility to do this? I already thought about a sem model or a mixed model with a year dummy but it doesn't work because the data points X are missing for Y. Thanks in advance,
Luise
Could you run an autoregressive/cross-lagged panel model via path analysis? In such an analysis, you can simultaneously consider several independent and dependent variables measured at different time points.
First, you need to build your model and then use appropriate statistical tools with missing data.
Saludos Luise. Te recomiendo que consideres dos caminos: El proceso cuantitativo matemático construyendo las matrices y resolviendo las ecuaciones diferenciales para "rellenar" los espacios vacíos en tu serie de tiempo no uniforme; tranquila, te estoy diciendo que SPLINE CUBICO lo resuelve, pero solo llega a la intrapolacion.
La segunda alternativa es que utilices RISK SIMULATOR (Simulador estadístico) que procesa en el entorno de Excel miles de datos, y aquí simplemente accedes al modelo Spline Cubico y realizas con mucha facilidad la intrapolacion y extrapolación.
Cuando ya tengas la data completa pronosticada Y* entonces realizas la correlación simple, correlación múltiple o la que explique mejor la distribución que estés estudiando.
Hello,
sorry for the late response... For the moment I simply omitted the "missing" things and worked on the available items. F.e. for question 1 I took years Z with variables X available. But a "full analysis" would be better though. Christian Geiser do you have some good literature at hand about that. I tried to build sems but the fit measures were not that great (or just bad so that I wasn't sure if I'm telling anything). Then I switched to mediation and moderation analysis.
Best, Luise
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression