**3.2 Model estimation**

Variance Inflation Factors (VIF) value, β coefficient, standard error, OR value, and confidence intervals related to the binary logistic regression model are shown in **Table 2**. Before model estimation, the issue of multicollinearity between variables should be investigated. Variables with a VIF value over five caused mid-level multicollinearity, and variables with a VIF value over 10 caused high multicollinearity [24]. As seen in **Table 2**, no variable in the model has a VIF value of five or above. Accordingly, no variable that causes multicollinearity between variables in the model exists.

According to the binary logistic regression analysis, when OR < 1, the estimated factor (according to the reference category) had little effect on the investigated state. When OR > 1, it had an increasing effect compared to the reference category [25]. As a result of the analysis, compared to the individuals surveyed in 2008, the odds ratio of individuals who participated in the survey in 2014 (OR = 0.770; 95% CI = 0.615–0.964) and 2016 (OR = 0.782; 95% CI = 0.617–0.991), was lower. In addition, men (OR = 2.246; 95% CI = 1.822–2.769) had higher odds of having a work accident than women. Considering the age variable, compared to the 15–24 group, the age ranges of 25–34 (OR = 0.795; 95% CI = 0.611–1.035), 35–44 (OR = 0.601; 95% CI = 0.450–0.803), 45–54 (OR = 0.486; 95% CI = 0.355–0.665), 55–64 (OR = 0.300; 95% CI = 0.197–0.458) and 65+ (OR = 0.296; 95% CI = 0.162–0.542) had a lower odds ratio of experiencing work accidents.

In terms of educational status, it was seen that primary school graduates (OR = 1.714; 95% CI = 1.208–2.434), secondary school graduates (OR = 1.554; 95% CI = 1.087–2.222), and high school graduates (OR = 1.612; 95% CI = 1.160–2.238) had higher odds ratio of work accident than university graduates. When the occupational groups were examined, technicians/assistant professional members (OR = 2.008; 95% CI = 1.225–3.292), service/sales staff (OR = 1.848; 95% CI = 1.189–2.875), qualified agriculture/forestry/aquaculture workers (OR = 3.031; 95% CI = 1.922–4.781), craftsmen/related workers (OR = 4.270;


**139**

for those who did not.

15–24 age range.

**3.3 Average direct elasticity**

99.9, and 121.1% higher than managers.

resulting in injuries in Turkey are provided in **Table 3**.

*Modeling the Factors That Affect Work Accidents with Binary Logistic Regression…*

**Error**

Medium 4.61 −0.242 0.155 0.120 0.785 0.579 1.065

Yes 1.05 0.495 0.140 0.000a 1.641 1.246 2.160

Yes 1.09 0.286 0.084 0.001a 1.331 1.130 1.568

**P OR 95% CI**

**Low. Up.**

95% CI = 2.786–6.542), plant-machine operators/assemblers (OR = 2.774; 95% CI = 1.773–4.340), and those who work in jobs that do not require qualification (OR = 3.459; 95% CI = 2.228–5.370) have higher odds of having a work accident than managers. When general health status is examined, the odds ratio of experiencing work accident of those with very good health (OR = 0.522; 95% CI = 0.386–0.706) is lower than those with poor health status. People who receive psycho-social support/ are depressed (OR = 1.641; 95% CI = 1.246–2.160) had higher odds of having a work accident than others. Finally, the odds ratio of experiencing work accidents for participants who used alcohol (OR = 1.331; 95% CI = 1.130–1.568) was higher than

*Binary logistic regression estimation results of socio-demographic and economic factors that affect whether* 

Average direct elasticities and standard errors in the socio-demographic and economic factors that influence whether individuals experience work accidents

When the education levels are analyzed, primary school graduates, secondary school graduates, and high school graduates are 52.4%, 42.9%, and 46.5% more

When looking at the occupational groups, the probability of technicians, service/sales staff, qualified agricultural workers, craftsmen, plant/machine operators, and those who do not work in qualified jobs is, respectively, 68.5, 60.4, 108.4, 141,

When the general health status is examined, those with very good general health status are 62.4% less likely to have a work accident than those who have poor health. In addition, those who receive psycho-social support/are depressed are 47.6% more likely to have a work accident than other individuals. Those who use alcohol are

likely to have a work accident than university graduates, respectively.

27.6% more likely to have a work accident than those who do not.

For marginal effects, the probability of experiencing work accidents was lower in other years compared to 2008. In terms of gender, the probability of men experiencing work accidents was 78.9% higher than women. Also, as age increased compared to the age range of 15–24, the probability of work accidents decreased. The probability of individuals within the age groups of 25–34, 35–44, 45–54, 55–64, and 65+ are 22%, 49%, 69,6%, 116,8% and 118,2%, respectively, lower than the

*DOI: http://dx.doi.org/10.5772/intechopen.93872*

**Alcohol use (reference category: no)**

*individuals experience work accidents.*

*a p < .01 b p < .05 c p < .10.*

**Table 2.**

**Variables VIF β Std.** 

**Psycho-social support/depression (reference category: no)**

*VIF, variance inflation factor; Std. Error, standard error; Low., lower; Up., upper.*

*Modeling the Factors That Affect Work Accidents with Binary Logistic Regression… DOI: http://dx.doi.org/10.5772/intechopen.93872*

