While doing research in cross sectional data set, first of all we should estimate OLS (ordinary least square) model then do the post regression diagnostics, based on the presence of problem use the appropriate model specified in the table below.
Diagnostic  Source  Model / Solution  
Cross Sectional Data  Non Normal Residuals
(Jarque Bera Test) 
Too many outliers (Kurtosis >> 3)  Use Logarithm Model 
Dependent is count variable  Use Poisson Regression  
Multicollinearity (Variance Infaltion Factor) 
Sample Too small  Use single variable regression  
Variables too many  Stepwise regression  
Indirect independent variables  Indexed variables using Principle Factor analysis  
One independent is catalysis for another  Moderator approach  
Few independent variables  Ridge Regression  
Hetroskedasticity (Breusch Pegan Test) 
Nonlinear variances  Use Logarithm Model  
Dependent is binomial dummy  Logit / Probit model  
Dependent is multinomial dummy  Multinomial logit model  
MisSpecification
(Ramsey RESET test) 
Nonlinear form missing  Incorporate nonlinear form or
Use Logarithm Model 

Instability (CUSUM and CUSUMsq)  Data might have two or more different qualities  Use independent dummy variables  
Autocorrelation
(Durbin Watson) 
Some quality is interconnecting the cross sectional residuals  Add more variables or use bootstrap approach or
use robust regression 

Endogeneity
(Hausman Wu test) 
Valid Regression is other way around  Reverse the regression or use IV / GMM regression  
Contemporaneous Correlation / simultaneity  Regression is two way  Use SEM  
Two equations are interrelated  Use SEM 
By practice researcher can directly used the appropriate model as by construction he knows what kind of problem is there.
Note: This chart is underdevelopment phase and open for suggestions and amendments.
Advertisements