**2.3 Study population and sample size**

Migrant labor workers who are engaged in precarious work activities in the poor urban setting of Agbogbloshie served as the population for the study. An enumeration undertaken in 2009 shows that the area has a total population of 79,684, with a population density of 2,424 persons per hectare [57]. For the quantitative, an initial screening of rural-urban migrant slum households' heads or their representatives was conducted. In total, 119 of the respondents were identified to represent the total population of the respondents. The small sample size that was identified during the screening was a result of the fact that data were collected during the upsurge of COVID-19; hence, most of the migrants had gone back to their places of origin in

order to avoid any threat posed by the virus. Yamane's formula was used to determine the sample size of 113. The formula is stated below:

$$\mathbf{n} = \left[ \mathbf{N} / \mathbf{1} \mathbf{+N} (\mathbf{e}) 2 \right] \mathbf{j}$$

Determination of sample size is based on the estimated population size (n = 119). N - The sample size N - The population size e - The desired level of precision or level of acceptable error = 0.05) Total sample size (n)= [119/(1+119(0.05)2]=[119/(1+119×0025) =[119/1+0.05] =[119/1.05] =113 For the qualitative aspect, an estimated twenty-one (21) respondents were selected

for the data collection. But due to saturation, a total of twelve (18) respondents were selected. Twelve (12) in-depth interviews (IDI) and eight (8) key informant interviews (KII) were held with the migrants. The use of the qualitative approach enabled us to select participants who were engaged in precarious work activities and were knowledgeable about the issue under investigation. It is also to allow the migrants to share their experiences about migration and precarious working conditions and how these have affected their health outcomes.

#### **2.4 Sampling frame, sampling technique**

In the absence of any existing reliable sampling frame, a household listing exercise was conducted in the migrant communities and a recording of the contacts of the household representatives (or heads) was made, to whom the questionnaires were administered. The simple random sampling technique was used to gather quantitative data. This was accomplished by labeling code numbers to the households, which were then placed in a box and shuffled. The labels were then picked randomly from the box to form the sample in a series of draws. The aim was to ensure that each household head had an equal chance of being selected [58].

For the qualitative study, the purposive sampling technique was used to select the respondents. Since the majority of the respondents did not understand English, the services of a translator were engaged to help with translation into the English language, and this served as the main mode of communication between the researcher and the migrants. To understand what the participants were saying, the translator used Dagbani and Mamprusi (i.e., their local dialects) to convey the information to the participants. Also, a key informant interview was conducted at the Ministry of Employment and Labor Relations, which is a government department. This was to enable us to understand the efforts being made by the government to address the precarious working conditions of migrant labor workers in Ghana.

#### **2.5 Ethical consideration**

We adhered to the national protocol on ethical issues in conducting research involving subjects. Permission was sought from the opinion leaders before entering the community by announcing the study. Also, written consent was obtained from the participants after carefully explaining the key issues to be considered before giving consent. Participants who did not want to participate or withdraw from the study could do so at any time.

While noting the caution in giving incentives to research participants, some incentives, such as soaps and the cost of transportation fare, were given at the end of every meeting.
