Collection errors, like when the interviewee gives incorrect information, or processing errors, like when the interviewer
registers the answer incorrectly, are tried kept to a minimum by automatically controlling the answers during the interview.
Nevertheless, there may still be some errors left. The number and kind of errors are not known.
The unit non-response has inceased somewhat over the years, and is approximately 40 per cent. Non-response errors are errors
caused by unit non-response, i.e. that the unit (for instance an individual or a company) has failed to respond, or item non-response,
i.e. that the unit has failed to respond to some but not all the questions in the survey.
Bias
Sampling bias may arise when the distribution of some variables in the sample is not the same as the corresponding distribution
in the population. The results in this survey are based on the information provided by those responding. Non-response is mainly
due to the interviewee not wanting to participate or that he/she was not reached by the interviewer. If the interviewee was
not reached due to he/she being on holiday, the distribution of this variable among those interviewed and the non-respondents
may differ and sampling bias may arise.
Variance
The Travel survey is carried out among a sample of persons resident in Norway between 16 and 79 years of age. The uncertainty
of findings based on only a part of the population, (sampling variance) depends among other factors on the number of observations
in the sample and on the distribution of the current variable in the whole population. An estimate of this uncertainty (standard
deviation) can be constructed by means of the observations in the sample. There is not made an exact calculation of the standard
deviation for this survey.
