The variables I have:
Independant: Taking an academic degree.
The Dependant: 6 yearly salaries, three prior to the degree, and three subsequent.
The question is whether the degree had any effect on salary increase.
The test is ARIMA, but I am unable to do this in SPSS
This is to find whether the attainment of education has an impact on the salary earned. According to Box and Tiao (1975) an ARIMA model is to be fitted to the preintervention thirty six monthly salaries. On the basis of this model forecasts are made for the posts intervention period. The difference between these forecasts and their corresponding posts intervention data are modelled to yield the transfer function. I usually do this with eviews.
Sage publishing co. has a book on Interrupted Time Series and their books usually use SPSS. You might start there. The following link may be helpful though I would suggest R might be better software for this project.
https://www.google.com/search?q=Interrupted+time+series+spss&ie=utf-8&oe=utf-8&client=firefox-b-1
in R
https://www.google.com/search?client=firefox-b-1&ei=JJ06W7WEJumijwSN25fACQ&q=Interrupted+time+series+r&oq=Interrupted+time+series+r&gs_l=psy-ab.3..0l9.4543.6931.0.10003.5.5.0.0.0.0.85.388.5.5.0....0...1.1.64.psy-ab..0.5.384...0i67k1j0i22i30k1.0.GHnzQaqxNZo
Best wishes, David Booth
If all you have are effectively 6 annual observations then ARIMA modelling is inappropriate, as would be regression.
What do you suggest Brian?
I have three observations prior to the degree, and three following. Each one is taken yearly.
Do you have panel data for each year, or only an annual average? If the latter I doubt you can draw any credible statistical inferences. If the former, you can pool the data pre and post change and test for significant mean differences (perhaps even for differing individual salary bands pre and post, although you may/may not have a constant sample in terms of individuals).
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS