Ordered Logistic Model interpretation. need help.

For econometric discussions not necessarily related to EViews.

Moderators: EViews Gareth, EViews Moderator

Epelikh
Posts: 1
Joined: Tue Jul 29, 2014 5:44 am

Ordered Logistic Model interpretation. need help.

Postby Epelikh » Tue Jul 29, 2014 6:13 am

Hello!

I am currently doing my dissertation that involves some stats. computation.

Background: The topic of my work is to find dependencies between some core skills of individuals ( such as leadership, teamwork etc) and video game genres. For the following I gather data with help of questionnaire and have a sample of 240 individual. I filtered the data to get rid of 'bad' responses and shortened sample to 202 individuals. I have quite a lot of dummy variables such as genre, education etc. The aim of work itself is to

1) analyse the effect of game genres in absolute terms on a skill - how game genres influence the skill value if the time spending per week is increased by 1h;
2) and analysis of the effect in relative terms - how playing specific game genre may increase Skillj relatively to other skills.

I have ranking of each skill from 1 to 8 on each individual by using Ordinary Ranking Method, so I dont have any gaps.
So in other words: I am trying to understand if person plays Shooter games, will it increase Rank of Leadership_skill(or any other skill) from current rank or decrease it?


Basically, I have completed the step1 with OLS.
The step 2, I have calculated Ordered Logistic Regression. I have Beta coefficients as presented below for one of the skills (Strategic Thinking) calculated in Eviews 7:

AG1825 = dummy for age of 18 to 25.
MSTATUSD = dummy for marital status *divorced*
COLH = dummy for collection hobbies
COMH = dummy for competition hobbies
SECONDARY = dummy for education
ACADEMIC, ARTS, Military = dummy for prof sphere *Academic:Biology, Mathematics, Statistics etc"
FR02 = dummy for number of friends from 0 to 2
STEALTH, FIGHTING,MOOGS, Simulators = Variable of Hours Spend Per Week on the following Game Genre

Variable Coefficients:


AG1825 1,04 0
AG2537 1,07
MSTATUSD 1,80
CANADA 1,03
US 0,75
COLH 0,55
COMH 0,66
SECONDARY 1,95
ACADEMIC -1,48
ARTS -1,04
MILITARY 2,97
FR02 -0,82
STEALTH -0,19
FIGHTING 0,21
MMOGS -0,47
SIMULATOR 0,30


Limit Points


LIMIT_2:C(17) -1,57 0,36 -4,35 0,00
LIMIT_3:C(18) -0,44 0,32 -1,37 0,17
LIMIT_4:C(19) 0,84 0,32 2,61 0,01
LIMIT_5:C(20) 1,78 0,34 5,19 0,00
LIMIT_6:C(21) 2,79 0,38 7,42 0,00
LIMIT_7:C(22) 4,05 0,44 9,10 0,00


Pseudo R-squared 0,094     Akaike info criterion 3,65
Schwarz criterion 4,014     Log likelihood -343,14
Hannan-Quinn criter, 3,798     Restr, log likelihood -378,85
LR statistic 71,415     Avg, log likelihood -1,72
Prob(LR statistic) 0,000


Could you help me to interpret coefficients? I do not clearly understand how to interpret the negative signs.
And also what do I need Limit_1, Limit_2 ... for?
I would appreciate it very very much as well if someone also suggest tests that I need to run with Ordered Logistic model in order to make an analysis sound and regression specification valid.

Thank you!

Return to “Econometric Discussions”

Who is online

Users browsing this forum: No registered users and 1 guest