Mixed-Mode Surveys

About

Due to decreasing response rates around the world, survey designers must explore new ways of recruiting respondents for their surveys. One popular option is to offer different modes for survey participation, thus allowing target persons to participate in the survey mode that suits them best. While mixing survey modes for data collection can have positive effects on response rates, sample balance, and survey costs, the question arises of whether the data from different survey modes can be easily pooled and compared. This question also applies to data from long-standing survey programs that, in the course of the COVID 19-pandemic, had to refrain from face-to-face surveys and to collect data in other (mostly self-administered) survey modes.

In this short course, we provide an overview of empirical evidence related to the benefits and drawbacks of using multiple modes for data collection and outline some recommendations for the implementation of mixed-mode surveys. Specifically, we will cover topics such as mixed-mode-specific questionnaire design considerations, experimental evidence on comparing mixed-mode surveys with single-mode surveys, and optimal strategies of how to implement them.

The practical part of the course provides an introduction of statistical methods for testing mode effects and measurement equivalence of multi-item scales for mixed-mode and mixed-device survey data. We employ confirmatory factor analysis with multiple groups representing different modes and use regression and structural equation models to test mode effects. Furthermore, we demonstrate its use with prepared examples using the programs R-lavaan, Mplus Automation, and Mplus.

Course leader

Sven Stadtmüller, HAWK Hildesheim/Holzminden/Göttingen, Germany Henning Silber, GESIS – Leibniz Institute for the Social Sciences, Germany Peter Schmidt, University of Giessen & University of Mainz, Germany Yannick Diehl, University of Marburg

Target group

You will find the course useful if:
- you plan or conduct your own mixed-mode survey,
- you plan or use data from mixed-mode surveys,
- you eventually plan to test mode effects with you own datasets,
- you aim for a better understanding of the peculiarities of mixed-mode surveys.

Course aim

By the end of the course, you will:
- have a general understanding of advantages and disadvantages of survey modes and of mixing them within a data collection,
- have a general understanding of the peculiarities of mixed-mode surveys regarding nonresponse and measurement error as well as questionnaire design,
- have a general understanding of procedures for testing measurement equivalence of survey data collected via different survey modes/devices and its implementation in R-lavaan and Mplus.

Prerequisites:
- Basic knowledge in quantitative social sciences; practical experience in conducting surveys will be beneficial
- Basic knowledge of multivariate statistics, esp. confirmatory factor analysis
- Experiences with R/R Studio and/or MPlus

Fee info