• Tidak ada hasil yang ditemukan

Validity and Reliability are strongly related concepts that represent various measuring instrument features. In general, a measuring instrument can be accurate without being valid;

nevertheless, if a measuring instrument is valid, it is more likely to be accurate. On the other hand validity measures, how well a measuring instrument performs its function and is concerned with whether the measuring instrument measures the quality that it is supposed to measure (Sürücü, and Maslakçı 2020). On the other hand, reliability, as highlighted by Sürücü and Maslakçı (2020) refers to the consistency and stability of the measuring instrument used over time. In other words, it is the ability to measure instruments to produce similar results when used at different times.

The validity of a questionnaire, as emphasized by Bolarinwa (2015) can be set up utilizing a board of specialists, who investigate theoretical constructs as they appear in the questionnaire.

This type of validity exploits how well the possibility of a theoretical construct is represented in an operational measure (questionnaire). The validity of the questionnaire can be proven with the use of another study in a form of experiment or pre-examination and this proves how a well given measurement connect to one or more external criterion. These structures could be the criterion related to validity and concept validity. Similarly, reliability is the extent at which observation, questionnaire, test or any measurement method creates similar outcomes in recurrent applications (Bolarinwa 2015:198).

Qualitative research aims at fostering understanding a phenomenon in a specific context and not generalizing findings to a wider population. To establish reliability and validity in

113

qualitative research, the five criteria of credibility, trustworthiness, dependability, transferability and conformability are used (Koonin 2014). These criteria are briefly explained below:

i. Credibility is about the truth-value. The strategies to maintain credibility are: continues observation, prolonged engagement, member check and triangulations. In other words, interpretive research can be regarded as credible if its conclusions are believable to readers;

ii. Trustworthiness is a matter of persuasion whereby the researcher is assumed to have made practices apparent and thus, auditable;

iii. Dependability: research can be regarded as reliable or authentic if two researchers independently arrived at the same conclusions evaluating the same phenomenon using the same set of evidence;

iv. Transferability: interpretive research refers to the extent to which the results can be applied to different situations; and

v. Conformability: is about ensuring that the researcher has acted in good faith whilst acknowledging that total objectivity is impossible in social research. It also refers to the extent to which others can independently confirm the findings reported in the research (Korsteins and Moser 2018). Reliability ensures the extent to which the scores measured by the research instruments are reliable. That is why Kumar (2014) emphasized that reliability could be the ability to generate consistent measurements whenever we administer an instrument to similar population under the same circumstances and achieve similar results; we say that instruments are reliable, reliability (extent of accuracy) and unreliability (extent of inaccuracy).

The measuring instrument is the sum of both the 'true score' that is not known and the 'error' in the measuring process. If the margins of errors are small and reporting research results is high, the research will be undoubtedly helpful (Mohajan 2017). In this regard, it is important for a researcher to test his or her survey instruments before using it for data collection (Abbas 2015).

This will ensure the consistency of a measure in the results obtained (Davies and Hughes 2014).

Research is only supposed to be reliable when a new researcher would come up with similar results if they were to conduct the same research study in a similar method (Davies and Hughes 2014). To safeguard the reliability and validity of the research instruments of the present study, the research instruments were given to professional librarians/academic staff at the Renaud

114

Cecil Library at the University of KwaZulu-Natal, Pietermaritzburg Campus who checked for any ambiguous questions that could lead to biased answers. The section below discusses the pre-testing of research instruments.

4.10.1. Pre-testing of research instruments

Pretesting, as noted by Hilton (2015) is a method or strategy for testing that questions function as proposed and are understood by those people who are probably going to respond to them.

Pretesting, as asserted by Hurst, Arulogun, Owolabi, Akinyemi, Uvere, Warth and Oviagele (2015) includes reproducing the official data collection procedure on a small scale to find real issues about data collection tools, sessions and techniques.

Pre-testing and piloting help to find questions that do not make any sense to respondents or problems with the instruments that may lead to biased responses (Abbas 2015). Babbie (2013) added that pre-testing is more related to instrument testing like questionnaire and interview approaches. Hence, regardless of how careful the researcher designs an instrument for data collection, such as a questionnaire, there is always the possibility of error. The most reliable protection against such mistakes is to fully or partially pre-test the questionnaire.

Some of the advantages of pre-testing, as identified by Hurst et al. (2015) includes:

i. Pretesting offers the opportunity to revise study materials and data collection procedures to ensure that appropriate questions are being asked and that questions do not make respondents uncomfortable and confused because they combine two or more critical issues in a single question;

ii. The value of pretesting can lead to detecting errors in cross-cultural language relevance and word ambiguity, as well as discovering possible flaws in survey measurement variables; and

iii. Pretesting can also provide a warning about how or why a primary research project can fail by indicating where research protocols are not followed or not feasible.

Therefore, the researcher conducted a pre-test of the research instruments to ensure their validity and reliability. To do so along with the interview schedule, they were also pre-tested at the University of Abuja, Nigeria. The head of the medical library there was interviewed accordingly and at the same time 30copies of the questionnaires were administered to the medical students of the University with the help of a research assistant and all were retrieved.

115

The choice of the University of Abuja was informed by the fact that it does not fall within the zone where the actual study was conducted. In the process of doing the interview; the researcher was able to make observations of the study area as a non-participant observer.

The data collected was coded and subjected to statistical analysis to determine the reliability of the instrument and the internal consistency of the items. The test was conducted with the Statistical Package for the Social Sciences (SPSS IBM version 23). The reliability of the instrument was established in the computation of Cronbach Alpha, the Spearman-Brown coefficient and the Guttman Split-Half Coefficients. The output for the two tests is attached as Appendix14. The observed reliability index for Cronbach Alpha, were 0.872 for the Part1 and 0.913 for Part 2. The Spearman-Brown coefficient was 0.718 and the Guttman Split-Half coefficient 0.698. The obtained reliability coefficients indicated that the instruments could be considered reliable for the study.