Landline Sample Research Articles

Experimenting with QR Codes and Envelope Size in Push-to-Web Surveys

Abstract Survey researchers are continually evaluating approaches to increase response rates, especially those that can be implemented with little or no costs. In this study, we experimentally evaluated whether or not including Quick Response (QR) codes in mailed recruitment materials for self-administered web surveys increased web survey participation. We also assessed whether mailing these materials in a non-standard envelope size (6 × 9 inch) yielded a higher response rate than invitations mailed in a standard, #10 envelope (4.125 × 9.5 inch). These experiments were embedded in a sequential mixed-mode (dual-frame phone and web) statewide survey. Including a QR code (in addition to a URL) significantly increased the response rate compared to invitations that only included a URL in our study. As expected, a consequence of including the QR code was an elevated number of completions on smartphones or tablets among households randomly assigned to the QR code condition. The use of a larger (6 × 9 inch) envelope did not affect the overall response rate but did significantly boost the response rate for the landline sample (envelopes addressed to “STATE resident”) while having little effect for the wireless sample (envelopes addressed by name). This study suggests that incorporating both QR codes and larger (6 × 9 inch) envelopes in mail recruitment materials for web surveys is a cost-effective approach to increase web participation.

1340A Trial of Modern Sample Frames and Modes: The WA Health and Wellbeing Surveillance System

Abstract Background The Health and Wellbeing Surveillance System (HWSS) monitors the health status of the WA population. Its reliance on a landline sample frame and Computer Assisted Telephone Interview (CATI) must shift to mobile phone sample frames and novel online completion options, improving the representativeness of the HWSS data to the WA population. Methods Concurrent surveys in 2020 were run using three sample frames, the 2013 Electronic White Pages (EWP), the WA Electoral Roll (WAER) and a Sensis consumer database. The survey modes explored included CATI, online, and dual mode. Responses were weighted to the WA population. Response rates were compared between frames and modes. Differences in demographics, health risk factors, conditions and behaviours were investigated by comparing prevalence estimates and logistic regression modelling. Results Response rates were 20% for EWP CATI, 41% for Sensis CATI, 16% for Sensis online, and 10% for WAER online. Compared to EWP, the WAER and Sensis frame respondents were younger, had higher incomes and better education. Online respondents had higher prevalence of high psychological distress and lower prevalence of smoking compared with CATI. Conclusions The WAER and Sensis consumer database are valid sample frame options for the future of the HWSS. CATI gave the highest response rates, yet respondent differences by frame and mode were evident. WA Health is further exploring these options of modern sample frames and survey modes in 2021. Key messages Population health surveillance systems must modernise their sample frames and data collection modes to continue to provide reliable health prevalence estimates.

Rural, urban, and suburban differences in influenza vaccination coverage among children

Influenza vaccination is the primary way to prevent influenza, yet influenza vaccination coverage remains low in the United States. Previous studies have shown that children residing in rural areas have less access to healthcare and lower vaccination coverage for some vaccines. Influenza vaccination coverage among children 6 months–17 years by rural/urban residence during the 2011–12 through 2018–19 influenza seasons was examined using National Immunization Survey-Flu data. The Council of American Survey Research Organizations response rates for National Immunization Survey-Flu ranged from 48% to 65% (2011–12 through the 2017–18 seasons) for the landline sample and 20%–39% (2011–12 through the 2018–19 seasons) for the cellular telephone sample. Children residing in rural areas had influenza vaccination coverage that ranged from 7.9 (2012–13 season) to 12.6 (2016–17 season) percentage points lower than children residing in urban areas, and ranged from 4.5 (2012–13 season) to 7.4 (2016–17 season) percentage points lower than children residing in suburban areas. The differences in influenza vaccination coverage among rural, suburban, and urban children were consistent over the eight seasons studied. Lower influenza vaccination coverage was observed among rural children regardless of child’s age, mother’s education, household income, or number of children under 18 years of age in the household. Rural versus urban and suburban differences in influenza vaccination coverage remained statistically significant while adjusting for selected sociodemographic characteristics. A better understanding of the reasons for lower childhood influenza vaccination coverage for children in rural and suburban areas is needed.

Impact of multiple sample frames on data quality of household travel surveys: the case of the 2016 Transportation Tomorrow Survey

ABSTRACT The landline sample frame has been the method for recruiting participants for household travel surveys for many years. However, more recently, the representativeness of populations has declined with the reduction of household landline use and the rise of cellphone use. As a result, interest has turned to construct the survey sample frame using multiple lists. To attempt to alleviate underrepresentation, the 2016 iteration of the Transportation Tomorrow Survey (TTS) – the regional household travel survey of the Greater Golden Horseshoe located in southern Ontario, Canada – used multiple sample frames consisting of address/phone, address-only, and phone-only lists. To analyze the impact of a multi-frame approach, sample frames are compared based on demographic and trip variables to investigate whether there are significant differences between sample frames. Trip variable analysis via chi-squared and interaction effects modeling show small but significant differences, particularly between address-only and address/phone frames. The multiple sample frame is shown to improve representation compared to the singular landline frame; however, some broad representation issues remain with respect to age and household-size distributions, particularly in rural areas. This points to the need for alternative methods to recruit younger individuals and larger households.

Telephone Survey Calling Patterns, Productivity, Survey Responses, and Their Effect on Measuring Public Opinion

This study examines the effect of telephone survey dialing patterns on lab productivity and survey responses. Using an original data set of paradata from 2010 to 2017 and a machine learning technique for variable selection, we find that early and late afternoon shifts are as productive as late evening shifts for both landline and cellphone Random Digit Dialing (RDD) samples. Also, early weekdays are more productive than the weekend for the cellphone RDD samples. Most importantly, time of the interview affects survey responses; therefore, survey practitioners and scholars should be cognizant of this effect when scheduling calls.

Estimation tools for reducing the impact of sampling and nonresponse errors in dual-frame RDD telephone surveys.

We discuss alternative estimators of the population total given a dual-frame random-digit-dial (RDD) telephone survey in which samples are selected from landline and cell phone sampling frames. The estimators are subject to sampling and nonsampling errors. To reduce sampling variability when an optimum balance of landline and cell phone samples is not feasible, we develop an application of shrinkage estimation. We demonstrate the implications for survey weighting of a differential nonresponse mechanism by telephone status. We illustrate these ideas using data from the National Immunization Survey-Child, a large dual-frame RDD telephone survey sponsored by the Centers for Disease Control and Prevention and conducted to measure the vaccination status of American children aged 19 to 35months.

Geographic Inaccuracy of Cellphone Samples

The geographic accuracy provided in cellphone samples has been an issue of particular concern for telephone surveys. Surveyors use this information for geographic sample stratification and analysis; however, the location generated using random digit dialing (RDD) methods is typically the rate center in which the phone was activated and not the owner’s current residence. Researchers have found the location is fairly accurate at the regional (census region) level. However, the error rate increases when moving to the state and local levels. Cellphone error rates are higher among ‘cell-only’ (cellphone users who do not have a landline telephone) respondents than for ‘dual-users’ (cellphone users who have a cellphone and landline telephone). Among respondents sampled from the landline frame, the error rate (difference between location on the sample file and current location of residence) is smaller. Past research estimates the state error rate is about 1% for a landline sample and 12% for the ‘cell-only’ group (Skalland and Khare 2013). This paper builds upon previous research by providing state level estimates of inaccuracy for both ‘cell-only’ and ‘dual-users.’ It also examines the trend of state-level inaccuracy estimates over 18 months using data from the Gallup Daily Tracking survey. Finally, we examine the characteristics of ‘cell-only’ and ‘dual-users’ and implications of geographic inaccuracy of cellphone samples on the state level dual-frame RDD estimates.

An outbreak of hepatitis A in Canada: The use of a control bank to conduct a case-control study.

An outbreak of 18 cases of hepatitis A virus infection across five Canadian provinces was investigated. Case onsets occurred between October 2017 and May 2018. A retrospective matched case-control study was conducted to identify the likely source of the outbreak. Three matched controls were recruited for each case using a previously established control bank, supplemented by landline and cell phone call lists. Univariate and multivariate matched analyses were conducted to identify a potential outbreak source. Seventy-two per cent of controls were recruited through the control bank, and required on average 25.5 calls per recruited control; 20% of controls were recruited through a landline sample and 8% of controls were recruited through a cell phone sample, requiring an average of 847.3 and 331.7 calls per recruited control, respectively. Results of the analysis pointed to shrimp/prawns (odds ratio (OR) 15.75, p = 0.01) and blackberries (OR 7.21, p = 0.02) as foods of interest, however, an outbreak source could not be confirmed. The control bank proved to be a more efficient method for control recruitment than random call lists. Expanding the control bank size and using alternative methods, such as online surveys, may prove beneficial for increasing the timeliness of a case-control study during an outbreak investigation.

Comparison response patterns on landline and cell phone in a call back survey: effects of demographic characteristics and lag days.

The Asthma Call-back Survey (ACBS) is conducted after the Behavioral Risk Factor Surveillance System (BRFSS) survey by calling BRFSS respondents who reported ever being diagnosed with asthma. To find response patterns and increase ACBS response rates, we first examined whether obtaining consents during the BRFSS survey could increase call back response rates by reducing the refusal and break-off. Then, we assessed how the lag days between BRFSS and ACBS interviews affected response rates. BRFSS cell phone respondents agreed more often to being called back than did landline respondents (75.5 vs. 70.9 percent). However, when respondents were contacted for ACBS, the cell phone response rate was lower than landline response rate (43.4 vs. 47.0 percent), except among males aged 25–34 years, for which the cell phone response rate was 2.1 percent higher than the landline response rate. ACBS response rate for landline and cell phone response were highest if the callback was within 2 days of BRFSS interviews (92.3 vs. 88.8 percent). As lag days increased, the response rate decreased. The cell phone response rate showed a sharper drop; after 2 weeks, the response rate gap between landline and cell phone samples reached 17.3 percent.

Within-Household Selection and Dual-Frame Telephone Surveys: A Comparative Experiment of Eleven Different Selection Methods

Numerous within-household selection methods have been tested in general population surveys since the advent of telephone interviewing. However, very few selection studies, if any, have been conducted with a dual frame (landline and cell phone) sample. Landline and cell phone frames are known to represent demographically different groups of respondents, and selection methods that may result in more representative demographics in a landline frame may actually skew the results when combined with the cell phone frame. This study tested 11 different within-household selection methods with approximately 11,000 landline respondents. A parallel cell phone sample was also collected with 1,000 respondents, and the frames were combined for analysis. The selection methods tested included one probability-based method, four quasi-probability methods and six nonprobability methods. The methods were evaluated on four criteria: response rates, accuracy, demographic representation and substantive results. The demographic representativeness of each method was examined for the landline frame only and for the dual (landline and cell phone) frame combination. The probability method had the lowest response rate, while the nonprobability at-home methods had the highest. Accuracy rates were lowest for the quasi-probability birthday methods. There were few demographic differences between selection methods, and no substantive differences, when combined with the cell phone sample.

Examination of associations between early life victimisation and alcohol's harm from others.

Study aims were to examine: (i) how physical and sexual victimisation in early life are associated with alcohol's harm from others; and (ii) whether respondents' current drinking is a mediator of the association between early life victimisation and alcohol's harm from others among men and women. Data were from national computer-assisted telephone interviews, using the landline sample (3335 men and 3520 women ages ≥18) from the 2010 US National Alcohol Survey. Harms from someone else's drinking included family/marital problems, financial troubles, assault and vandalism in the past 12months. Victimisation was measured with severe physical abuse or sexual assault before age 18. Severe physical or sexual victimisation before age 18 was reported by 3.4% of men and 8.1% of women. Significantly more men (5.2%) than women (2.4%) reported assault by other drinkers, and significantly more women reported family/marital (5.3%) and financial problems (2.8%) than did men (2.6 and 1% respectively). Severe early life victimisation was robustly associated with a greater likelihood of experiencing past-year harms from other drinkers for both men and women. Men's drinking partially mediated associations between early life victimisation and recent assaults and vandalism by other drinkers. Early life victimisation may increase risk of harms from someone else's drinking. Health services and interventions that screen for histories of victimisation may help decrease risk of later harms from others' drinking. Reductions in drinking among men with histories of victimisation also could help reduce their exposure to such harms. [Kaplan LM, Greenfield TK, Karriker-Jaffe KJ. Examination of associations between early life victimisation and alcohol's harm from others.

The impact of the mode of survey administration on estimates of daily smoking for mobile phone only users

BackgroundOver the past decade, there have been substantial changes in landline and mobile phone ownership, with a substantial increase in the proportion of mobile-only households. Estimates of daily smoking rates for the mobile phone only (MPO) population have been found to be substantially higher than the rest of the population and telephone surveys that use a dual sampling frame (landline and mobile phones) are now considered best practice. Smoking is seen as an undesirable behaviour; measuring such behaviours using an interviewer may lead to lower estimates when using telephone based surveys compared to self-administered approaches. This study aims to assess whether higher daily smoking estimates observed for the mobile phone only population can be explained by administrative features of surveys, after accounting for differences in the phone ownership population groups.MethodsData on New South Wales (NSW) residents aged 18 years or older from the NSW Population Health Survey (PHS), a telephone survey, and the National Drug Strategy Household Survey (NDSHS), a self-administered survey, were combined, with weights adjusted to match the 2013 population. Design-adjusted prevalence estimates and odds ratios were calculated using survey analysis procedures available in SAS 9.4.ResultsBoth the PHS and NDSHS gave the same estimates for daily smoking (12%) and similar estimates for MPO users (20% and 18% respectively). Pooled data showed that daily smoking was 19% for MPO users, compared to 10% for dual phone owners, and 12% for landline phone only users. Prevalence estimates for MPO users across both surveys were consistently higher than other phone ownership groups. Differences in estimates for the MPO population compared to other phone ownership groups persisted even after adjustment for the mode of collection and demographic factors.ConclusionsDaily smoking rates were consistently higher for the mobile phone only population and this was not driven by the mode of survey collection. This supports the assertion that the use of a dual sampling frame addresses coverage issues that would otherwise be present in telephone surveys that only made use of a landline sampling frame.

Assessing the Use of a Pre-Field Screening Service to Identify Nonworking Cellphone Numbers in Ohio

In recent years, most RDD telephone surveys have introduced a cellphone frame to augment the landline sample to improve coverage and reduce variance. Due to a number of challenges it can be more expensive to complete cellphone interviews and often, for this reason, the proportion of the total sample allocated to the cell phone frame has been kept artificially small to reduce costs. As more people shift away from using landlines it becomes increasingly important to increase the allocation of RDD sample to the cellphone frame. We investigate the effectiveness of the Cell-Wins service provided by Marketing Systems Group which flags cellphone numbers that are actively being used by a person at the time the flag is applied. If accurate, this flag could reduce data collection costs by identifying nonworking numbers prior to data collection. This paper presents an assessment of this flag in the State of Ohio by which we (1) assess the under coverage rate if identified inactive telephone numbers were excluded from the sampling population and (2) determine the cost efficiency associated with excluding telephone numbers identified as inactive. We found that the accuracy of the Cell-Wins flag was high for telephone numbers identified as inactive and only introduced an under coverage rate of 2.4%. Furthermore, we determined that utilizing the Cell-Wins flag was cost efficient decreasing data collection costs a net 12%.

Drinking context and alcohol’s harm from others among men and women in the 2010 US National Alcohol Survey

ABSTRACTBackground: Little is known about how drinking in different contexts is associated with harms from someone else’s drinking, including marital problems, financial problems, and assault. We examined how drinking in four different contexts was associated with alcohol’s harm from others (AHFO). Methods: We utilized the landline sample of the 2010 US National Alcohol Survey (n = 5885) to examine associations between drinking context and AHFO using weighted binary logistic regression. Results: For women, drinking when friends dropped over was positively associated with assault and financial troubles due to someone else’s drinking. Drinking when friends dropped over was negatively associated with assault for men. For men, drinking at a bar, party, or during a quiet evening at home were each significantly associated with more assault by someone who had been drinking. Bar drinking among women was significantly associated with more marital problems, whereas drinking at a party at someone else’s home was associated with significantly less marital problems. Conclusions: Context-specific drinking has differential associations with specific types of harms from someone else’s drinking for men and women. Additional research on drinking context, relationship to the harmer, and violence experienced by men and women is needed.

A Next Step in the Evolution of Landline Sampling: Evaluating the Assignment-Based Frame

A Next Step in the Evolution of Landline Sampling: Evaluating the Assignment-Based Frame

Differences Between Landline and Mobile Phone Users in Sexual Behavior Research.

This study investigated differences between the demographic characteristics, participation rates (i.e., agreeing to respond to questions about sexual behavior), and sexual behaviors of landline and mobile phone samples in Australia. A nationally representative sample of Australians aged 18years and over was recruited via random digit dialing in December 2011 to collect data via computer-assisted telephone interviews. A total of 1012 people (370 men, 642 women) completed a landline interview and 1002 (524 men, 478 women) completed a mobile phone interview. Results revealed that telephone user status was significantly related to all demographic variables: gender, age, educational attainment, area of residence, country of birth, household composition, and current ongoing relationship status. In unadjusted analyses, telephone status was also associated with women's participation rates, participants' number of other-sex sexual partners in the previous year, and women's lifetime sexual experience. However, after controlling for significant demographic factors, telephone status was only independently related to women's participation rates. Post hoc analyses showed that significant, between-group differences for all other sexual behavior outcomes could be explained by demographic covariates. Results also suggested that telephone status may be associated with participation bias in research on sexual behavior. Taken together, these findings highlight the importance of sampling both landline and mobile phone users to improve the representativeness of sexual behavior data collected via telephone interviews.

Weighting Strategies for Combining Data from Dual-Frame Telephone Surveys: Emerging Evidence from Australia

Abstract Until quite recently, telephone surveys have typically relied on landline telephone numbers. However, with the increasing popularity and affordability of mobile phones, there has been a surge in households that do not have landline connections. Additionally, there has been a decline in the response rates and population coverage of landline telephone surveys, creating a challenge to collecting representative social data. Dual-frame telephone surveys that use both landline and mobile phone sampling frames can overcome the incompleteness of landline-only telephone sampling. However, surveying mobile phone users introduces new complexities in sampling, nonresponse measurement and statistical weighting. This article examines these issues and illustrates the consequences of failing to include mobile-phone-only users in telephone surveys using data from Australia. Results show that there are significant differences in estimates of populations’ characteristics when using information solely from the landline or mobile telephone sample. These biases in the population estimates are significantly reduced when data from the mobile and landline samples are combined and appropriate dual-frame survey estimators are used. The optimal choice of a dual-frame estimation strategy depends on the availability of good-quality information that can account for the differential patterns of nonresponse by frame.

Early Life Health, Trauma and Social Determinants of Lifetime Abstention from Alcohol.

Factors influencing lifetime abstention from alcohol may be relevant to the validity of analyses of alcohol's impact on health outcomes. We evaluate relationships between early life experiences, social factors, and demographic characteristics on lifetime abstainer status in models disaggregating by gender and, among women, race/ethnicity. Analyses use the landline sample (N=5382) of the 2010 U.S. National Alcohol Survey. Surveyed participants who reported never drinking alcohol were defined as lifetime abstainers. Additional variables assessed included demographics, dispositions to risk taking and impulsivity, and indicators of early life stress like economic difficulty, childhood trauma and early onset of health conditions. Logistic regression models predicting lifetime abstention were estimated. Lifetime abstainers are more likely to be women and, among women, to be non-White and Latina. Those reporting that their religion discouraged drinking and that religion was very important to them were more likely to be lifetime abstainers. Higher education levels were associated with reduced rates of lifetime abstention among women. Also among women, family problem drinking was associated with lower rates of lifetime abstention. However, childhood economic difficulty significantly predicted lower abstention only for White women, and childhood sexual abuse was significantly related to lower lifetime abstention only for Black women. Understanding the characteristics and determinants of individuals who never drink alcohol is relevant to any analysis of alcohol-related health outcomes. Results point to specific factors related to lifetime abstention with potential to bias such analyses if not included as control measures. Analyses evaluating relationships between early life experiences, social factors, and demographics with lifetime abstainer status identified characteristics associated with both poor health and with better health. These included lower risk taking and impulsivity scores and lower rates of family problem drinking, childhood economic difficulties and childhood sexual abuse.

Bias of health estimates obtained from chronic disease and risk factor surveillance systems using telephone population surveys in Australia: results from a representative face-to-face survey in Australia from 2010 to 2013.

BackgroundEmerging communication technologies have had an impact on population-based telephone surveys worldwide. Our objective was to examine the potential biases of health estimates in South Australia, a state of Australia, obtained via current landline telephone survey methodologies and to report on the impact of mobile-only household on household surveys.MethodsData from an annual multi-stage, systematic, clustered area, face-to-face population survey, Health Omnibus Survey (approximately 3000 interviews annually), included questions about telephone ownership to assess the population that were non-contactable by current telephone sampling methods (2006 to 2013). Univariable analyses (2010 to 2013) and trend analyses were conducted for sociodemographic and health indicator variables in relation to telephone status. Relative coverage biases (RCB) of two hypothetical telephone samples was undertaken by examining the prevalence estimates of health status and health risk behaviours (2010 to 2013): directory-listed numbers, consisting mainly of landline telephone numbers and a small proportion of mobile telephone numbers; and a random digit dialling (RDD) sample of landline telephone numbers which excludes mobile-only households.ResultsTelephone (landline and mobile) coverage in South Australia is very high (97 %). Mobile telephone ownership increased slightly (7.4 %), rising from 89.7 % in 2006 to 96.3 % in 2013; mobile-only households increased by 431 % over the eight year period from 5.2 % in 2006 to 27.6 % in 2013. Only half of the households have either a mobile or landline number listed in the telephone directory. There were small differences in the prevalence estimates for current asthma, arthritis, diabetes and obesity between the hypothetical telephone samples and the overall sample. However, prevalence estimate for diabetes was slightly underestimated (RCB value of −0.077) in 2013. Mixed RCB results were found for having a mental health condition for both telephone samples. Current smoking prevalence was lower for both hypothetical telephone samples in absolute differences and RCB values: −0.136 to −0.191 for RDD landline samples and −0.129 to −0.313 for directory-listed samples.ConclusionThese findings suggest landline-based sampling frames used in Australia, when appropriately weighted, produce reliable representative estimates for some health indicators but not for all. Researchers need to be aware of their limitations and potential biased estimates.Electronic supplementary materialThe online version of this article (doi:10.1186/s12874-016-0145-z) contains supplementary material, which is available to authorized users.

Is the only poll that counts the one on Election Day?

Is the only poll that counts the one on Election Day?