Non Sampling Errors

There are certain sources of errors which occur both in sample survey as well as in the complete enumeration. These errors are of common nature. Suppose we study each and every unit of the population. The population parameter under study is the population mean and the ‘true’ value of the parameter is \mu which is unknown. We hope to get the value of \mu by a complete count of all the units of the population. We get a value called ‘calculated’ or ‘observed’ value of the population mean. This observed value may be denoted by {\mu _{{\text{cal}}}} . The difference between {\mu _{{\text{cal}}}} and \mu (true) is called non-sampling error. Even if we study the population units under ideal conditions, there may still be the difference between the observed value of the population mean and the true value of the population mean. Non-sampling errors may occur due to many reasons. Some of them are:

  • The units of the population may not be defined properly. Suppose we have to carry out a study about skilled labor force in our country. Who is a skilled person? Some people do more than one job. Some do the secretariat jobs as well as the technical jobs. Some are skilled but they are doing the job of un-skilled worker. Thus it is important to clearly define the units of the population otherwise there will be non-sampling errors both in the population count and the sample study.
  • There may be poor response on the part of respondents. The people do not supply correct information about their income, their children, their age and property etc. These errors are likely to be of high magnitude in population study than the sample study. To reduce these errors the respondents are to be persuaded.
  • The things in human hand are likely to be mishandled. The enumerators may be careless or they may not be able to maintain uniformity from place to place. The data may hot be collected properly from the population or from the sample. These errors are likely to be more serious in the population data than the sample data.
  • Another serious error is due to ‘bias’. Bias means an error on the part of the enumerator or the respondent when the data is being collected. Bias may be intentional or un-intentional. An enumerator may not be capable of reporting the correct data. If he has to report about the condition of crops in different areas after heavy rainfalls, his assessments may be biased due to lack of training or he may be inclined to give wrong reports. Bias is a serious error and cannot be reduced by increasing the sample size. Bias may be present in the sample study as well as the population study.