Two-Stage regression estimation with probit model

Moderators: statman, Analyst Techy, andris, Fierce, GerineL, Smash

jrbrugger
Posts: 1
Joined: Sun Aug 23, 2015 10:37 pm

Two-Stage regression estimation with probit model

Postby jrbrugger » Sun Aug 23, 2015 11:29 pm

Hey everyone.

I am stuck as to how I can attain similar results to this paper:

B.I. Situmeang, F., A.A.M. Leenders, M., & M. Wijnberg, N. (2014). The good, the bad and the variable. European Journal of Marketing, 48(7/8), 1466–1486.

It is using a two-stage regression approach with a first-stage probit model in SPSS.

Now: Here are some questions....

Is there anyone that could tell me which models I would use in SPSS to generate the 1st-stage and 2nd-stage?
In Situmeang (2014) what models in SPSS are they using to get the results in table III and V?
If Table V was via 2-Stage Least Squares in SPSS... how did they get the Betas?
Where would they have entered each of those variables in the SPSS models (covariates, factors, dependent, etc.)?

I am quite a beginner at SPSS and statistics in general... I'm also confused as to where to put my variables. If anyone can help or provide me with a link to a tutorial as to how to do a model like this (a for-dummies version) that would be great.
GerineL
Moderator
Posts: 1477
Joined: Tue Jun 10, 2008 4:50 pm

Re: Two-Stage regression estimation with probit model

Postby GerineL » Mon Aug 24, 2015 8:35 am

They seem to be logistic regression models. By glancing the paper (didn't read it in detail), it looks like they first estimate the probability of a sequal coming out using logistic regressions, then they save those probabilities and enter those as a variable for subsequent analyses for estimations on the sequels.

They refer to a paper by Ho et al (2009), you might want to take a look at that one to see what they do.

Then table III contains the estimations of probability based on these analyses.

Table IV contains means and standard deviations (obtainable via analyze -> descriptives) and correlations (via analyze -> correlate).

This seems like pretty complicated stuff for someone who has no exprience with statistics, I would advise you to find someone to help you with this because it may not be feasable for someone who knows nothing about statistics.

For youtube videos: look at logistic regression. 

Who is online

Users browsing this forum: No registered users and 3 guests

cron