**2. Material and method**

#### **2.1 Data**

The micro data set obtained from the Turkey Health Survey carried out by the TSI in 2008, 2010, 2012, 2014, and 2016 were used in this study. The Turkey Health Survey, which was first carried out in 2008, had been conducted every two years. The Turkey Health Survey was last conducted in 2016. With this survey, the aim is to minimize the information gap in the current structure by obtaining information for the health indicators that have a large share in the development indicators that show a country's level of development. In addition to being a survey that reflects the whole country, it is also important in enabling both international comparisons and shedding light on national needs. The scope of this survey is households that are located in all settlements within the borders of Turkey. Institutions including soldiers and permanent residents of dormitories, prison, nursing homes, hospitals, etc. are out of this survey's scope as well as locations (i.e. small villages, settlements of nomads, etc.) that are thought to be inadequate in terms of sample size (the number of population less than 20) have been excluded.

This survey was designed to give a total estimate for Turkey. A stratified twostage cluster sampling method was used to obtain the data. The first stage sampling unit was randomly selected blocks from the clusters (blocks) containing an average of 100 household addresses with a proportionate stratification, and the second stage sampling unit was the household addresses systematically and randomly selected from each selected cluster [18–22].

In this study, the data from a total of 35,019 employees over 15 years old were employed, including 5473 people who participated in the Turkey Health Survey in 2008, 5238 people who participated in this survey in 2010, 10,436 people who participated in 2012, 7415 people who participated in 2014, and 6457 people who participated in 2016.

**135**

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

The dependent variable of this study was a work accident of an individual measured by the question, "Have you had an accident that caused injury in the past 12 months?" The dependent variable was a binary variable. In the established binary logistic regression model, the dependent variable was categorized as 1 if the

The independent variables are survey year (2008, 2010, 2012, 2014, 2016), gender (male, female), age group (15–24, 25–34, 35–44, 45–54, 55–64, 65+), education level (did not finish school/illiterate, primary school, secondary school, high school, university), marital status (single, married), work schedule (part-time, full time), profession (managers, professional occupational groups, technicians/ assistant professional occupational groups, staff working in offices, service/sales staff, qualified agricultural/forestry/aquaculture workers, craftsmen/craft-related jobs, plant-machine operators/installers and those who work in jobs that do not require qualification), general health (very good/good, moderate, bad/very bad), psycho-social support/being depressed (no, yes), and alcohol use (no, yes). Ordinal and nominal variables were defined as dummy variables in order to observe the effects of the categories of all variables to be included in binary logistic regression

Survey statistics in Stata 15 (Stata Corporation) were used to account for the complex sampling design and weights. Weighted analysis was performed. First, frequency analyses of the variables in the model were performed. Then, chi-square independence tests were performed in order to detect the relationship between whether individuals had experienced a work accident and socio-economic and demographical factors. Last, factors which influenced the work accident experience

of individuals were determined with binary logistic regression analysis.

accidents resulting in injury in Turkey are presented in **Table 1**.

Socio-demographic and economic factors that are critical in work-related

19.2, 14.5, 28.9, 20, and 17.4% of those who experienced a work accident participated in the survey in 2008, 2010, 2012, 2014, and 2016, respectively. In terms of age range, 15.6% of employees who experienced work accidents were between 15 and 24 years old, 30.3% were between 25 and 34 years old, 29.2% were between 35 and 44 years old, 18.7% were between 45 and 54 years old, 4.5% were between 55 and 64 years old, and 1.8% were 65 years and older. In terms of education level, while 5.6% of workers, who had experienced work accidents were illiterate, 44.6% graduated from primary school, 21.4% were secondary school graduates, 20.3% were high school graduates, and 8.1% were university graduates. For occupational groups, while 3.2% of work accident victims were managers, 5.1% belonged to professional occupational groups, 5% were technicians and assistant members of professional occupations, 2.1% are office staff, 12.7% were service/sale staff, 15% were qualified agricultural/forestry/aquaculture workers, 26.9% were artists and related employees, 11.9% were facility-machinery operators/assemblers, and 18.1% were workers in non-qualified jobs.. While 67% of work accident victims had very good health, 26.3% had medium health, and 6.7%

**3.1 Descriptive statistics and chi-square test**

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

individual had had a work accident and 0 if not.

**2.2 Measures and variables**

model [23].

**3. Results**

**2.3 Research methodology**

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