RobertNiles.com
About Robert
Reporting Help
Finding Data on the Internet
Questions for Candidates
Stats Lessons
Mean
Median
Percent
Per capita
Standard Deviation
Margin of Error
Data Analysis
Sample Sizes
Stats Tests
Help Board
Bookstore


The Cartoon Guide to Statistics

This isn't some dumb-downed whitewash. It's a must-read for a beginner taking his or her first steps toward understanding stats.
More information
-->

Home > Statistics Every Writer Should Know > The Stats Board > Discusssion

Calculating Path Coefficients
Message posted by Mike Beers on October 21, 1999 at 12:00 AM (ET)

I'm trying to update a status attainment model that evaluates the direct and indirect influences of Father's Education, Father's Occupation, Son's Education, and Son's First Job on Son's Occupation. I have data from the the General Social Survey.

Here are the paths:
Father's Ed to Father's Occ (and vice versa)
Father's Ed to Son's Ed
Father's Occ to Son's First Job
Father's Occ to Son's Ed
Father's Occ to Son's Occ
Son's Ed to Son's First Job
Son's Ed to Son's Occ
Son's First Job to Son's Occ

Father's Occ and Ed are exogenous. This leaves me with three reduced form equations.

My problem is that after I run the first regression of Son's Ed on Father's Ed, I don't know how to proceed any further. In the next equation, where I solve for Son's First Job, the equation is something like:
FirstJob =(b1 + a2b2) Dad'sOcc + b2a1 Dad'sEd + b2e1 + e2
where b1 is the path coefficient from Father's Occ to FirstJob, a2 from Father's Occ to Son's Ed, b2 from Son's Ed to FirstJob, and a1 from Father's Ed to Son's Ed.

If anyone is still with me--my writing can be a bit opaque--I have values for a1, a2, and e1. What's the next step?

Thanks very much.

Mike Beers


READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: Calculating Path Coefficients
Message posted by Bill on October 21, 1999 at 12:00 AM (ET)

Could you tell us what path coefficients are and what is a status attainment model. I can't start to answer your question but I'm curious about your field. Is this something particular to sociology?


Re: Calculating Path Coefficients
Message posted by jg on October 21, 1999 at 12:00 AM (ET)

If father's ed and occ are exogenous, why are you finding the relationship between them ?


Re: Calculating Path Coefficients
Message posted by jg on October 21, 1999 at 12:00 AM (ET)

Path coefficients are an extension of linear regression. Path analysis is usually an extension of multiple regression and widely used in the social sciences.


Re: Status Attainment Model & Path Analysis
Message posted by Mike Beers on October 22, 1999 at 12:00 AM (ET)

The status attainment model is a sociological term or convention for understanding how people get what types of jobs and places in society, especially with respect to their background.

Let's say you want to understand why some people end as investment bankers and other people end up as shoe shiners. The logic would be as follows: For a given individual, get information on his occupation, his education, his parents' education, his parents' occupations, and, in this case, the individual's first job. Clearly, there is a relationship between an individual's education and his occupation; so too, however, will parents' occupations affect an individual's occupation. Moreover, and this is the key to path analysis, parents' occupations also affect the individual's education. So, you have parents' occupations both having a direct effect on the individual's occupation as well as an indirect effect on it, through education.

If you draw a path diagram

IndEduc
^ -
- -
- >
IndOcc
ParentOccs ------------>

The path coefficients are those that represent the relationship between each variable.

Mike



Your $5 contribution helps cover part the $500 annual cost of keeping this site online.

Niles Online Sites:RobertNiles.comTheme Park InsiderViolinist.com

RobertNiles.com™, the site, content and services 咀opyright 1996-2002, Robert Niles.
All rights reserved. Questions? Comments? Read my Privacy Policy, or E-mail me!