Notice: the WebSM website has not been updated since the beginning of 2018.

Web Survey Bibliography

Title The use of within-subject experiments for estimating measurement effects in mixed-mode surveys
Year 2014
Access date 13.04.2014
Full text

PDF (716 kb)

Abstract

The estimation of measurement effects (MEs) of survey modes in the presence of selection bias poses a great problem to methodologists. In practice, unit nonresponse is often assumed to be ignorable (MAR) conditional on socio-demographic auxiliary data for lack of other sample-level information. Since socio-demographics commonly are only weakly related to target variables or selection mechanisms this assumption often may not hold true.
We present a new method to estimate MEs by means of “within-subject designs”, in which the same sample is approached by two different modes at two subsequent points in time. This design allows ignoring nonresponse on repeatedly measured target variables implying weaker MAR assumptions, because repeated target variables typically are strongly related. Further assumptions of this design are discussed in detail, in particular (a) time-stability of target variables and response probabilities and (b) the independence of measurement occasions. In extensions of simple within-subject designs, an independent control group, to which the same mode is assigned at both occasions, is useful to test these assumptions and adjust for time-instability.

Access/Direct link

(Abstract & Full text)

Year of publication2014
Bibliographic typeConferences, workshops, tutorials, presentations
Reports, seminars
Print