Name
Professor
Subject
Date
5.1 SAMPLING PROCEDURES
A sample is a set of individuals or objects chosen from a statistical population through a well-defined procedure. A population-representative sample is unbiased or statistically significant if the parameter estimates of the sampling distribution have a mean equivalent to the population parameter underestimation (Wang et al., 2020). In this study, we will consider a sample obtained from the health workers in Monrovia, Liberia.
The concept of sampling is an essential tool in research because the main objective of sampling is forming a small group from a population that is similar to the larger population as possible. Sampling as a tool is essential to research in the following ways:
- Time-saving. By quickly taking a part of the population to represent the more significant population, a researcher can carry out the research process in a limited amount of time. In this study, sampling eased the summarization procedure (Pandey & Chawla, 2016).
- Reduced cost. Sampling reduces the cost of researching in that it is less pricey to get data from a portion of a population than from the entire population. In this research, the reduced cost enabled accurate data collection; hence a proper accurate representation was obtained (Fröhlich et al., 2018).
- Sampling ensures that data collection and analyses are done conveniently; hence accuracy.in this specific study, sampling brought ease by enabling easier summarization.
Target Population. A target population is the overall number of individuals or groups from which a sample is selected (Etikan & Bala, 2017). The target population in this research study were the workers in the Liberian Ministry of health. The researcher was interested in establishing how the leaders from the various department in the ministry responded to the outbreak emergencies. It is a case study of Ebola.
Determination of sample size. The determination of sample size is the process of selecting several replicates or observations in a sample. Before the determination of sample size, the following factors were put into consideration.
- Population size. This factor puts into consideration the number of individuals engaged in the research in total. The population was the entire workforce of the ministry of health I Liberia.
- Level of confidence. It is how confidently a researcher wants the average to be within the marginal error.
- Standard deviation. Standard deviation is the deviation of the response from one another and the average number.
- Degree of variability. It refers to varying levels of the attribute under study. The higher the degree of heterogeneity, the larger the sample size required to get the desired level of precision. And the more the homogeneity, the more a small sample size will be used.
This research reverted to the use of formulas approach as the strategy to determine the required sample size. In specific Taro, Yamane’s sample determination formula of 1967 was used in this process.
Yamane’s method of sample size calculation was invented by Taro Yamane to get a representative sample of a given population. The mathematical representation of the formulae;
n=N/(1+N(e)2)
denotations,
n – desired sample size
N – the target population.
E- accepted margin of error (in most cases, 0.05)
Liberia has a total number of five thousand (5000) health workers, with 3.8 million people. The study’s desired error margin is 0.05
Therefore, the estimated sample size is n=5000/ (1+5000(0.05)2)
n=5000/(13.5)
n=370
from the results above, it can be seen that the sample size is 370 people out of a population of 5000.5000 is the lower number of the respondent to maintain the 95% statistical level of the confidence interval.
Cluster sampling will be an ideal sampling technique for the research. Cluster sampling is defined as a probabilistic sampling method where the targeted group is subdivided into several subgroups called clusters (Makela & Gelman, 2018). A group is randomly selected using either a simple random sampling technique for the collection and analysis of data. In this sampling technique, a researcher selects a sample by analyzing the sample population with multiple sample parameters such as demographics, habits, or other human attributes that may be of interest. The multiple clusters of people from the population target indicate homogenous traits and pose equal chances of being included in the sample.
Cluster sampling is classified into single-stage cluster sampling and two-stage cluster sampling. Single-stage cluster sampling is based on the number of stages used to come up with the sample, while the two-stage cluster sampling is based on the representation of the group on the entire group (Sedransk et al., 2018).
- Single-stage cluster sampling. In this technique, sample clusters are randomly selected, and the data collected from every unit in the clusters. In Single-stage cluster sampling with equal-sized clusters, the total population contains N clusters, equally sized in terms of the number of units within the cluster (Wang, 2020). Simple random sampling is used to select n clusters, measurement of all population units within clusters is done. The unbiased estimator for the entire population mean is the average of cluster means in the selected sample. Analysis of variance can be used to show that clustering outshines simple random sampling whenever the mean square is for the clusters is less than the population variance.
- Two-stage cluster sampling. Rather than selecting all elements of a cluster, simple random sampling is used to select units from each group. A sample of clusters is randomly selected, and from the sampled cluster, a sample is selected randomly from the units. Two-stage cluster sampling is mostly used when the cluster size appears too large, making it challenging to observe all the units inside clusters. Two-stage cluster sampling can be extended to more stages (Eldardiry et al., 2020).
Steps in Cluster Sampling
- After deciding on the target population, the sample size is predetermined.
- The second step is Creating and evaluating sampling frames, done using either an already existing framework or by having a new framework for the target population.
- The third step is Determining the number of groups done by including similar average members in each cluster. The groups should be distinct from each other.
- The selection of clusters follows it by application of random sampling.
- The creation of sub-types depends on the size of the cluster.
Cluster sampling consumes less time as compared to other sampling techniques. Despite choosing large samples, researchers can conveniently access several clusters and therefore enhances accessibility. Data accuracy is enhanced since large samples in each cluster compensate for lost accuracy of information per unit. It is easier to implement cluster sampling than other sampling techniques (Loukas et al., 2020).
5.2 DATA COLLECTION PROCEDURES
Data collection is how data on variables of interest are gathered systematically to answer research-related questions, test the hypothesis, and make inferences based on the outcomes (Mirbabaie et al. ,2020). Data collection procedures describe all necessary steps that are used to carry out the effective methods of the study.
This study will be carried out in Monrovia (the capital city of Liberia) at the ministry of health. It will employ the use of focus groups and one on one interviews when collecting data. The mode of data collection will entirely depend on the respondent’s reaction. One on one interviews will be used in cases where the respondents do not feel comfortable to participate in focus groups due to privacy and confidentiality. Under the case of the focus groups, the respondents will voluntarily give their responses to the already established questionnaire after discussion. Each participant will have an equal chance to present their views regarding the questions in the questionnaire. Additionally, at one on one discussions, the respondents will have a chance to respond to a maximum of 10 questions covering the same topic as the focus group. The focus group interview will take a maximum of 90 minutes; the one-on-one meeting will last for more than 45 minutes.
The focus group questionnaire will contain eighteen inquiries. The majority will be open-ended, and all the questions will be formulated to capture our research question (How does the Ebola outbreak of 2014 influence Liberian Health leaders’). Furthermore, it is more important to put more weight on the participants signing the consent form, although participating in the interview will be optional. The primary purpose of signing the consent form is that to ensure the rights of the respondents are protected as well as to guarantee the researcher is safe from any legal imputation that may come out during the research.
However, under the current situation emanating from the outbreak of coronavirus, it is more difficult to conduct face-to-face interviews. Zoom technology will be the only platform available for the meeting. Therefore, all the interviews, both the focus group and the one-on-one meeting, will be conducted via zoom technology (Washington et al., 2020). Zoom program, which is a cloud-based video conferencing service used for virtual meetings, provides a chance of hosting virtual discussions where the respondent can interact with themselves and the researcher virtually, therefore, making the data collection process easy (Lawless et al., 2019). Zoom was selected because it provides privacy and security for the users, making it a more reliable platform for any virtual meeting. During the meeting, all interviews will be audio-recorded, and the information collected will be used to fill in the questionnaire questions. Since the questions are similar to the two groups, all the information recorded will be used during the analysis.
The data used in both the focus group and the one on one interview will both be collecting independently. Each of the two data collection procedures will have a different procedure. Interviews are considered to be the most convenient means to conduct qualitative data collection- The steps that will be involved in interviews are;
- Determination of the goal. The first step in an interview is to set the goal or the intended purpose of the study. In our case, we intend to collect data that will help in a qualitative analysis of the leadership in the health system. We will specifically study emergency preparation in the West Africa region.
- Target people to interview- Setting the set goals to guide the target population. A target population is a set of persons from whom data will be collected. In our case, the goal is to examine the leadership in the health sector. The best people to interview would be health ministry workers. The workers will provide the required information on the quality of leadership and how preparations were done towards emergency response in western Africa.
- The set objectives guide designing of the interviewing questions. The questions set are clear, unambiguous, and are easy to understand. Open-ended questions are best for interviews to give the respondents a room to explain their views. The questions must align with data collection goals and the intended method of analysis.
- Preparation for the interview. Preparation ensures that interview is effective. A guide helps the interviewer to stay on track and collect only the required data. The interviewer should rehearse the interview question and draft the answer path questions. Before commencing the study, the researcher is encouraged to create a rapport with respondents to make them feel comfortable while answering the questions.
- Conduct the interview, which is the data collection point. Guided by the questions and the interview guide, the researcher, obtains the required data recording it in a reliable form. The interviews should not take too long to avoid boring respondents.
- Transcription and analysis of the interview; It is the final step, and it involves scrutinizing the data to ensure that the information obtained is reliable. You should check to ensure that the answered questions met the goal of the interview.
The steps involved in a focus group interviewing are less similar to those taken in interviews.
- The first step is writing down the research goals. Before you commence on the data collection, you need to set the goal, which is the intended purpose of organizing the focus group.
- Defining your target audience- Our target audience here is the health sector workers. We, therefore, will include them in the focus group. Health sector leaders also need to be considered for the focus group participation.
- Venue selection- The venue selected should not be a fancy one. It should be a place easy to find and access and where communication will go on undisrupted. An example of a suitable venue is a church hall.
- Participants recruitment. The focus group is set with ideal size. The recommended size of a focus group is usually 10-15 people. It is also essential to offer gifts like a voucher to participants to lure them into accepting the deal.
- Questionnaire designing. The main aim of using a focus group is to stimulate a discussion. It is, therefore, essential to choose open-ended questions that are debatable. Probing questions should be used, or a question should be followed by a probe to ensure a discussion arises.
- Group moderation. The focus group consists of a moderator and an assistant moderator. The two take a video of the discussion and take notes.
- Analysis of the response. It is the last step and is to check the data collected and countercheck it with the research goal.
5.4 DATA ANALYSIS
Data analysis is the process by which inspection, cleansing, and transformation and modelling of the data are conducted to discover important information that can be used in making conclusions and support decision making. The data analysis method used depends on the type of data collection procedures used. For the case of this study, a primary data collection procedure will be used. To be precise, questionnaires in interviews and focus group discussions will be used.
The data analysis method that will be applied will include but is not limited to empirical data analysis. Empirical data analysis is a study based on evidence to study and make inferences of information. It relies on real data (Locke et al., 2020). The method allows iterations of examination data, hand-coding, and writing memos to aid in error detection. It identifies errors that may arise in the analysis process. The first step in the data analysis procedure is data preparation. The empirical method involves the reading and coding of information contained in the interview transcripts as well as combining them with other data sources. It will help detect errors and is also intended to help in the analysis of data.
Yin Research recognizes this exercise. They were especially emphasizing on the matching trends and the strategies relating to coming up with the explanation, which involves the identification of several themes and the trends that occur between cases and also being able to interpret the data so that to be able to illuminate how specific facts occurred (chai et al., 2016). It is essential to keep in mind that because the case study is a non-probabilistic process, the data collected from various sources will be analyzed collectively but within the parameter of study. All the interview transcripts will be scrutinized and also the written word to find out the similarity in the themes. After identifying the common themes, they will be restricted to sub-themes. It will be used to identify the respondent’s competence about Liberia’s Health system, leadership, and emergency preparation intervention, which in turn will help in coming up with examples and reliable information on interpretation of the data collected.
Additionally, the data analysis procedure will also help in the combination of the data that are in similar groups. Also, it will help with grouping broader themes into narrow ones to enhance the overall analysis of the study. While using memos and hand-coding will help with the creation of themes and streamlining of data significantly in the way of improving the entire success of the study, therefore, it is essential to recognize that hand-coding and writing of memos have several limitations. It has been shown that unless the researcher creates a process by which hand-coding and writing of memos may be counterchecked, and the number of errors may be high. According to Chandra & Shang (2019), the data analysis will include a process by which hand-coding and writing of memos will be counter checked. However, as themes are channelled from broader themes, the chances of errors in the data collection process could be minimal. After the hand-coding and writing of themes, then the data collected will be entered into excel for correct presentation and correction of errors that arose during collection. Since the data will be primary, then a sample will be used to analyze and conclude the entire study population after the analysis.
Afzalan, M., Jazizadeh, F., & Eldardiry, H. (2020). Two-Stage Clustering of Household Electricity Load Shapes based on Temporal Pattern & Peak Demand. arXiv preprint arXiv:2008.04293.
Archibald, M. M., Ambagtsheer, R. C., Casey, M. G., & Lawless, M. (2019). Using Zoom videoconferencing for qualitative data collection: perceptions and experiences of researchers and participants. International Journal of Qualitative Methods, 18, 1609406919874596.
Bhatta, D., Nandram, B., & Sedransk, J. (2018). Bayesian testing for the independence of two categorical variables under two-stage cluster sampling with covariates. Journal of Applied Statistics, 45(13), 2365-2393.
Chai, D. S., Jeong, S., Kim, J., Kim, S., & Hamlin, R. G. (2016). Perceived managerial and leadership effectiveness in a Korean context: An indigenous qualitative study. Asia Pacific Journal of Management, 33(3), 789-820.
Chandra, Y., & Shang, L. (2019). Qualitative research using R: A systematic approach. Springer Singapore.
Etikan, I., & Bala, K. (2017). Sampling and sampling methods. Biometrics & Biostatistics International Journal, 5(6), 00149.
Fröhlich, H., Balling, R., Beerenwinkel, N., Kohlbacher, O., Kumar, S., Lengauer, T., … & Rebhan, M. (2018). From hype to reality: data science enabling personalized medicine. BMC medicine, 16(1), 150.
Locke, K., Feldman, M., & Golden-Biddle, K. (2020). Coding Practices and Iterativity: Beyond Templates for Analyzing Qualitative Data. Organizational Research Methods, 1094428120948600.
Makela, S., Si, Y., & Gelman, A. (2018). Bayesian inference under cluster sampling with probability proportional to size. Statistics in Medicine, 37(26), 3849-3868.
Pandey, S., & Chawla, D. (2016). Using qualitative research for establishing content validity of e-lifestyle and website quality constructs. Qualitative Market Research: An International Journal.
References
Stieglitz, S., Meske, C., Ross, B., & Mirbabaie, M. (2020). Going back in time to predict the future-the complex role of the data collection period in social media analytics. Information Systems Frontiers, 22(2), 395-409.
Tremblay, N., & Loukas, A. (2020). Approximating spectral clustering via sampling: a review. In Sampling Techniques for Supervised or Unsupervised Tasks (pp. 129-183). Springer, Cham.
Wang, L. (2020). Improving the external validity of epidemiologic analyses by incorporating data from population-based surveys (Doctoral dissertation).
Wang, L., Graubard, B. I., Katki, H. A., & Li, Y. (2020). Improving the external validity of epidemiologic cohort analyses: a kernel weighting approach. Journal of the Royal Statistical Society Series A, 183(3), 1293-1311.
Washington, C. D. (2020). Increased Access to Information Communication Technologies and Work-Life Balance with K-12 Leaders.