Skip to main content

Advertisement

Table 1 Comparison of survey methods, 2011 NSW Population Health Survey and 2012 NSW Population Health Survey

From: Inclusion of mobile phone numbers into an ongoing population health survey in New South Wales, Australia: design, methods, call outcomes, costs and sample representativeness

Procedures and Protocols 2011 NSW Population Health Survey (Landline phone numbers) 2012 NSW Population Health Survey
Landline phone numbers Mobile phone numbers
Sample generation Landline RDD sample frame for each of the administrative strata were generated using “best fit” postcodes for the geography (exchange district and charge zone) associated with the Australian Communications and Media Authority (ACMA) phone number ranges for NSW [25]. The sample was then randomly ordered within each strata and each number was tested using proprietary software [26] to identify valid and invalid numbers. The resulting valid numbers were used for the study. Same as for previous landline survey The RDD mobile sample frame was developed using all known Australian mobile prefixes and then using proprietary software [27] each number was tested to identify valid and invalid numbers. A random sample of valid mobile numbers was then provided for the study.
Sample design Stratified two-stage cluster sample design, with: strata defined by health administration areas; simple random sampling of clusters (household telephone numbers) within each stratum; and simple random sampling of population elements (household residents) within each cluster. Same as for previous landline survey Two-stage cluster sample design with simple random sampling of the mobile telephone numbers (adult population element) and simple random sampling of children in household (child population elements).
Questionnaire The questionnaire included questions on: health behaviours, health status, social determinants, demographics (including number of adults and children in the household) and landline phone ownership ("How many residential telephone numbers do you have? Do not include mobile phone numbers or dedicated FAX numbers or modems."). The actual questions in the questionnaire are available on the survey website. Same as for previous landline survey except for the addition of two questions on mobile phone ownership ("How many mobile phone numbers do you personally have?" and "Is/are your residential telephone number/s listed in the White pages?") Same as for previous landline survey the addition of two questions on mobile phone ownership ("How many mobile phone numbers do you personally have?" and "Is/are your residential telephone number/s listed in the White pages?")
Sample 3000 persons per quarter with equal numbers in each of the strata 2000 persons per quarter 1000 persons per quarter
Ineligible Business landline numbers, non-NSW residential numbers Same as for previous landline survey Business mobile numbers, non-NSW residential mobile numbers or mobile numbers owned by a child under the age of 16 years.
Data collection Data collection was undertaken using SAWTOOTH WinCati version 4.2 and trained interviewers from the in-house NSW Ministry of Health’s CATI facility. Same as for previous landline survey Same as for previous landline survey
Calling protocol The interviewers rang the randomly ordered landline numbers consecutively to try and contact households and convince the household and the respondent to participate in the survey. Up to 12 attempts were made to establish contact and if possible secure an interview with the selected respondent within a household. Same as for previous landline survey The interviewers rang the randomly ordered mobile phone numbers consecutively to try and contact the owner of the phone. Because mobile numbers could be located anywhere in Australia initial calls were timed to accommodate different time zones across Australia. Up to 12 attempts were made to establish contact and if possible secure an interview with the mobile phone holder.
Participant selection One person from the household was randomly selected for inclusion in the survey. If the selected respondent was a child under the age of 16 years, a parent or carer completed the interview on their behalf. Same as for previous landline survey The mobile phone holder was selected. If the owner of the mobile phone was a parent of a child under 16 years of age they were asked at the end of the interview if they or the main carers would agree to being contacted at a later date to undertake an interview about one of their children chosen at random.
Weighting (probability of selection) Adjust for differences in the probabilities of selection among subjects (using household size and number of landline phones in household). Same as for previous landline survey except for the inclusion of ratio of landline sample to landline phone populations for each strata. Adjust for differences in the probabilities of selection among subjects (using number of mobile phones owned by respondent and ratio of mobile phone sample to mobile phone population and number of children in the household).