*3.2.2 Secondary data*

Desk review has been conducted to collect data from various secondary sources. This includes reports and project documents at each manufacturing sectors (more on medium and large level). Secondary data sources have been obtained from literatures regarding OSH, and the remaining data were from the companies' manuals, reports, and some management documents which were included under the desk review. Reputable journals, books, different articles, periodicals, proceedings, magazines, newsletters, newspapers, websites, and other sources were considered

on the manufacturing industrial sectors. The data also obtained from the existing working documents, manuals, procedures, reports, statistical data, policies, regulations, and standards were taken into account for the review.

In general, for this research study, the desk review has been completed to this end, and it had been polished and modified upon manuals and documents obtained from the selected companies.

#### **4. Population and sample size**

#### **4.1 Population**

The study population consisted of manufacturing industries' employees in Addis Ababa city and around as there are more representative manufacturing industrial clusters found. To select representative manufacturing industrial sector population, the types of the industries expected were more potential to accidents based on random and purposive sampling considered. The population of data was from textile, leather, metal, chemicals, and food manufacturing industries. A total of 189 sample sizes of industries responded to the questionnaire survey from the priority areas of the government. Random sample sizes and disproportionate methods were used, and 80 from wood, metal, and iron works; 30 from food, beverage, and tobacco products; 50 from leather, textile, and garments; 20 from chemical and chemical products; and 9 from other remaining 9 clusters of manufacturing industries responded.

#### **4.2 Questionnaire sample size determination**

A simple random sampling and purposive sampling methods were used to select the representative manufacturing industries and respondents for the study. The simple random sampling ensures that each member of the population has an equal chance for the selection or the chance of getting a response which can be more than equal to the chance depending on the data analysis justification. Sample size determination procedure was used to get optimum and reasonable information. In this study, both probability (simple random sampling) and nonprobability (convenience, quota, purposive, and judgmental) sampling methods were used as the nature of the industries are varied. This is because of the characteristics of data sources which permitted the researchers to follow the multi-methods. This helps the analysis to triangulate the data obtained and increase the reliability of the research outcome and its decision. The companies' establishment time and its engagement in operation, the number of employees and the proportion it has, the owner types (government and private), type of manufacturing industry/production, types of resource used at work, and the location it is found in the city and around were some of the criteria for the selections.

The determination of the sample size was adopted from Daniel [5] and Cochran [6] formula. The formula used was for unknown population size Eq. (1) and is given as

$$m = \frac{\omega^2 \mathcal{P}(1-\mathcal{P})}{\mathbf{d}^2} \tag{1}$$

**31**

*Research Design and Methodology*

*DOI: http://dx.doi.org/10.5772/intechopen.85731*

Bureau of Labour and Social Affair.

**5.1 Primary data collection methods**

*5.1.1 Workplace site observation data collection*

*5.1.2 Data collection through interview*

**5. Data collection methods**

be satisfactory and representative for the data analysis.

are devised and prepared with their proper procedures.

elaborate how the data were obtained from the primary sources.

**4.3 Workplace site exposure measurement sample determination**

The expected sample number was 267 at the marginal error of 6% for 95% confidence interval of manufacturing industries. However, the collected data indicated that only 189 populations were used for the analysis after rejecting some data having more missing values in the responses from the industries. Hence, the actual data collection resulted in 71% response rate. The 267 population were assumed to

The sample size for the experimental exposure measurements of physical work environment has been considered based on the physical data prepared for questionnaires and respondents. The response of positive were considered for exposure measurement factors to be considered for the physical environment health and disease causing such as noise intensity, light intensity, pressure/stress, vibration, temperature/coldness, or hotness and dust particles on 20 workplace sites. The selection method was using random sampling in line with purposive method. The measurement of the exposure factors was done in collaboration with Addis Ababa city Administration and Oromia Bureau of Labour and Social Affair (AACBOLSA). Some measuring instruments were obtained from the Addis Ababa city and Oromia

Data collection methods were focused on the followings basic techniques. These included secondary and primary data collections focusing on both qualitative and quantitative data as defined in the previous section. The data collection mechanisms

Primary data sources are qualitative and quantitative. The qualitative sources are field observation, interview, and informal discussions, while that of quantitative data sources are survey questionnaires and interview questions. The next sections

Observation is an important aspect of science. Observation is tightly connected to data collection, and there are different sources for this: documentation, archival records, interviews, direct observations, and participant observations. Observational research findings are considered strong in validity because the researcher is able to collect a depth of information about a particular behavior. In this dissertation, the researchers used observation method as one tool for collecting information and data before questionnaire design and after the start of research too. The researcher made more than 20 specific observations of manufacturing industries in the study areas. During the observations, it found a deeper understanding of the working environment and the different sections in the production system and OSH practices.

Interview is a loosely structured qualitative in-depth interview with people who are considered to be particularly knowledgeable about the topic of interest.

where *n* = sample size, *Z* = statistic for a level of confidence, *P* = expected prevalence or proportion (in proportion of one; if 50%, *P* = 0.5), and *d* = precision (in proportion of one; if 6%, *d* = 0.06). *Z* statistic (*Z*): for the level of confidence of 95%, which is conventional, *Z* value is 1.96. In this study, investigators present their results with 95% confidence intervals (CI).

#### *Research Design and Methodology DOI: http://dx.doi.org/10.5772/intechopen.85731*

*Cyberspace*

from the selected companies.

**4.1 Population**

industries responded.

of the criteria for the selections.

results with 95% confidence intervals (CI).

**4.2 Questionnaire sample size determination**

**4. Population and sample size**

on the manufacturing industrial sectors. The data also obtained from the existing working documents, manuals, procedures, reports, statistical data, policies, regula-

In general, for this research study, the desk review has been completed to this end, and it had been polished and modified upon manuals and documents obtained

The study population consisted of manufacturing industries' employees in Addis Ababa city and around as there are more representative manufacturing industrial clusters found. To select representative manufacturing industrial sector population, the types of the industries expected were more potential to accidents based on random and purposive sampling considered. The population of data was from textile, leather, metal, chemicals, and food manufacturing industries. A total of 189 sample sizes of industries responded to the questionnaire survey from the priority areas of the government. Random sample sizes and disproportionate methods were used, and 80 from wood, metal, and iron works; 30 from food, beverage, and tobacco products; 50 from leather, textile, and garments; 20 from chemical and chemical products; and 9 from other remaining 9 clusters of manufacturing

A simple random sampling and purposive sampling methods were used to select the representative manufacturing industries and respondents for the study. The simple random sampling ensures that each member of the population has an equal chance for the selection or the chance of getting a response which can be more than equal to the chance depending on the data analysis justification. Sample size determination procedure was used to get optimum and reasonable information. In this study, both probability (simple random sampling) and nonprobability (convenience, quota, purposive, and judgmental) sampling methods were used as the nature of the industries are varied. This is because of the characteristics of data sources which permitted the researchers to follow the multi-methods. This helps the analysis to triangulate the data obtained and increase the reliability of the research outcome and its decision. The companies' establishment time and its engagement in operation, the number of employees and the proportion it has, the owner types (government and private), type of manufacturing industry/production, types of resource used at work, and the location it is found in the city and around were some

The determination of the sample size was adopted from Daniel [5] and Cochran [6]

(1)

formula. The formula used was for unknown population size Eq. (1) and is given as

where *n* = sample size, *Z* = statistic for a level of confidence, *P* = expected prevalence or proportion (in proportion of one; if 50%, *P* = 0.5), and *d* = precision (in proportion of one; if 6%, *d* = 0.06). *Z* statistic (*Z*): for the level of confidence of 95%, which is conventional, *Z* value is 1.96. In this study, investigators present their

tions, and standards were taken into account for the review.

**30**

The expected sample number was 267 at the marginal error of 6% for 95% confidence interval of manufacturing industries. However, the collected data indicated that only 189 populations were used for the analysis after rejecting some data having more missing values in the responses from the industries. Hence, the actual data collection resulted in 71% response rate. The 267 population were assumed to be satisfactory and representative for the data analysis.
