Population and sampling
Population of the study is all types of road users in the western province that is because each individual in the western province comes under some category of road users. According to the department of census and statistics the entire population in the western province is 5865 000 (Registrar General's Department, 2010), most valid and reliable data would come from every member of the population this is very hard to achieve it is costly and time consuming (10 Step Guide to Sampling by RCU, 2010), for those reasons research team choose not to use a census therefore a sample of the population was selected to perform the focus group survey and the questionnaire.
...view middle of the document...
Table1: Sample composition for questionnaire
Road users Pedestrians Passengers Cyclists Riders Drivers(heavy vehicle) Drivers (other) Total
Colombo 20 15 10 5 15 20 85
Gampaha 20 10 5 10 10 20 75
Kalutara 10 5 5 5 5 10 40
Total 50 30 20 20 30 50 200
The sample sizes were between 5-30 ensuring each exceeds the minimum sample size of 5 and allows for none response. A relatively larger sample size was selected for pedestrians and drivers(other) in the city as they emerge as the yardstick when determining the influence advertising have over behavior of general public in terms of road safety. It is also true they demonstrate an array of different road user behavior as opposed to some types in the survey and hence a large size. They also emerged as great source of information when researching into road safety issues. No apparent variation is was expected in the other categories in terms of road behavior and their response rate was relatively high. Therefore a small sample size had been taken from these categories.
Although it is likely to see the whole western province as the population of the research the research aim is based on the effectiveness of advertising road safety thus need for a large sample was not needed to measure the effectiveness of advertising.
By using a sample to estimate the value of a parameter in the population a sampling error occurs. Sampling error is the difference between the sample statistics and the actual statistics .which is always present until a census is carried out. This is inherent in the process of sampling and is reflected in the in the accuracy of the estimates about the total population. Due to this a disproportionate number of uncharacteristic observations are made. While the non-sampling errors was reduced through in-depth briefing, careful analysis of inconsistency and by providing thorough field guide lines, the sampling error was greatly eliminated through calculation...