Skip to main content

Table 10 Probit model addressing potential endogeneity (unmatched samples)

From: Can producer groups improve technical efficiency among artisanal shrimpers in Nigeria? A study accounting for observed and unobserved selectivity

  EXTENSION CREDIT
Coeff Coeff
AGE  − 0.02
(0.19)
 − 0.01
(0.01)
EXPEREIENCE 0.04**
(0.02)
0.01
(0.01)
EDUCYEAR 0.05
(0.04)
0.05**
(0.02)
REPEAT 0.20
(0.24)
0.07
(0.16)
FEMLABSHR  − 0.08
(0.12)
0.10
(0.06)
FEMAS 0.76**
(0.32)
0.18
(0.21)
AKWA-IBOM  − 0.64
(0.56)
 − 1.55***
(0.37)
ONDO    − 0.19
(0.20)
MOBILE    − 0.60**
(0.26)
EXTENSION    − 0.47
(0.40)
CREDIT  − 0.69*
(0.40)
 
CUSTOMERS 0.01
(0.08)
 − 0.07
(0.05)
SHOCK  − 0.12
(0.30)
0.80***
(0.20)
TAROAD 0.28
(0.43)
0.65**
(0.30)
LEADER 0.48
(0.38)
 − 0.33
(0.27)
COOP 0.63*
(0.42)
0.70***
(0.25)
UsefulExtension 4.30**
(1.59)
 
Creditworthiness    − 0.71***
(0.18)
Constant  − 3.00***
(0.86)
 − 0.65
(0.54)
Log-likelihood  − 49.59 154.86
LR chi2(14 and 16) 43.03 77.68
Number of obs 353 353
  1. ***p < 0.01, **p < 0.05, *p < 0.1. ONDO and MOBILE were dropped in extension model because they perfectly predicted failure to access extension. Standard errors are presented in parentheses
  2. Source: Authors’ calculation based on survey data