Computers play an important role in everyday multitasking. Within this context, we focus on respondent multitasking (RM) in web surveys. RM occurs when users engage in other activities while responding to a web survey questionnaire. The conceptual framework is built on existing literature on multitasking, integrating knowledge from both cognitive psychology and survey methodology. Our main contribution is a new approach for measuring RM in web surveys, which involves an innovative use of the different types of paradata defined as non-reactive electronic tracks concerning respondents' process of answering the web questionnaire. In addition to using questionnaire page completion time as a measure of RM, we introduce "focus-out" events that indicate when respondents have left the window containing the web questionnaire (e.g., to chat, email, browse) and then returned. The approach was tested in an empirical study using a web survey on a student sample (n=267). The results indicate that 60% of respondents have multitasked at least once. In addition, they reveal that item nonresponse as an indicator of response quality is associated with RM, while non-differentiation is not. Although this study confirms that a paradata-based approach is a feasible means of measuring RM, future research on this topic is warranted.
COBISS.SI-ID: 33672285
In the research of online communities and web survey methodology little is known about how elements in email invitations to list-based web surveys can be used to obtain higher response rates. In the present work, we investigated whether making authority, plea for help, and sense of community salient in email invitations determines the response of survey participants. Drawing from both survey methodology and recent research on online communities, this study also tested a hypothesis on the relationship between activity in an online community and survey response. Using a full-factorial experiment based on a simple random sample of 2500 members from the largest online health community in Slovenia, the results support only the hypothesis that plea for help is an effective response inducing element in email invitations. Furthermore, the results support the hypotheses that online community activity, related to the frequency of visits and number of posts to an online community, are positively associated with response in list-based web survey. Since this study also shows that combining more than one element in email invitations does not necessary improve response rates, web survey research and practice may benefit from future research on this topic.
COBISS.SI-ID: 33758045
Although the number of older adults with access to landline and mobile phones has grown in recent years, little research exists that investigates how older adults use and potentially combine the two technologies for communication with the members of their social support networks. Therefore, this study explored how the composition of older adults' social companionship networks is related to the frequency of landline and mobile phone communication with their network members. In addition, the study examined the association between landline and mobile phone communication as well as their relationship with in-person communication in such networks. The survey data were gathered using an ego-centered social support approach in a nationwide representative sample of retired older adults in Slovenia. The results of multiple regression analyses revealed that the composition of a social companionship network better predicts the frequency of landline rather than mobile phone communication of older adults with their network members. The non-significant relationship between the frequency of landline and mobile phone communication suggested that the two technologies very likely have a different role in the maintenance of social companionship networks. The study also found a difference between the importance of control variables related to the socio-demographic characteristics of older adults for the frequency of landline and mobile phone communication. While age, education, and living alone significantly predicted the frequency of mobile communication with older adults' network members, significant predictors of their landline communication included gender and health impairments.
COBISS.SI-ID: 33921629
The chapter of an edited book on survey methodology published by one of the leading publisher of scientific literature provides a comprehensive overview of differences between survey modes and their implications. It summarises advantages and disadvantages of most commonly used survey modes and presents various approaches to combining survey modes at different stages of a survey project. The key focus of the chapter is on elaboration of potential impacts of different modes and their combinations on data quality. For this purpose, it presents a novel typology of distinguishing survey modes by their inherent characteristics while also considering specific survey implementation approaches and contextual factors of data collection. On this basis, it critically evaluates the impacts of different survey modes on measurement quality and representativeness of survey data and provides guidelines for effective use of mixed-mode survey strategies.
COBISS.SI-ID: 34229853
The sampling theory was basically developed for probability sampling, where all units in the population have known and positive probabilities of inclusion. This definition implicitly involves randomization, where the units are selected according to their inclusion probabilities. In probability sampling the randomized selection is used instead of arbitrary or purposive sample selection of the researcher, or, instead of various self-selection processes run by respondents. Within this context, the notion of non-probability sampling denotes the absence of probability sampling mechanism. In this chapter we first reflect on the practice of non-probability samples. Second, we introduce probability sampling principles and observe their approximate usage in the non-probability setting and we also discuss some other strategies. Third, we provide a closer look at two contemporary – and perhaps also the most exposed – aspects of non-probability sampling: online panels and weighting. Finally, we summarize recommendations for deciding on probability–nonprobability sampling dilemmas and provide concluding remarks.
COBISS.SI-ID: 34230109