Internet Service Provider 2017

The Internet service provider survey is a postal survey of all businesses that meet the population selection criteria. Since 2016, data used by mobile phone Internet connections was collected and published. Population size The Internet Service Provider (ISP) Survey is a survey sent to all New Zealand-based Internet service providers. The target population for the ISP Survey in 2017 was 83 businesses. This decreased from 2016, when 92 businesses were surveyed. Such changes in the population can be explained by: • existing businesses merging or ceasing • businesses being excluded from the population after not reporting ISP activity in the previous survey This decrease is slightly offset by new business being created. Not all businesses identified in the survey population ultimately report ISP activity. Population scope Businesses are included in the population for the ISP survey if they reported activity as an internet service provider in the 2016 ISP survey. Further businesses are considered for addition to the population if their main business activity or main income sources include any of a list of internet related keywords, and their industry sector is one typical for an internet service provider (such as telecommunications). Such businesses will be added to the population if research indicates that they currently provide any internet connection services. Scope error is possible if a business acts as an ISP, but their reported industry sector is not typical of an ISP. This error is mitigated by subject matter experts manually adding businesses to the population once they are publically renowned. Response rates The overall target response rate for ISP 2017 was 85 percent, which was met with a response rate of 91 percent. Some businesses were identified as key units if their total number of connections made significant contribution in the previous ISP survey. The target response rate for key businesses was 100 percent, which was achieved. Non-Response The ISP Survey is a census; therefore the data is not subject to sample error. Unit non-response occurs when a business does not return the questionnaire. While weighting is commonly used in other Stats NZ surveys, it is not applied to the ISP Survey. This is because there are no external (non-survey) variables that allow us to group businesses in a way that they are likely to provide similar survey responses, and therefore be representative of one another. To minimise the impact of unit non-response on the outputs, key respondents are targeted with 100 percent response rate targets. Therefore, we do not expect overall figures to be significantly affected by unit non-response. Data for businesses that did not respond to the survey was not imputed. Item non-response is not applied to the ISP survey either, for the same reasons as above. en-NZ