Characteristic Positivist View Interpretive View Nursing Essay

Introduction

This chapter describes the research design and methodology that was used in this research, including sampling, population, establishing rigour during and after data collection and how data was derived from primary and secondary sources, validity and reliability and ethical considerations. The fundamental aim of this chapter was to determine what the key factors are that affect project quality at Koeberg. The ultimate objective was to find an answer to the research question, as defined in Chapter 1, Paragraph 1.2, which reads "the erosion of project quality and the inconsistencies in delivering quality projects have an adverse effect on modifications and projects implemented on the nuclear power plant in South Africa

Information to determine factors that influence project quality and what in the project environment allows these to persist, was collected from the Nuclear Project Management Department (NPM), departments that have a direct influence on NPM processes as well as contractors. While project quality information was obtained via a survey (primary source), historical data related to project quality and its influence on the project lifecycle at Koeberg was gleaned from databases and archives (secondary resources).

The survey environment

NPM has to perform various duties in the project management of modifications to the plant. These duties include the execution of projects to the plant asset creation process, plant related modifications motivated predominantly for operational needs and in accordance with requirements set by international standards for operating plants and the National Nuclear Regulator (NNR). These modifications have to be implemented in accordance with the Project Lifecycle Model (PLCM). The functional areas within NPM, which served as the research survey environment included:

Project Development

Project Execution (Operational)

Construction Management

Programmes Management

Monitoring and Support

Contracts Management

Strategic Projects

External departments having a direct impact on NPM include:

Engineering (Plant Engineering and Nuclear Engineering) Page 1 Done

Finance

Procurement (Project Sourcing and Project Quality Engineering)

Quality Assurance / Quality Control

Outage Management

Maintenance

Suppliers/Contractors/Vendors

All researchers have different beliefs and ways of viewing and interacting within

their surroundings. As a result, the way in which research studies are conducted vary.

However, there are certain standards and rules that guide a researcher’s actions and

beliefs. Such standards or principles can be referred to as a paradigm. To gain a

better understanding of why and how the researcher chose the methodological

approach in this study, an initial discussion will be completed about the paradigm

that best fits the focus of this study.

Following a discussion about the research paradigm, the aim of this chapter is to

discuss the research design and methodology utilised in this study. In order to

describe the variety of research activities undertaken during this study, the data

collection activities and associated analysis methods will be systematically discussed

under four phases. For ease of discussion, the study activities will be described in the

order in which the researcher completed them. The order of the study activities have

been outlined in Figure 3.

Research Paradigm

According to Taylor, Kermode, and Roberts (2007, p. 5), a paradigm is "a broad

view or perspective of something". Additionally, Weaver and Olson’s (2006, p. 460)

definition of paradigm reveals how research could be affected and guided by a

certain paradigm by stating, "paradigms are patterns of beliefs and practices that

regulate inquiry within a discipline by providing lenses, frames and processes

through which investigation is accomplished". Therefore, to clarify the researcher’s

structure of inquiry and methodological choices, an exploration of the paradigm

adopted for this study will be discussed prior to any discussion about the specific

methodologies utilized in this study.

This study utilised a triangulation approach to explore and guide the development

and evaluation of a clinical forensic nursing educational package. The use of both the

qualitative and quantitative methodologies was necessary to encompass the different

aspects of forensic science and nursing’s holistic approach to patient care. According

to Lynch (2006), providing forensic patient care requires objectivity and neutrality

while attending to the various human dimensions of health and well-being. To

address the diversity and complexity of such nursing and forensic issues, a mixed

methodology was necessary.

According to Weaver and Olson (2006), the paradigms most commonly utilised in

nursing research are positivist, postpositivist, interpretive, and critical social theory.

The quantitative methodology shares its philosophical foundation with the positivist

paradigm (Weaver and Olson). The positivist paradigm arose from the philosophy

identified as logical positivism and is based on rigid rules of logic and measurement,

truth, absolute principles and prediction (Halcomb and Andrew, 2005; Cole, 2006;

Weaver and Olson). The positivist philosophy argues that there is one objective

reality. Therefore, as a consequence, valid research is demonstrated only by the

degree of proof that can be corresponded to the phenomena that study results stand

for (Hope and Waterman, 2003).

In this study, such rigid principles lend themselves more to the scientific forensic

aspects such as scientific knowledge, logic and measurement incorporated into this

study (Weaver and Olson, 2006; Lynch, 2006). However, such inflexible beliefs did

not have the capacity to accommodate the investigatory aspects of this study that

dealt with the social and human experiences. As a result, qualitative methodologies

were also incorporated into the research design (see Table 3.1).

The qualitative methodology shares its philosophical foundation with the interpretive

paradigm which supports the view that there are many truths and multiple realities.

This type of paradigm focuses the holistic perspective of the person and environment

which is more congruent with the nursing discipline (Weaver and Olson, 2006).

Additionally, the interpretive paradigm is associated more with methodological

approaches that provide an opportunity for the voice, concerns and practices of

research participants to be heard (Cole, 2006; Weaver and Olson). Cole further

argues that qualitative researchers are "more concerned about uncovering knowledge

about how people feel and think in the circumstances in which they find themselves,

than making judgements about whether those thoughts and feelings are valid" (p.

26).

Table 3.1: Summary of the Research Paradigms

Characteristic Positivist View Interpretive View

Purpose The researcher will predict and

explain changes in forensic

knowledge of HospC

participants

The researcher will interview the

stakeholders and recognise the

value and depth of the individual

content

Beliefs • One truth exists

• Must be objective

• Many truths and realities

• Different people have different

perceptions, needs and

experiences

Research Methods Quantitative Qualitative

What Study Data

is Based Upon

Measurable outcomes from

questionnaire data

Descriptive, explanatory and

contextual words of interview data

Study Sample Clear and precise inclusion and

exclusion data

Representatives who are able to

provide expertise from different

points of view.

Due to the complex nature of the research study, there was no single paradigm that

could satisfactorily deal with all of the required methodological aspects. Therefore,

the researcher found it necessary to combine the quantitative/positivist paradigm

with the qualitative/interpretive paradigm. The blending of both paradigms provided

the researcher with the ability to statistically analyse the scientific data whilst also

recognizing the complex psychosocial and emotional factors that influence patient

care issues. The discussion that follows will further elaborate and describe in detail

how each paradigm and methodological approach was implemented in this study.

Research Design

In this descriptive study, qualitative and quantitative data collection techniques were

used including; semi-structured interviews, chart audits, pre and post-test

questionnaires, focus group interviews, and the researcher’s field notes of personal

observations and conversations. Additionally, to provide a more complete and multidimensional understanding of the issues, a triangulation methodology design

was employed (Taylor, Kermode, and Roberts, 2007). In the section below, the

discussion will be divided into two main headings; that of descriptive research and

triangulation.

Descriptive Research

In order for the researcher to gain different perspectives and draw attention to

different factors that affect forensic practice in Western Australia, descriptive

research methods were employed in this study. According to Polit, Beck, and

Hungler (2001, p. 180), descriptive methods are used when the researcher seeks to

"describe, observe, and document a naturally occurring phenomenon which cannot

readily be ascribed an objective value". In other words, descriptive research deals

with questions that look to explain what things are like and describe relationships but

do not predict relationships between variables or the direction of the relationship.

Depending on what is to be described, descriptive research can be very concrete or

more abstract (DeVaus, 2002). At a concrete level, data collected is often strongly

quantitative in nature (Polit, Beck, and Hungler, 2001). In this study, data will be

collected in the form of participant demographics, chart audit data, monitoring of

implementation tools, and data collected from the pre and post-test questionnaires. In

addition, more abstract descriptive research, in the form of stakeholder interviews,

was also included. According to Morse and Richards (2002), qualitative descriptive

approaches are extremely helpful because evidence of experience and knowledge can

be easily missed when quantitative methods are used.

In this study, semi-structured interviews were incorporated into the study design

because the researcher believed that open ended questions would be the most

efficient way to collect data from stakeholder participants. Open ended questions are

thought to allow an individual time and scope to discuss their perception and

knowledge (Morse and Richards, 2002). DeVaus (2002) believes that descriptive

research can play a key role in highlighting the existence and extent of problems

which can stimulate interventions and actions that lead to policy change.

The intent of stakeholder interviews was to investigate and describe current forensic

practices occurring in Western Australia. According to Taylor, Kerrmode and

45

Roberts (2007), qualitative interviews attempt to "make meanings" from individual

accounts and experiences. Forensic patients are usually treated in partnership by

medical and legal professionals (Lynch, 2006). Therefore forensic and healthcare

stakeholders who work in the field have the best ability to contribute, enhance links

and increase the successful integration of services (Haddow, O’Donnell, and Heaney,

2007). The incorporation of stakeholders in this study was to enhance the

understanding of the current issues and experiences confronting forensic and

healthcare professionals working with and providing care to forensic patients.

Triangulation

Multiple triangulation methods were utilised in this descriptive study. Triangulation

involves the application and combination of several research methodologies in one

study (Schneider, Elliott, Lo-Biondo-Wood, and Haber, 2003; Taylor, Kermode, and

Roberts, 2007). There are four common types of triangulation discussed within the

literature including: data triangulation that involves time, space, and persons;

investigator triangulation which uses multiple observers; theory triangulation that

uses more than one theoretical perspective to interpret the study phenomenon; and

methodological triangulation that involves using more than one methodological

strategy during data collection. According to Halcomb and Andrew (2005), the use

of multiple data sources and methods to cross-check and validate findings increase

the depth and quality of the results and also provides valuable guidance to nursing

practice.

Triangulation provides in-depth data, increases the confidence in the research results

as well as enables different dimensions of the problem to be considered (Barbour,

2001; Jones and Bugge, 2006). A combination of methods is thought by some to

improve the consistency and accuracy of data by providing a more complete picture

of the phenomenon (Roberts and Taylor, 2002; Halcomb and Andrew, 2005;

Williams, Rittman, Boylstein, Faircloth, and Haijing, 2005; Jones and Bugge, 2006).

Morse (1991) cited in Minichiello, Sullivan, Greenwood, and Axford, (1999, p. 258)

believes that triangulation is a means by which the researcher is able to "capture a

more complete and holistic portrait of the phenomena under study".

46

In this study, the researcher employed methodological, data, and unit of analysis

triangulation. Each of these aspects of triangulation will be discussed individually

below and study examples provided to help illustrate the concepts. Firstly,

methodological triangulation will be explored which can be sub-divided into within

and across-method triangulation (Schneider, et al., 2003; Halcomb and Andrew,

2005).

Methodological triangulation

Methodological triangulation, according to Taylor, Kermode, and Roberts (2007),

involves using two or more research methods in one study at the level of data

collection or design. Across-method triangulation involves combining research

strategies usually qualitative and quantitative methods. Such an approach is common

in nursing studies (Jones and Bugge, 2006; Halcomb and Andrew, 2005). In this

study, for example, data from stakeholders interviews were utilised to reinforce and

complement the data from quantitative chart audits because concepts mentioned by

the stakeholders were checked during the chart audits. Complementary findings in a

study make a more valid contribution to theory and knowledge development,

enhance diversity, and enrich the understanding surrounding the study’s objectives

and goals (Schneider, et al., 2003; Macnee and McCabe, 2008).

Data triangulation

Data triangulation can be described as the use of multiple sources of data to obtain

differing views about a situation in a single study (Roberts and Taylor, 2002). For

example, in this study, data was collected from various interviews, pre and post-test

questionnaires and by reviewing nurse participant’s documentation within patient

medical records. Multiple data sources help validate the findings by exploring

different views of the situation under investigation (Taylor, Kermode, and Roberts,

2007). Data triangulation can be divided into categories of time, space, and person

(Roberts and Taylor).

Time triangulation involves researchers collecting data at different points in time

such as time of day; at different days of the week, or at different months of the year

(Rinaldi, Carpenter, and Speziale, 2006). In this study, however, the goal was not to

compare participant knowledge between shifts or from one month to the next.

47

Instead, the researcher was interested in evaluating an educational intervention over

time. Therefore, for this study, only two types of data triangulation were utilised:

space and person.

Space triangulation involves the collection of data from multiple sites (Roberts and

Taylor, 2002). In this study, for example, data was collected from two hospitals

emergency departments. Analysis from both sites helped evaluate the effectiveness

of Phase III activities of this research and also increased the validity and

strengthened the study (Begley, 1996; Halcomb and Andrew, 2005).

Person triangulation implies that data was collected from more than one category of

person (Roberts and Taylor, 2002; Taylor, Kermode, and Roberts, 2007). For

example, in this study, participants included ED nurses as well as key forensic and

healthcare stakeholders. The use of various legal and healthcare professionals

provided greater insight into a variety of issues including: hospital administration,

staffing, costing concerns; medical practices; Western Australian legal requirements

and governmental policies; current evidentiary processes; as well as existing

investigatory practices. Such data was utilised to support, supplement, and validate

the information gained from published forensic material as well as the research data.

Unit of analysis triangulation

The unit of analysis triangulation as described in Begley (1996) is the use of two or

more analysis approaches to validate the same set of data. In other words, the use of

differing qualitative techniques or different families of statistical tests helps verify

results. The researcher rarely found this type of triangulation discussed in current

literature; however, there was some dated literature that described this topic (Kimchi,

Polivka, and Stevenson, 1991; Begley, 1996; Bergen and While, 2000). In this study,

to evaluate the effectiveness of the forensic education package, several levels of

analyses were conducted. For example, by comparing pre and post questionnaire

responses and then interviewing and analysing the interviews the effectiveness of the

educational package was assessed at a participant level. In addition, data from the

chart audits and focus group interviews also provided qualitative and quantitative

data which assisted towards the analysis and evaluation of the package effectiveness.

Methodology

Due to the complexity of this research project, a true experimental design was not

able to be conducted. However, a quasi-experimental design is similar to that of a

true experimental design except that the participants are not randomly assigned to the

control and treatment groups (Schneider, et al., 2003; Taylor, Kermode, and Roberts,

2007). It was therefore decided to employ a descriptive a pre-test, post-test type of

design. Details of how this was utilised in this study is explained below.

This descriptive research study employed a multiple triangulation methodology

design in order to develop and evaluate the effectiveness of a forensic educational

package (see Figure 4). Theoretical guidance was sought from Bandura’s (1977)

Social Cognitive Theory, Malcolm Knowles (1980) Adult Learning Principles and

Lynch’s (1990) forensic nursing integrated practice model. Participants included 49

treatment and control group nursing participants from two metropolitan West

Australian hospitals. In addition, 22 forensic and hospital stakeholders from 10

forensic specialty areas were also involved. Qualitative and quantitative data was

collected across four phases from semi-structured interviews, policy manual reviews,

audits of nursing documentation, pre and post-test questionnaires, focus group

interviews, and the researcher’s observations. The following sections will describe

the research sites, the sampling and the data collection tools.

Questionnaire development

No previously tested questionnaire was available for this research study; therefore,

the researcher was required to develop and validate the pre and post-test

questionnaires before their use. The questionnaires were to be used to evaluate the

effectiveness of the implementation activities (workshops A, B, and C). The research

instruments were constructed after a thorough review of the available published

literature, consultation with local and international forensic professionals and

reflection upon the researcher’s knowledge and professional experience.

The researcher was confronted with two major issues when developing the pre and

post-test questionnaires. Firstly, the researcher needed to develop a tool that would

accurately assess whether there was any difference in forensic knowledge amongst

the treatment group participants after attending the three intervention workshops.

Secondly, the researcher needed the tool to be consistent when used on multiple

occasions with different groups of participants. These two important and

fundamental characteristics of a measurement tool (validity and reliability) need to

be proven before its use (DeVaus, 2002; Schneider, et al., 2003). How the researcher

addressed the issues of validity and reliability during the questionnaire development

will be explored below.

Validity

Validity is the most fundamental consideration in instrument development and refers

to the degree that the instrument measures what it claims to measure (DeVaus, 2002).

There are three basic ways in which to assess the validity of an instrument; criterion,

content and construct validity. The criterion validity approach compares the new tool

to an existing well-accepted instrument that measures the same concept (DeVaus;

Schneider, et al., 2003). Since no other instrument could be found in the published

literature, this approach could not be used to test the rigor of this instrument.

Therefore, the following discussion will focus on the issue of content and construct

validity.

Content validity refers to the ability of the instrument’s items to represent the content

of the given construct (DeVaus, 2002; Schneider, et al., 2003). When the researcher

was developing the instrument, the concern was whether the measurement tool and

the items it contained were representative of general forensic knowledge which was

what the researcher intended to measure. To tackle the issues of content validity, the

researcher approached forensic and forensic nursing experts to examine the

questionnaire’s content. The researcher wanted to ensure that the tool focused on

fundamental and essential forensic nursing concepts (DeVaus; Schneider, et al.).

Forensic nurse specialists living in the US and all of the legal and forensic

stakeholders interviewed for this study were asked to review and examine the pre and

post-implementation questionnaires for accuracy and content. The experts in the US

were all e-mailed copies of the questionnaires and an information sheet explaining

the purpose of the study. All of the Western Australia legal and forensic stakeholders

were provided a research information sheet and the questionnaires in person. The

information sheet discussed the objectives of the study and of the questionnaire.

Comments on items and their relevance were clarified and modified according to the

comments from the reviewers. Minor modifications to the layout and wording were

made prior to its use in the study.

For example, a clinical forensic nurse specialist working in the US made suggestions

about the wording of some of the questions. The US expert commented about

sentence structure and the presence of language inconsistencies. For example, a

suggestion to include the statement "Please tick all that apply" at the end of some

questions and to have other sentences begin with "How many of the choices listed

below". Therefore, the language and sentence structure of every question was

examined carefully so that the participants would not be confused by the content of

the questions.

In total, two West Australian lawyers, five US forensic nurse specialists, and 10

forensic specialists agreed to review the questionnaire for content and to provide

answers to questions specific to their forensic specialty. For example, the Forensic

Biologist, Forensic Scientist and Forensic Pathologist reviewed and answered

questions 10 and 12 which dealt with specimen collection. Such professional

feedback allowed the researcher to develop a more accurate marking key and

subjected the questionnaire to further scrutiny prior to its distribution and use in the

research study.

The last type of validity that required discussions is construct validity. Construct

validity refers to the extent in which the instrument measures a theoretical trait

(DeVaus, 2002; Schneider, et al., 2003). This type of validity is difficult to achieve

and was not used in this study as there was no single, well established theory

associated with forensic nursing suitable for this study. Therefore, the researcher

utilised three different theoretical models to deal with the complexity of the study.

The establishment of construct validity can be a complex process that often involves

many studies and several different approaches (DeVaus). Furthermore, DeVaus

believes that there is no ideal way of determining validity and that the researcher

must choose the method best suited for the situation.

Reliability

As well as the issue of validity, it was essential to consider the reliability of the pre

and post-test questionnaires. Reliability addresses the ability of a measuring tool to

provide the same result on repeated occasions (DeVaus, 2002; Schneider, et al.,

2003). The method of test-retest reliability addresses the question of consistent

answers from multiple occasions of use. Depending on the text, the suggested

interval at which the retest should be administered varied from two to six weeks

(DeVaus; Golan, and Weizman, 1998; Zwart, Frings-Dresen, and vanDuivenbooden,

2002; Taylor, et al., 2001). DeVaus suggested that a trial of the instrument be

undertaken on a smaller but similar practice sample to that being used in the study.

To address the issue of questionnaire reliability in this study, the test re-test method

of reliability testing was used. Twelve experienced Clinical Nurses were asked to

complete the questionnaire twice. None of these nurses came from HospB or HospC.

Eight weeks after completing the questionnaire, the 12 Clinical Nurses were asked to

complete the same questionnaire again. During the eight-week time period, the

nurses were asked not to research information about any of the questions or talk to

anyone about the contents. After the eight week time period, 10 of the 12 nurses

completed and submitted the second questionnaire. There were two nurses who

declined further involvement in the reliability testing.

The scores from both questionnaires were evaluated and the tool assessed for

consistency and reliability of answers. A comparison of test scores was expressed by

a Pearson correlation coefficient, r. The magnitude of the coefficient (r = 0.85)

provided support regarding the tool’s stability. An r equal or greater than 0.7 is

considered an acceptable value for a tool to be viewed as reliable (Burns and Grove,

2007). Therefore, this result indicated that the questionnaire was a reliable tool.

The final contents of both the pre and pot-test questionnaires included short answer

and tick box response questions as well as a demographic cover sheet on the pre-test

questionnaire (see Appendix 5 and 6). The researcher developed the questionnaire by

modifying and selecting information that was considered fundamental forensic

knowledge. All of the information was derived from published literature (Wick,

2000; Meserve, 1992; Easter and Muro, 1995; Pavlik, 2004), Western Australia legal

codes (Western Australia Coroner’s Act; Privacy Act of 1988; Western Australia

Criminal Code), and real life patient scenarios experienced by the researcher and

other forensic experts. The questions were designed to target and explore forensic

issues that related to patient assessment, law and ethics, and nursing practice.

In total, the pre-test questionnaire contained 25 questions including five demographic

questions that were not repeated on the second questionnaire. To minimise any

confusion between the two instruments during analysis, the first pre-test

questionnaire was printed on white paper and the second post-test questionnaire on

orange paper. Participants were given the pre-test questionnaire only after the

researcher received a signed consent form.

66

Participants were identified only by a code number located in the top right hand

corner of the questionnaire. Each code began with a capital "H" for hospital, then a

capital "B" or "C" to identify which hospital data was collected from, and finally a

number (1-27 for HospB participants and 1-22 for HospC participants). The number

at the end of each participant’s code was assigned according to when the participant

agreed to participate. Participants could only be associated with their code through

their consent form which had their personal code, their name and signature on the

single document. Only the researcher had access to such information.

In addition to the questionnaire data, the researcher reviewed all of the policy and

procedure manuals located at Hospital B, and C. Such information was used to assist

the researcher during the scoring of the questionnaire. Details of the methods used

during this study activity will be outlined below.

Policy and procedure manual review

There were two types of policy and procedure manuals that were reviewed at HospB

and HospC; the main hospital policy and procedure manual and the ED nursing

policy and procedure manual. It was necessary for the researcher to review both sets

of manuals because nurses must follow, and are accountable for, practicing under all

items discussed within individual area policy and procedure manuals as well as the

more general hospital wide policies.

The focus of each review was to evaluate the policy and procedure manuals for any

forensic related issues. Each forensic related policy was examined for the inclusion

of treatment guidelines, clinical pathways, legal implications of guidelines, and

suggested referral agencies. During each hospital’s policy and procedure manual

review the researcher noted if the above items were included in the policies (see

Appendix 7). Any conflict between research protocols and hospital policy and

procedure information needed to be considered before finalising workshop content,

marking participant questionnaires, and analysing final data.

According to De Vaus (2001:9-16) the function of a research design is to ensure that the evidence obtained enables the answering of the initial question as unambiguously as possible. The author further writes that research design is the structure of an enquiry; a logical rather than a logistical matter where the type of evidence answers the research question convincingly so it is not only consistent with a particular theory but must be found to have the potential to disprove preferred explanations.

Saunders et.al (2007:147-152) believe that research design has to do with the credibility of research findings with Watkins (2008:42) defining it as "…the logical sequence that connects the empirical data to study’s initial research question and ultimately, to its conclusion". Research design is therefore an action plan to get from here to there where ‘here’ may be defined as the initial set of questions to be answered, and ‘there’ is some set of conclusions (answers) to these questions.

Crotty (2003:3) is of the opinion that research methodology is the strategy, plan of action, process or design lying behind the choice and use of particular methods and linking them to the desired outcomes while Collis and Hussey (2003:55) state that it refers to the overall approach to the research process, from the theoretical underpinning to the collection and the analysis of data.

The thinking of the researcher in this study was that all components (plant, people, processes and procedures, the project lifecycle, project quality, survey environment and respondents) should fit together in such a way that meaningful information and answers would be obtained. To achieve this goal, the researcher drew up a design strategy (as discussed from paragraph 3.4 onwards) that assisted in obtaining answers to the research questions that were raised in Chapter 1 paragraph 1.4.1. The adopted strategy shaped the choice and use of particular methods and linked them to the desired outcomes. This approach provided insight into factors that affect project quality at Koeberg.

Page Done

Positivism/Post-Positivism

Constructivism

Emancipatory

Ontology (Nature of reality)

One reality

Reality knowable within probability

Multiple constructed realities

Multiple realities which include the social, political, cultural, class, economical, gender, etc.

Epistemology (Nature of knowledge, relationship between knower and what can be known)

One "body of knowledge"

Objective is important

Researcher controls and observes in an objective dispassionate manner

Knowledge individually or socially constructed

Framework/values of researcher acknowledged/made viable/explicit

Interactive link between researcher and participants

Knowledge is socially, historically, politically, culturally situated

Interactive/activist link between researcher and participants/context

Methodology (Purpose)

Predict

Test

Measure

Prove

Disprove

Understand

Describe

Construct meaning

Understand from participants’ perspectives

Promote social change

Liberate

Emancipate

Critique

Take political action

Methodology (Purpose)

Quantitative

Interventionist

Deductive

Design

Single group,

Experimental

Quasi-Experimental, etc.

Qualitative

Inductive (discovery of patterns)

Hermeneutical (Interpretive)

Dialectical

Contextual features important

Qualitative (Primarily)

Quantitative (Can be used)

Contextual/historical features important as they relate to oppression

Axiology (Value and Judgement)

Value free/theoretically influenced

Suspend judgement until statistical tests prove/disprove

Judgement is based upon consensus of participants and researcher

Varies upon theoretical framework/values held by researcher

Judgement is based on experienced oppression by participants

Framed by beliefs/values of all participants

Can be theory driven

Figure 3.1: Philosophical grounding of paradigms in research – adapted "freely" from Mortens (1998), Kmitta (2000) and Guifoyle (2005)

Methods of data collection

There are two types of data; primary data - collected for the first time and secondary data - those which have already been collected and analysed by someone else. Kumar (2005:142) explains that interviewing, observation and questionnaires are the three main methods of data collection classified under primary sources. All other sources, where the information required is already available are called secondary sources. The choice of a particular method for collecting data is important itself for ensuring the quality of information.

The researcher used the tree types of data collection methods, Watkins (2008:53), (citing Emory and Cooper 1995), namely:

Personal interviewing;

Telephone interviewing;

Self-administered questionnaires/surveys.

In this data collection exercise, respondents were asked questions, in order to determine what they think, feel or did in terms of a particular statement as it related to project quality at Koeberg. An indirect approach was used with the personal and telephone interviews and this gave the researcher insight as to factors that the researcher was not consciously aware of as it relates to project quality and the plant. Personal interviews allowed for the identification of key issues within the target environment, which were not readily identifiable using a survey questionnaire. Through this, the researcher was able to interpret responses immediately and allow the respondent to elaborate on significant information. Telephone interviews, while there were few, were also used as a quick method which allowed the researcher to explain questions not understood by the respondent. This method was used where respondents did not have access to computers during the survey period.

In addition to personal and telephone interviews, this study used self-administered questionnaires. Since there are many ways to ask a question, the questionnaire was very flexible and attempted to cover crucial aspects of project quality throughout the project life cycle. There are three basic types of questionnaire: closed-ended, open-ended or both. Closed ended questions include all possible answers or prewritten response categories and respondents are asked to choose among them, while open-ended questions allow respondents to answer in their own words. The questionnaire was designed such that respondents had to indicate the level of agreement to a particular statement with the final question giving them the opportunity to provide general comment or critically express their overall view.

"Question design is determined by data to be collected. It is more appropriate to explore the nature of a problem, issue or phenomenon without quantifying it." (Saunders et al 1997:156). However self-administered questionnaires are also subject to a number of disadvantages (DeVaus 1996:108):

There is no control over who responds to the questionnaire and whether or not that person "consults" with colleagues while completing it;

The response rate may be low, giving rise to bias;

Misunderstandings cannot be cleared up;

Sampling is subject to error.

Since the researcher was mindful of the pittfals of self-adminstered questionnaires, an e-mail was sent out to identify persons who had to complete the questionnaire so as to ensure that the appropriate persons completed them. This was to overcome or minimise the disadvantages of self-administered questionnaires. The questionnaire was also pre-tested with colleagues who were conducting research to identify problems so as to avoid confusion in terms of the wording or layout. Follow up phone calls were made that served as a reminder for participants to complete the questionnaire.

The construction of a research instrument or tool for data collection is vital because determines the nature and quality of the information. Findings or conclusions are based upon the type of information and the data collected is entirely dependent upon the questions asked of respondents. The data collection methods used in the survey falls within the context of a survey, defined as: "A sample of subjects being drawn from a population and studied to make inferences about the population" Watkins (2008:59) citing Hussey and Hussey (2003). Questions were designed such that they cover the most crucial aspects of project quality throughout the project life cycle. This ensured that questions had a direct link to the objective of factors affecting project quality at Koeberg. The researcher employed all three primary data collection methods with self-administered questionnaires serving as the primary data collection method. This allowed the researcher to probe deeply and uncover new clues, open up new dimensions round the factors that affect project quality at Koeberg and to secure accurate and inclusive accounts that are based on personal experiences of those in the employ of NPM, whether permanent, temporary, seconded or contracted. Please refer to Appendix XXX for an example of the questionnaire.

Ethical consideration

According to Saunders et al (2000:103),"… ethics refers to the appropriateness of your behaviour in relation to the rights of those who become the subject of your work, or are affected by it". The following will be the Researcher’s behaviour and conduct guide when conduction the research:

Informed consent: Participants should be given the choice to participant or not to participate and furthermore be informed in advance about the nature of the study.

Right to privacy: The nature and quality of participant’s performance must be kept strictly confidential.

Honesty with professional colleagues: Findings must be reported in a complete and host fashion, without misrepresenting what has been done or intentionally misleading others as to the nature of it. Data may not be fabricated to support a particular conclusion.

Confidentiality / anonymity: It is good research practice to offer confidentiality or anonymity, as this will lead to participants giving more open and honest responses (Saunders et al., 2001:103).

To this end the researcher obtained a letter of consent from the Senior Manager of Nuclear Project Management granting permission for such study to be conducted in this business area. The researcher further issued an informed consent letter making all respondents adequately aware of the type of information that is required of them, why it is being sought, what purpose it will be put to, how they are expected to participate in the study and how it will directly or indirectly affect them. The consent was voluntary and without pressure of any kind. Please refer to Appendix XXX for an example of the informed consent letter that accompanied each questionnaire.

Determining sample design and the choice of sampling method

Researchers usually draw conclusions about large groups by taking a sample. A sample is a segment of the population selected to represent the population as a whole. Ideally, the sample should be representative and allow the researcher to make accurate estimates of the thoughts and behaviour of the larger population. The survey design asks:

Who will be surveyed? (Sample)

How many people will be surveyed? (Sample size)

How should the sample be chosen? (Sampling)

There are three types of probability samples:

Simple random sample: Every member of the population has an equal chance of being selected;

Stratified random sample: Population is divided into mutually exclusive groups and random samples are drawn from each group;

Cluster sample: The population is divided into mutually exclusive groups and the researcher draws a sample of the group to interview. Find a reference

Hussey and Hussey (1997:148) are of the opinion that there is no ideal of prescribed sample size as it depends on the discipline, the level of confidence expected in the answers and the anticipate response rate. In order that each identifiable stratum of the population is taken into consideration (Collis and Hussey, 2003:157, Easterby-Smith, Thorpe and Lowe, 1996:48), respondents were randomly selected from each stratum.

Corbetta (2003:210-218) is of the opinion that with probability samples:

Each unit has a non-zero probability of selection;

Probability of selection for all units is known; and

The selection is completely random.

Corbetta (2003:210-218) further believes that the accuracy of sampling estimates depends on, among other things, sample size and the degree of variability in the distribution of the phenomenon studied within the reference population. "…the two elements that are more important that any others in survey research are randomisation and bias" Leedy and Omrod (2005:208)

The various functional areas, listed in paragraph 3.2, served as the individual strata for the research survey and ensured that all identifiable strata of the population were taken into account. Respondents were divided into categories in terms of the project life cycle, which served as the individual strata for the survey. The project lifecycle model for Eskom is:

Concept Release Approval (CRA)

Definition Release Approval (DRA)

Execution Release Approval (ERA)

Finalisation Release Approval (FRA)

The sampling method utilised by the researcher allowed for stratified sampling of the participants. The sample design was organised in three phases as per Corbetta (2003:210-218):

The reference population of NPM, which currently has 84 employees (both technical and non-technical), was subdivided into sub-populations called strata (paragraph 3.2) that are homogeneous as possible to be studied

A sample was selected from each stratum by means of random procedure

The sample drawn from each stratum was pooled in order to produce an overall sample

It is to be noted that while the above external departments have a huge and separate staff compliment, only those staff members seconded to NPM or those working on the larger projects as part of the project team formed part of the sample. They are therefore considered as part of NPM for the purposes of this research.

Survey design

Watkins (2008:140), (citing Collis and Hussey 2003), believes that if research is to be conducted in an efficient manner and make the best of opportunities and resources available, it must be organised. Furthermore, if it is to provide a coherent and logical route to a reliable outcome, it must be conducted systematically, using appropriate methods to collect and analyse the data.

The researcher used the descriptive survey design and according to Watkins (2008:140), (citing Collis and Hussey 2003), a descriptive survey is frequently used in business research in the form of attitude surveys. The descriptive survey as defined by Ghauri, GrØnhaug and Kristianslund (1995:60), has furthermore the characteristics to indicate how many members of a particular population have a certain characteristic.

The statements within the survey have been designed with the following principles in mind (Watkins, 2008:143):

Avoidance of double-barrelled statements;

Avoidance of double-negative statements;

Avoidance of prestige bias;

Avoidance of leading statements;

Avoidance of the assumption of prior knowledge.

Questions and statements were formulated in such a manner that respondents could respond to the questionnaires ranging from strongly agree to strongly disagree, and to determine the factors that affect project quality at a nuclear power plant from various perspectives.

Target population

Lavrakas (2008) states that a target population for a survey is the entire set of units for which the survey data are to be used to make interpretations. In any survey the target population must be clearly defined. Collis and Hussey (2003:155-160) define a population as follows: "A population is any precisely defined set of people or collection of items which is under consideration". A sampling frame is defined as a list or record of the population from which all the sampling units are drawn.

For this survey, the researcher selected a random sample of respondents within the entire NPM population who fitted the profile and represented the sampling frame. As it was not necessary to sample the entire population of 84 employees; 46 employees consisting of project managers, project leaders, project supervisors, contracts managers, buyers and quality assurers formed the sample size. It is to be noted that while the external departments have a huge and separate staff compliment, only those staff members seconded to NPM or those working on the larger projects as part of the project team formed part of the target population. They were therefore considered as part of NPM for the purposes of this research. The sample also included 4 contractors, the business partners in ensuring Eskom meet their objectives of keeping the lights burning. The contractors that were targeted are those who have the most influence and impact on project quality. This transposed into different numbers of respondents from different functional areas being randomly selected from the identified research strata ensuring representativeness as the sample across the various departments, within NPM, that influence project quality at Koeberg. This target population was specifically chosen to validate the practicality of the concepts as presented within this research. This sample is therefore considered a representative of the target population. The risk of bias, which cannot be statistically eliminated, was recognised by the researcher based on the very definition of the target population as well as the limitations introduced by the number of respondents selected in each category. To this end, the researcher took the necessary steps to not deliberately conceal or highlight something that could be seen as the researcher introducing a vested interest thereby not drawing conclusions to the best of the researcher’s ability.

Page Done

Measurement scales

According to Kumar (2005:146-151) attitudinal scales measure attitudes towards an issue and their strength lies in their ability to combine attitudes towards different aspects of an issue and to provide an indicator that is reflective of an overall attitude. There are three types of scale that measure attitude: the Likert, Thurstone and Guttman scales. The Likert scale is the most common because it is easy to construct.

Kumar (2005:146-151) further explains that the Likert Scale does not measure attitude per se. but it does help to place different respondents in relation to each other in terms of the intensity of their attitude towards an issue; it shows the strength of one respondent’s view in relation to that of another.

During the survey respondents were asked to respond to statements based on the Likert scale. The reason for choosing the Likert scale is that it can be used in both respondent-centred (how responses differ between people) and stimulus-centred (how responses differ between various stimuli) studies. It was most appropriate to gather data in support of the research problem and best extract respondents’ views round factors that influence project quality at Koeberg. It was also structured such that respondents had to personalise and think of these factors when they implemented a project. The advantages in using the Likert scale, according to Watkins (2008:140), (citing Emory and Cooper 1995) are:

Easy and quick to construct;

Each item meets an empirical test for discriminating ability.

Qualitative research strategy

Greenhalgh and Taylor (1997:740-743) point to the positive characteristics that are inherent in qualitative research, amongst others, qualitative methods have the objective to make sense of or interpret phenomena in terms of the meanings people bring to them.

Golafshani (2003) in citing (Patton, 2001:39) explains that qualitative research uses a naturalistic approach that seeks to understand phenomena in context-specific settings, such as "real world setting where the researcher does not attempt to manipulate the phenomenon of interest"

Creswell (2003:182) argues that qualitative research is fundamentally interpretive with the researcher making interpretations of the data and then drawing conclusions about their meaning, personally and theoretically. Leedy and Ormrod (2005:134-135) believes that qualitative research comprises the following characteristics, namely: description; interpretation; verification and evaluation. Furthermore, the qualitative researcher collects data on an instrument or gathers information on a behavioural checklist (Creswell, 2003:17).

While the researcher acknowledges that a number of strategies can be applied in similar research projects, the well-known concepts of objectivity, reliability, etc., inherited from the empirical analytical paradigm, are suggested for business research in more or less the traditional way. The researcher used qualitative methods of research to unearth the factors that affect project quality at Koeberg.

Survey sensitivity

There were particular challenges to the researcher, especially in research conducted in areas of a sensitive nature such as a nuclear power plant. This survey as well as the empirical data gleaned by interrogating the databases speaks to the factors that affect project quality at a nuclear power plant. This survey is therefore sensitive in nature and the following guidelines from various academics serve to illustrate the mitigation process, which can be deployed in an instance where research is conducted in areas of a sensitive nature:

Oskowitz and Meulenberg-Buskens (1997:83), conducted a qualitative investigation of a particularly sensitive nature. They stated the importance of handling critical issues in these words: "Thus any type of qualitative investigation could benefit from the researchers being skilled and prepared, and the sensitive nature of an investigation into a stigmatising condition made the need for such an undertaking even more imperative in the current study".

Meulenberg-Buskens (1997:94), held the view that the sensitivity of certain issues that impacted negatively on the research in the environments being evaluated required intimate personal involvement. It also demanded ‘personal and practical experience’ by the researcher. They regarded it as imperative in order to support the authenticity of the qualitative research that is to be performed. Checkland (1989:152), elaborates that the "… researcher becomes a participant in the action, and the process of change itself becomes the subject of research"

The validity and reliability

It is the understanding of most scholars that validity means that correct procedures have been applied to find answers to a question while reliability refers to the quality of a measurement procedure that provides repeatability and accuracy. Emory & Cooper (1995:156) states that a strategy of empirical analysis of data collected that is used in business research will deliver results that are:

Practical: Results will be economical, convenient, and interpretable;

Valid: The extent to which the test measures that we actually wish to measure represent the real situation. There are three subsets to the concept of validity. These are construct validity, internal validity and external validity (Yin, 2003:34).

Reliable: The accuracy and precision of the measurement procedure.

Saunders et al (1997:156) state that the validity and reliability of collected data depend on the design of the questions, the structure of the questionnaire and the diligence of pilot testing. The questionnaires for this research were designed to achieve the research objectives and obtain additional information on factors that affect project quality at Koeberg.

Golafshani (2003:598-599) in citing Joppe (2000:1) defines reliability as: …"The extent to which results are consistent over time and an accurate representation of the total population under study is referred to as reliability and if the results of a study can be reproduced under a similar methodology, then the research instrument is considered to be reliable. The author further states that validity determines whether the research truly measures that which it was intended to measure or how truthful the research results are.

Leedy and Ormrod (2005:28) believe that the validity and reliability of measurement instruments influence the extent to which something can be learned about the phenomenon being studied, the probability that statistical significance in data analysis will be obtained and the extent to which meaningful conclusions can be drawn.

Kumar (2005:147-151) therefore believes that it is important for a researcher to attempt to establish the quality of results as other researchers can also ask you to establish the validity of the procedures adopted for finding answers to research questions. As inaccuracies can be introduced into a study at any stage, the concept of validity can thus be applied to the research process as a whole or to any of its steps. Broadly, there are two perspectives on validity:

Is the research investigation providing answers to the research questions for which it was undertaken?

If so, is it providing these answers using appropriate method and procedures?

Davies (2007:241) states that the concept of reliability is related to the rigour with which the researcher has approached the tasks of data collection and analysis with reliability being equated with methodological ‘accuracy’.

Reliability and validity are conceptualised as trustworthiness, rigor and quality in qualitative paradigm. It is also through this association that the way to achieve validity and reliability of a research get affected from the qualitative researchers’ perspectives which are to eliminate bias and increase the researcher’s truthfulness of a proposition about some social phenomenon using triangulation. Then triangulation is defined to be "a validity procedure where researchers search for convergence among multiple and different sources of information to form themes or categories in a study" Creswell & Miller (2000:126). Therefore, reliability, validity and triangulation, if they are to be relevant research concepts, particularly from a qualitative point of view, have to be redefined in order to reflect the multiple ways of establishing truth.

The researcher will ensure validity and reliability by the use of the method called triangulation so as to have an understanding as to how to test or maximise the validity and as a result the reliability of this study. If the validity or trustworthiness can be maximized or tested then more "credible and defensible result" Johnson (1997:283) may lead to generalisability. Triangulation is typically a strategy or test for improving the validity and reliability of research or evaluation of findings. Mathison (1988:13) elaborates this by saying: "Triangulation has risen as an important methodological issue in naturalistic and qualitative approaches to evaluation [in order to] control bias and establishing valid propositions because traditional scientific techniques are incompatible with this alternate epistemology." Patton (2001:247) advocates the use of triangulation by stating "triangulation strengthens a study by combining methods. This can mean using several kinds of methods or data, including using both quantitative and qualitative approaches" Furthermore triangulation may include multiple methods of data collection and data analysis, but does not suggest a fix method for all the researches. The methods chosen in triangulation to test the validity and reliability of a study depend on the criterion of the research.

Validity and reliability of measurement instruments influence the extent to which something can be learned about the phenomenon being studied, the probability that statistical significance will be obtained in data analysis and the extent to which meaningful conclusions can be drawn from collected data. For this survey the researcher has developed 2 separate survey questionnaires which contain statements requiring a response that has been designed to measure the attitude towards and experience of respondents within NPM who have direct influence on project quality. One questionnaire focussed on the NPM project staff while the other was designed for the contractors. These questionnaires were issued to Senior Management, Project Managers, Project Leaders/Supervisors, Contract Managers, Quality Assurers, Buyers and Contractors within NPM. These respondents influence project quality in different ways so it will be accurately measure factors that affect project quality at a nuclear power plant from the NPM and contractors’ perspective.

Data analysis

In accordance with Leedy and Omrod, data analysis will be conducted in the following manner:

Organisational: The researcher will use software to assemble, categorise, code, integrate and search potentially huge data sets (e.g., survey open-ended responses, qualitative interview data). Data will be organised. Coded by marking the segments of data with symbols, descriptive words, or category names. The researcher will continue this process until all data is segmented and initial coding has been completed. A master list will be kept (i.e., a list of all the codes that are developed and used in the research study). Then the codes will be reapplied to new segments of data each time an appropriate segment is encountered. Once complete, the data will be integrated.

Conceptual: The researcher will use software to write and store ongoing reflections about data and construct theories that integrate research findings. Here the researchers will transcribe data from questionnaires, interviews, observational notes, memos, etc. into word processing documents. It is these transcriptions that are later analysed using a qualitative data analysis computer programme. The researcher will also record reflective notes about what is being learnt from data so when ideas and insights are gained, include those as additional data to be analysed.

Statistical: The researcher will utilise statistical and spread sheet software packages used to categorize and analyse various types of data sets

Graphic production: The researcher will employ software used to depict data in graphic form to facilitate interpretation. These graphical images will illustrate how project quality is meant to work and clarify the relationship between project quality and the factors that affect project quality at Koeberg.

Corroborating and validating of results: This is an essential component of data analysis and the qualitative research process and should be done throughout the qualitative data collection, analysis, and write-up process. This is essential because in presenting trustworthy results otherwise, there is no reason to conduct a research study. The results will be validated using a computer programme.

Conclusion

When you say that you are undertaking a research study to find answers to a question, you are implying that the process;

is being undertaken within a framework of a set of philosophies ( approaches);

uses procedures, methods and techniques that have been tested for their validity and reliability;

is designed to be unbiased and objective .

The researcher will ensure that the research is effective in the following ways:

Rigorous: the researcher will be scrupulous in ensuring that the procedures followed to find answers to questions are relevant, appropriate and justified;

Systematic: the researcher will adopt a procedure that follows a certain logical sequence;

Valid and verifiable: The researcher will ensure that whatever conclusions are made on the basis findings is correct, valid and can be verified;

Empirical: The researcher will ensure that conclusions drawn are based on hard evidence gathered from information collected from real life experiences, observations and relevant databases and archives;

Critical: The researcher will employ critical scrutiny of the methods employed to ensure they are fool proof and are able to withstand critical scrutiny.

In this chapter, the ‘knowledge management’ survey design and methodology was addressed and the researcher believes that the above will be covered in order to ensure that this research is effective