• Tidak ada hasil yang ditemukan

According to Kimberlin and Winterstein (2008:2276), the key elements in measuring the quality of data collection instruments are the reliability and validity of the measures. These elements are used to reduce error in the data collection process. These elements are broken down below to provide clearer perspectives of what they. Morse et al. (2002:13) state that research is insignificant if it is not executed thoroughly. Great emphasis is placed on reliability and validity, and without these elements, research becomes fictional, and its value remains insignificant (Morse et al., 2002:14).

3.10.1 Reliability

Heale and Twycross (2015:66) define reliability as the steadiness of a measure. Reliability relates to the results of the information acquired from a source that is constant over a period, which is dependable and can provide the conclusions of a grouping that represents the population in question (Boyatzis, 1998). Heale and Twycross (2015:66) further explains that a researcher distributing data collection tools designed to measure a certain phenomenon should receive more of the same replies at the end of each data collection exercise. To analyse the internal stabilities of the research measures within the ambit of this study, the coefficient of Cronbach’s alpha was used to calculate the values for each element within each theme, and each theme overall. According to Kimberlin and Winterstein (2008:2277), reliability coefficients range from 0.00 to 1.00, with higher coefficients signifying increased reliability. The questionnaire used to collect data should, according to Nunnally (1978:245) represent a consistency of 0.700 reliability or better.

Table 3.1. Attributes of Reliability

Attributes Description How the attribute was achieved for this study

Homogeneity (or internal consistency)

The extent to which all the items on a scale measure one construct

Cronbach’s Alpha was used to test for internal consistency. Collectively, the scores are justified in Table 4.34 in Chapter 4.

52

Attributes Description How the attribute was achieved for this study

Stability

The consistency of results using an instrument with repeated testing

Cronbach’s Alpha was used to test the reliability of stability. Collectively, the scores are justified in Table 4.34 in Chapter 4.

Equivalence

Consistency among responses of multiple

users of an

instrument, or among alternate forms of an instrument

Cronbach’s Alpha was used to test the reliability of Equivalence. Collectively, the scores are justified in Table 4.34 in Chapter 4.

Source: Heale and Twycross (2015:66)

3.10.2 Validity

According to Kimberlin and Winterstein (2008:2278), Validity is the degree to which an instrument evaluates what it purports to measure. Boyatzis (1998) also defines Validity as the results or information that is the closest to the truth, or falsity of a targeted area of research. Spearman’s Rank Correlation Coefficient was used as the measure to determine the Validity for this research study, of which the results indicate significant correlation between the elements.

7The Spearman’s Rank Correlation Coefficient, can take a value between +1 to −1 where, A ⍴-value of +1 means a perfect association of rank.

7 https://www.questionpro.com/blog/spearmans-rank-coefficient-of-correlation

53

A ⍴-value of 0 means no association of ranks. Closer the ⍴-value to 0, weaker is the association between the two ranks.

A ⍴-value of −1 means a perfect negative association between ranks.”

Table 3.2 The strength of relationships

Size of r Interpretation

± (0.000 - 0.190) very weak or no relationship

± (0.200 - 0.390) weak relationship

± (0.400 - 0.590) moderate relationship

± (0.600 - 0.790) strong relationship

± (0.800 – 1.000) very strong relationship

+(1.000) Perfect association between ranks

Heale and Twycross (2015:66) define three types of validity in Table 3.3

Type of

validity Description How this was achieved for this study

Convergent Validity

The extent to which a research instrument accurately measures all aspects of a

Convergent Validity was measured using Spearman’s Rank Correlation Coefficient.

The results are depicted in Chapter 4, Table 4.35.

54 Type of

validity Description How this was achieved for this study

construct

Construct Validity

The extent to which a research instrument (or tool) measures the intended construct

Construct Validity for this study was measured using Cronbach’s Alpha, who measures the reliability of the data collection tool. In this case, the reliability data collected was also determined by Cronbach’ Alpha, and the values are stated in Chapter 4, table 4.34.

Discriminant Validity

The extent to which a research instrument is related to other instruments that measure the same variables

Discriminant Validity was measured using Spearman’s Rank Correlations Coefficient. The results are depicted in Chapter 4, Table 4.35.

Source: Heale and Twycross (2015:66) 3.11 Ethical consideration

The researcher considered the following ethical considerations:

Informed consent: The researcher sought the consent of research participants; thus, no participants were coerced in participating. Further, to ensure that participants were aware of what they are participating in, the researcher clearly explained the nature of the study and its objectives to the participants before collecting information from them. All information was obtained/collected by voluntary participation of participants, and at their discretion.

55

Harm to participants: The researcher ensured that this study will not result in any harm or damage to the research participants and their related companies. This study, therefore, is purely for academic purposes and while no harm can be foreseen, the researcher will not be held responsible for unforeseen circumstances that could result in harm. All information/data collected are treated confidentially and remain private.

Invasion of privacy: The researcher ensured that no private information is used for the study unless authorised by the participants. As such, the researcher requested exclusive permission where the data obtained is private. Further to this, the researcher did not use the data gathered for the personal benefit of the researcher. The study was conducted in adherence to the University’s ethical standards.