4.8 Pilot Study
A pilot test is considered to be like preparation in which a small – sample of the study is tested prior to testing the whole sample (Gorondutse & Hilman, 2012b; Gay, Mills & Airasian, 2006). Thus, this study carried out pilot test to (i) assess validity and internal consistency of instrument; and (ii) find out real condition of the main study, thereby enabling researcher to predict and cater potential issues during full scale research. Validity is referring to degree of which instrument is measuring what it is supposed to measure and not something else, whereas reliability measure the extent to which an instrument is error free, and hence consistent and stable across time and also across various items in the scale (Sekaran & Bougie, 2010).
The sample size for a pilot study is usually small in nature; starting from 15-30 respondents although it could be increased substantially if the test involved a number of levels (Malhorta, 2008). An attempt was made to conduct factor analysis but due to the nature of cases/samples it was not possible to conduct factor analysis. This has been stated in previous literature that factor analysis is a responsive to the number of cases/samples and hence, with a few sample sizes the correlation co-efficient among the variables is less reliable and tends to vary. Researchers have different views with regards to sample size that is sufficient for factor analysis. They agree that a larger sample is better.
Tabachinik and Fidell (2007) based on rule of thumb, suggest that a sample size of at least 300 is required, while others suggest a sample size based on the measure of subject to items. Nunnally (1978) suggests that measure of 10 cases for each item is enough for factor analysis. In line with the above information this study only tested for validity and reliability of the instrument. In any case, it is expected that having a valid and reliable instrument leads to the reduction of measurement error to a large extent.
4.8.1 Validity of the Instrument
In this section, a number of questions were re-worded/re-phrased in order to measure the appropriate variables and also to be reasonable to the potential respondents. Hence, the process of face validity looking for expert opinion was completed within two weeks’ time. Following the guidelines of using the scales laid down by the developers, minor modification was made in order to suit the location of this research and also to suit the sector as well. Based on the foregoing, more especially considering the fact that the instrument was verified by a number of fairly educated people who are well-known in with the area of the study, it could be said that the instrument is appropriate for the context of the study. Therefore, subsequent to obtaining opinions and the observations of the specialists the researcher developed an enhanced/revised version of the instrument, which was eventually administered for the pilot test (Gorondutse & Hilman, 2012). Total 70 copies of questionnaires were circulated and 62 were returned and measured for pilot test analysis (Gorondutse &Hilman, 2012). A returned completed questionnaire was received after the cut-off dateline and thus, was not incorporated in the trial test analysis. Hence, high response rate of about 88.6% was achieved perhaps because of distributing and collecting the questionnaire personally. Some friends of the researcher assisted in the distribution and collection of questionnaires. This process was completed within four weeks. It should be noted that the pilot study was done in the month of March, 2016 to April, 2016.
4.8.2 Reliability of the Instrument
There are different types of reliability tests; the most widely used technique in many researches is internal consistency reliability (Litwin, 1995). The Cronbach‘s alpha coefficient test was carried out to measure the internal consistency reliability. A pilot test was conducted with the respondents to test the reliability of the instruments. A total 350 respondents participated in the pilot study, but only 226 were returned and analysed. The following table shows the summary of the reliability results:
Table 4.8.2.1
Result of Reliability Analysis for Pilot Study
S/N | Variable | Variable Scale Statistics | N of Items | Cronbach’s Alpha | ||
Means | Variances | Std. Deviation | ||||
1. | Pressure (P) | 55.80 | 124.836 | 11.173 | 17 | .850 |
2. | Opportunity (OP) | 41.05 | 63.828 | 7.989 | 11 | .867 |
3. | Rationalization (RT) | 18.59 | 15.469 | 3.933 | 5 | .760 |
4. | Capability (CAP) | 16.64 | 18.143 | 4.260 | 5 | .720 |
5. | Islamic Corporate Governance (ICG) | 40.4823 | 44.935 | 6.70338 | 10 | .882 |
6. | Fraud (FRD) | 20.27 | 13.938 | 3.733 | 5 | .756 |
The results of pilot test in Table 4.8.2.1 indicate that the Cronbach‘s alpha for the variables under examination are all above 0.70. Therefore, based on the recognized yardstick of 0.70 all the variables have internal consistent reliability and consequently there was no need to remove any items. Further reliability analysis was performed in the actual study based on larger samples size (refer to chapter 5).
Tagged with: Banking Fraud, Capability, Islamic Corporate Governance, Moderating Role, Opportunity, Pressure, Rationalization