Reliability

Reliability

A reliable questionnaire produces consistency in findings when measurements are recorded at two different points in time (Gray, 2009). The reliability of a Data Collection Services relates to the consistency of this instrument (Bryman & Cramer, 2001). In relation to the data collection methods used in this research, Saunders et al. (2000) and McKinnon (1988) listed a number of factors that are likely to threaten reliability including: subject error; subject bias; observer –caused effects and observer bias. There follows a discussion of each factor and the procedures carried out in this research to counteract these threats to reliability.

Firstly, subject error refers to the tendency of the respondents to provide responses that differ from the true facts. This is most likely to happen if the researcher does not choose an appropriate time during the day to collect data (Saunders et al., 2000). As an illustration, if the data is collected early at the beginning of a working day, respondents may be keen to respond. Whereas, collecting data at the end of a working day is likely to drive respondents to provide irrational responses because they suffer fatigue resulting from workload which subsequently will affect the reliability of the data collected. To overcome this threat, the researcher tried to chose ‘neutral’ times for data collection when respondents were neutral in their feelings (e.g. during midday) when this was possible to make.

Secondly, subject bias refers to the tendency of respondents to provide responses that differ from the true facts because they are obliged to do so or due to the firm’s policy which restricts publishing sensitive or confidential information (Saunders et al., 2000). To overcome this threat, the researcher assured the respondents that both data collected from the questionnaire and the open ended questions would be analysed with complete confidentiality and anonymity and would not be used for other purposes than this research. Thirdly, observer-caused effects are those effects which result from the observer’s (i.e. the researcher’s) presence in the phenomenon under study and which are likely to influence the respondent’s behaviour, conversation, and data he/she provides. This type of threat occurs when the role attributed to the researcher by the respondents is such that it drives them to change their normal behaviour (McKinnon, 1988). To overcome this threat, questionnaire were preceded by opening statements and clarification of the role of the researcher in order to build confidence and trust between the researcher and the respondents (Saunders et al., 2000).

Further, of “paramount importance” (Cohen et al., 2007) to reliability is design, and wording of the questionnaire. Two process to aid the design and piloting and sampling. In the present study in order to taken care of reliability, pilot study was conducted among 5 teachers to obtain a feedback on format, inter alia, perceptions on sensitive questions, time for completion, redundant items, and ambiguities in wording. Based on the results of pilot study some of the items questions are reworded and two questions are combines considering the time. Admittedly, even if the aforementioned procedures were carried out in order to counteract the influence of subject error, subject bias, observer-caused effects and observer bias on the requirements of reliability, absolute reliability cannot be guaranteed.