Academia.eduAcademia.edu

Figure 8 – uploaded by Segun Idowu Adeniyi PhD

See full PDFdownloadDownload figure
Note: (1) { } are probability values; Parentheses () are Z-statistic Table 7. Binary regression output for the samples of health, conglomerate and agriculture

Table 7 Note: (1) { } are probability values; Parentheses () are Z-statistic Table 7. Binary regression output for the samples of health, conglomerate and agriculture

Related Figures (7)

Source: Researcher, 2020. Table 1. Operationalisation of variables
Source: Researcher, 2020. Table 1. Operationalisation of variables
Table 2. Descriptive statistics
Table 2. Descriptive statistics
Table 3 above shows the correlation result for the study. AQ was noticed to be positively correlated with BIND (r = 0.128, FOWNER (r = 0.334) and AFSIZE (r = 0.500) for the full sample. For sector sample of the consumer goods, AQ was found to be positively related with FOWNER (r = 0.281, AFSIZE (r = 0.528) while it was negatively correlated with BIND (r = -0.091). All the variables exhibited relationship that is significant with AQ at 5% significant level except BIND. For the sector sample of oil & gas and other natural resources, AQ was correlated positively with FOWNER (r = 0.500) and AFSIZE (r = 0.459). All the other variables exhibit a relationship that is significant with AQ at 5% significance level. Finally, for the sample of the consumer goods sector, AQ was correlated positively with FOWNER (r = 0.219) and AFSIZE (r = 0.518) while it was correlated negatively with BIND (r = Table 3. Correlation results
Table 3 above shows the correlation result for the study. AQ was noticed to be positively correlated with BIND (r = 0.128, FOWNER (r = 0.334) and AFSIZE (r = 0.500) for the full sample. For sector sample of the consumer goods, AQ was found to be positively related with FOWNER (r = 0.281, AFSIZE (r = 0.528) while it was negatively correlated with BIND (r = -0.091). All the variables exhibited relationship that is significant with AQ at 5% significant level except BIND. For the sector sample of oil & gas and other natural resources, AQ was correlated positively with FOWNER (r = 0.500) and AFSIZE (r = 0.459). All the other variables exhibit a relationship that is significant with AQ at 5% significance level. Finally, for the sample of the consumer goods sector, AQ was correlated positively with FOWNER (r = 0.219) and AFSIZE (r = 0.518) while it was correlated negatively with BIND (r = Table 3. Correlation results
Table 4. Binary regression result for the full sample Table 4 above revealed the outcome of the logistic regression for the full sample. logit, probit as well as Gompit are noticed. To choose from the models, the Log Likelihood (LL) was then applied. The consent is that the higher the LL value, the better the outcome. Each of the binary model were specifically estimated: control variable was first excluded; and secondly, control variable (AFSIZE) was included. The result was noticed to be better in every model were control variable (AFSIZE) was included because the different independent variables equally give explanation of more changes in AQ which is the dependent variable. McFadden R-squared result for binary regressions with the inclusion of the control variables is: logit= 37%; probits= 37.3%; and gompit= 38.3%. The LR statistic for the three models indicates that they are 5% significant in explaining the dependent variable outcome. The Likelihood Huber/White Heteroskedasticity test indicates that the results of the binary regression are
Table 4. Binary regression result for the full sample Table 4 above revealed the outcome of the logistic regression for the full sample. logit, probit as well as Gompit are noticed. To choose from the models, the Log Likelihood (LL) was then applied. The consent is that the higher the LL value, the better the outcome. Each of the binary model were specifically estimated: control variable was first excluded; and secondly, control variable (AFSIZE) was included. The result was noticed to be better in every model were control variable (AFSIZE) was included because the different independent variables equally give explanation of more changes in AQ which is the dependent variable. McFadden R-squared result for binary regressions with the inclusion of the control variables is: logit= 37%; probits= 37.3%; and gompit= 38.3%. The LR statistic for the three models indicates that they are 5% significant in explaining the dependent variable outcome. The Likelihood Huber/White Heteroskedasticity test indicates that the results of the binary regression are
Table 5. Results of the binary regression for the samples of consumer goods
Table 5. Results of the binary regression for the samples of consumer goods
Table 6. Binary regression output for the samples of oil & gas and natural resources
Table 6. Binary regression output for the samples of oil & gas and natural resources
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved