100%(1)1 out of 1 people found this document helpful
This preview shows page 42 - 44 out of 91 pages.
understanded. Efforts were made to construct clear and precise items in order to avoidambiguity. Construct validity was measured by administering a few questionnaires to somerespondents and analyzing the results to evaluate whether the questionnaire measures what itis required to measure. Criterion validity was measured by analyzing outcome provided bythe data collected using the questionnaires. 3.6.1 Reliability of the Instruments Reliability is the extent or level of uniformity with which an instrument assesses a variable(Mugenda & Mugenda, 1999). In order to test the dependability of the tool to be used in thestudy, the split-half method was used where the investigator randomly divided all items thatintend to evaluate the same paradigm into two sets. The researcher administered the entiretool to a chosen number of people and worked out the count for either half. The split halfdependability assessment is basically the relationship between these two counts. Theresearcher randomly divided the test into two portions using an ordinary-out of the ordinaryapproach. Each half had the same number of questions as the other. Reliability index wascalculated using the Cronbach alpha (rα). A dependability of 0.8 and higher is said by manyscholars to be okay. rα = ( k ) ( 1 – Σαj2 ) k-1 α2 where αj = variance of one test item, k represents the total numberof test items, Σ is the sum 30
3.7 PILOT TESTING. To ensure that the instruments are reliable and valid, a pilot study involving ten (10)randomly selected respondents with similar characteristics as the target population obtainedfrom SMEs was carried out. This was based on the recommendations of Mugenda andMugenda (2003) that a sample of 10% is adequate for piloting. The pilot study assisted theresearcher in making the necessary corrections on the instruments to ensure the questionswere accurate and consistent and correctly capture the objectives of the study beforecarrying out the actual research. These respondents were not included in the final sample forthe study. This helped eliminate any typing errors. The questionnaire should be precise and to the pointto curb some respondents from refusing to comply and fatigue. The results of the pilot phasewas used to improve the questionnaire and offer feasibility test hence ensure improvedresults. After all this was done research process proceeded followed by screening forms thatwas randomly distributed to determine eligible respondents for the survey. The data shouldnot be collected for a period exceeding 4 weeks. In course of the research I randomlyapproached and administer the questionnaires to potential predetermined respondents withinKutus town. 3.8 DATA PROCESSING AND ANALYSIS Preliminary editing was done where the data was checked for accuracy and errorscommitted. Clarity and legibility of all questions was established and questions withambiguous responses eliminated. The collected data was examined by use of expressivestatistics which consisted of weighted average and percentages. Correlation and regressionwas used to find out the relationships between the variables. For the unstructured type of