Instrument Reliability, By Saunders Et Al (2009)
3.8.2 Instrument reliability
Saunders et al (2009) contend that although for a questionnaire to be valid it must be reliable, this on its own is not sufficient. They note that a situation may arise where respondents may consistently interpret a question in a questionnaire in one way, when the researcher meant something else. Therefore, even though the question may be reliable, it does not really matter as it has no internal validity and hence will not enable the research question to be answered. They thus explain that reliability is concerned with the strength of a questionnaire and, in particular, whether or not it will produce consistent findings at different times and under different conditions, such as with different samples or, in the …show more content…
The survey questionnaires were administered to a group of mobile phone users with a view to determining their ability to answer the questions in the instrument. A sample of 40 subscribers were used for the pilot test since previous studies (Ngugi, 2012; Igoki, 2011) indicate that the size of the pilot group may range from 20 to 100 subjects or be 1% of the target population. Construct validity was tested by subjecting the data to a convergence and discriminant validity test and the results showed an Alpha coefficient greater than 0.7. To assess the consistency of the obtained data, and how consistent they are for each individual from one administration of an instrument to another as well as from one set of items to another, the study used Cronbach’s alpha coefficient test on data obtained from all the pilot test respondents (Magutu, 2013). The overall Alpha coefficient for the sample was set at a recommended value of at least 0.70 and compared with those obtained from pilot studies in previous studies. All the study’s Cronbach’s alpha coefficients were greater than 0.70 as shown on Table 3.1 below. The results obtained thus indicated that the study’s research instrument is reliable. These findings are consistent with results from previous studies by Namada (2013) and Magutu (2013) which showed an alpha coefficient greater 0.7, indicating reliability of the research instruments used. Another study by Osoro (2012) showed that even a minimum alpha score of 0.6 is
Saunders et al (2009) contend that although for a questionnaire to be valid it must be reliable, this on its own is not sufficient. They note that a situation may arise where respondents may consistently interpret a question in a questionnaire in one way, when the researcher meant something else. Therefore, even though the question may be reliable, it does not really matter as it has no internal validity and hence will not enable the research question to be answered. They thus explain that reliability is concerned with the strength of a questionnaire and, in particular, whether or not it will produce consistent findings at different times and under different conditions, such as with different samples or, in the …show more content…
The survey questionnaires were administered to a group of mobile phone users with a view to determining their ability to answer the questions in the instrument. A sample of 40 subscribers were used for the pilot test since previous studies (Ngugi, 2012; Igoki, 2011) indicate that the size of the pilot group may range from 20 to 100 subjects or be 1% of the target population. Construct validity was tested by subjecting the data to a convergence and discriminant validity test and the results showed an Alpha coefficient greater than 0.7. To assess the consistency of the obtained data, and how consistent they are for each individual from one administration of an instrument to another as well as from one set of items to another, the study used Cronbach’s alpha coefficient test on data obtained from all the pilot test respondents (Magutu, 2013). The overall Alpha coefficient for the sample was set at a recommended value of at least 0.70 and compared with those obtained from pilot studies in previous studies. All the study’s Cronbach’s alpha coefficients were greater than 0.70 as shown on Table 3.1 below. The results obtained thus indicated that the study’s research instrument is reliable. These findings are consistent with results from previous studies by Namada (2013) and Magutu (2013) which showed an alpha coefficient greater 0.7, indicating reliability of the research instruments used. Another study by Osoro (2012) showed that even a minimum alpha score of 0.6 is