This chapter covers the methods that were used in conducting the research. The chapter begins by describing the conceptual framework for the study and a description of the research approach. The population for the survey is also identified as well as how the sample of participants was selected. The instrument for data collection is described, including what procedures were undertaken to ensure content validity and internal consistency reliability of the research. The chapter also describes how the data was analysed. Finally the research procedure is outlined to describe the flow of activities for the research.
Research Framework
Research framework outlined the collection of items to ask and things to observe in particular contexts, along with contextually appropriate techniques for doing so. Research framework also included processes for integrating research or data from other practice areas as well as specific methodologies for making meaning of the raw research (Fischler, 1981).
Another definition noted that the research framework would present a structure for supporting or enclosing the data collected especially since the research began as a skeletal support that would be used as the basis for the model that was being constructed. Therefore, research framework consisted of a set of assumptions, concepts, values, and practices that constituted a way of viewing reality (www.thefreedictionary.com).
The research was influenced by the work of Çalişkan (2010), in his research titled
“The impact of strategic human resource management on organizational performance.”Çalişkan (2010), stated that the relationship between HRM and firm performance has been a hotly debated topic over the last two decades and both organizations and academics are striving to prove that HRM has a positive impact on bottom line productivity. The research generally reports positive statistical relationships between the greater adoption of HR practices and business performance. Therefore, since the relationship between SHRM and SHRD as previously discussed, has already highlighted the fact that SHRM practices encompass the very function of SHRD. As a result, adopting SHRM practices that improve organizational performance will cascade down to inform the SHRD strategies that will in turn enhance individual performance.
30
STRATEGIC HRD
DEMOGRAPHICS
Figure 3.1 Conceptual Framework
Research Approach
The study took a quantitative approach. Quantitative research consists of studies in which the data concerned can be analysed in terms of numbers. The research was based more directly on original plans and the results were more readily analysed and interpreted.
Quantitative research is, as the term suggests, concerned with the collection and analysis of data in numeric form. Quantitative research tends to emphasize relatively large-scale and representative sets of data, and is often, falsely presented or perceived as being about the gathering of facts (Hughes, 2006).
The researcher used the quantitative approach to the study because the research was a comparative survey that sought to explain and predict the relationship between barriers to
Perceived Barriers to Enhanced Performance Age
Gender Educational Level
Marital Status Position Years of Experience
Communication Systematic Training
Accountability &
Ownership Culture Building
Quality Performance Management
31
enhanced employees as perceived by both the supervisors and employees. Moreover, the quantitative approach was the most appropriate for the study since the researcher gathered data from a large number of people.
Some of the strong characteristics of quantitative research are that the research can have precision through quantitative and reliable measurement. In addition, the researcher can exercise control through sampling and design (Hughes, 2006).
Population, Sampling and Subject Selection Population
Some authors define population as the total number of units from which data can be collected such as individuals and organisations. They further describe a population as all elements that meet the criteria for inclusion in a study (http://uir.unisa.ac.za).
The target group that made up the population for the study were public sector employees based in the Swazi government Ministries’ headquarters that are all located in Mbabane, Swaziland. For purposes of accessibility, time and financial constraints, including every public sector employee was not possible as some of the government offices are remotely located.
Sampling Procedure
Multi-stage cluster sampling was used to shortlist the government departments that are located in the preferred location, Mbabane. The criterion for selecting participants was based on the fact that they were available and easy to access. The sample represented the most convenient and available participants in addition to being easy and inexpensive to access.
Sample size
Smith (2013), advised that before calculating a sample size, researchers should consider a few things that need to be determined about the target population and the sample needed. Regarding the population size, the author advised that researchers should consider the number of people that fit the researcher’s target population.
32
In addition, good research should be able to state the desired margin of error, popularly known and referred to as confidence interval. However, researchers advise that no sample will be perfect, thus, the researcher should individually make the decision on how much error will be allowed in their study. The confidence interval determined how much higher or lower than the population mean the researcher was willing to let their sample mean fall. Furthermore, after determining the confidence interval, the researcher had to state the confidence level. The confidence level basically highlighted how confident the researcher intended to be that the actual mean fell within their confidence interval.
The Swazi government has over 35, 000 public sector employees in 100 Departments housed within 18 Ministries spread out in the four administrative regions of the country.
However, for purposes of accessibility, time and financial constraints, the research covered only those in Mbabane. The researchable population for this region was 1375 employees. In order to achieve a confidence level of 95% while allowing a margin of error of 5%, the size of the sample selected for the study was 300 participants. Since the sample constituted a subgroup of the target population that the researcher intended to study, the findings will then be generalized to the larger population. The table below shows the distribution of the Ministries and Departments from where the sample of 300 participants was selected.
Table 3.1
Randomly Selected Departments
Ministry Department
Tourism Tourism, Wildlife and Forestry.
Foreign Affairs Protocol.
Natural Resources and Energy Accounts, Energy, Property Valuation.
Agriculture Administration, Economic Analysis &
Planning, Data Investigation Unit.
Economic Planning and Development Aid Coordination, Central Planning, Macro Economic Unit.
Housing and Urban Development Urban Government, Housing Settlements.
Commerce and Industry Handicraft Promotion, Trade Development.
Education and Training Teaching Service Commission.
Finance Budget & Economic Affairs, Fiscal &
(continued)
33 Table 3.1 (continued)
Labour and Social Security
Monetary Affairs, Public Enterprise.
Labour, National Employment Services.
Public Service Management Services Division, HRD,
Administration, Civil Service Commission.
Communication, Information and Technology Computer Services.
Home Affairs Immigration
Instrument for Data Collection
The SHRD practices identified in the literature review formed the basis for the statements that were used to explore and describe the perceptions of the supervisors and employees of barriers to enhanced performance. A questionnaire with forty-five (45) closed-ended statements was used to collect information from the target population. The Likert scale was be used to rate the responses from6 indicating the highest level of agreement (strongly agree) to1, the lowest level indicating non-agreement (strongly disagree) for a series of statements. The symmetric scale helped to simplify and quantify the participants’ responses.
Paper-pencil questionnaires were hand delivered to the selected participants to ensure the control of non-response error. People usually find a sense of commitment through personal face to face interaction and feel somewhat obliged to respond since they already agreed to take part in the survey. Furthermore, people are generally more truthful when responding to anonymous questionnaires regarding controversial issues in particular due to the fact that their responses cannot be identified with them or traced back to them. However, drawbacks like a majority of the people who receive the questionnaire not returning or responding are possible (Leedy & Ormrod, 2001).
Validity and Reliability Analysis
Reliability according to Phelan & Wren (2006), refers to the degree to which an instrument produces stable and consistent results. Each time the instrument is used for data collection it should measure consistently whatever it sets out to measure. On the other hand, validity refers to whether an instrument measures what it purports to measure. A valid instrument will truly measure that which it was intended to measure. Research pointed out
34
that reliability alone was not sufficient; therefore, for an instrument to be reliable it must also be valid (Golafshani, 2003).
For purposes of controlling the measurement error in a survey, a professional in the human resource area of the Swaziland government reviewed the instrument. The researcher’s thesis advisor including other academic experts at the National Taiwan Normal University also reviewed the instrument. Peer review by academic colleagues at the National Taiwan Normal University and National Chengchi University as well as professional colleagues at the Ministry of Public Service in Swaziland, was also conducted to ensure content validity.
In addition, Cronbach’s alpha was calculated for the different SHRD domains to estimate the internal consistency reliability for the specific items in the questionnaire. A general overview of the results indicate that the different variables, estimated separately achieved a below acceptable level of Cronbach’s alpha (see Table 3.2) below. Furthermore, the means of the variables were used to estimate Cronbach’s alpha to ascertain if there will be a difference. Indeed, the variables combined yielded an acceptable Cronbach’s alpha as shown in Table 3.3 below.
Table 3.2
Reliability Statistics for Domains
Variables Cronbach’s Alpha N of Items
MCOM .638 7
MST .653 6
MAO .525 6
MCB .497 6
MQS .303 6
MPM .496 6
MPB .301 8
Data Collection and Analysis
Data Collection
As previously mentioned, data were collected from three hundred 300 public sector employees in the government of Swaziland. A Likert scale of six levels was used for forty-five (45) statements. For demographic data, only seven items were required from the
35
respondents. The data collection period ran from January 20 to February 21, 2014. The researcher personally delivered the questionnaires to all the randomly selected departments and respondents, with a two weeks deadline to complete the form.
Data Analysis
Quantitative analysis of the data was used because the study took a quantitative approach. According to Babbie (2008), quantitative analysis is the numerical representation and manipulation of observations for the purpose of describing and explaining the phenomena that those observations reflect. The data were analysed using Statistical Programme for Social Sciences (SPSS) to answer the research questions and test the hypotheses. Statistics employed are descriptive analysis, t-test, ANOVA, Pearson correlation and regression.
The statistical analysis helped to describe the relationships among the variables identified in the study. Descriptive statistics were used to describe the demographic variables of the population being studied. Since the target population for the study is large, descriptive statistics will provide a powerful summary that will enable comparisons across people or other units, thus providing quantitative descriptions in a manageable form (www.socialresearchmethods.net). The table below shows how the data were coded into different categories according to the different variables and for easy reference when conducting the analysis.
Table 3.3 Coding Guide
Variable/Domain Code Number Of Items
SHRD/Communication COM 7 (1-7)
SHRD/Systematic Training ST 6 (8-13)
SHRD/Accountability &
Ownership
AO 6 (14-19)
SHRD/Culture Building CB 6 (20-25)
SHRD/Quality Strategy QS 6 (26-31)
SHRD/Performance Management
PM 6 (32-37)
Perceived Barriers PB 8 (38-45)
Demographics DEM 7 (46-52)
36
37