Incomplete data is a common problem of survey research. Recent work on multiple imputation techniques has increased analysts' awareness of the biasing effects of missing data and has also provided a convenient solution. Imputation methods replace non-response with estimates of the unobserved scores. In many instances, however, non-response to a stimulus does not result from measurement problems that ...
In:
Quality & Quantity
40 (2006), 2, S. 225-244
| Martin Kroh
As a dependent variable, party choice did not lend itself to analysis by means of powerful multivariate methods until the coming of discrete-choice models, most notably conditional logit and multinomial logit. These methods involve estimating effects on party preferences (utilities) that are post hoc derived from the data, but such estimates are plagued by a number of difficulties. These difficulties ...
In:
Electoral Studies
25 (2006), 3, S. 424-447
| Cees van der Eijk, Wouter van der Brug, Martin Kroh, Mark Franklin
Although left-right items are a standard tool of public opinion research, there remains some difference of opinion on the optimal response format. Two disputes can be identified in the literature: (a) whether to provide respondents with a small or large number of answer categories and (b) whether or not to administer the response scale including a midpoint. This study evaluates the performance of the ...
Using an experimental design, we compare two alternative approaches to dependent interviewing (proactive and reactive) with traditional independent interviewing on a module of questions about sources of income. We believe this to be the first large-scale quantitative comparison of proactive and reactive dependent interviewing. The three approaches to questioning are compared in terms of their impact ...
2005| Peter Lynn, Annette Jäckle, Stephen P. Jenkins, Emanuela Sala
This report derives from the project "Improving survey measurement of income and employment (ISMIE)" which investigates measurement error in survey data on income and employment, using a UK sub-sample of the European Household Community Panel (ECHP). In this paper we describe the process of collecting validation data and the outcomes of the process. Validation data were obtained from two sources: employers' ...
2005| Annette Jäckle, Emanuela Sala, Stephen P. Jenkins, Peter Lynn
Linkages of household survey responses with administrative data may be based on unique individual identifiers or on survey respondent characteristics. The benefits gained from using unique identifiers need to be assessed in the light of potential problems such as non-response and measurement error. We report on a study that linked survey responses to UK government agency records on benefits and tax ...
2005| Stephen P. Jenkins, Peter Lynn, Annette Jäckle, Emanuela Sala
We analyse consent patterns and consent bias in the context of a large general household survey, the 'Improving survey measurement of income and employment' (ISMIE) survey, also addressing issues that arise when there are multiple consent questions. Using a multivariate probit regression model for four binary outcomes with two incidental truncations, we show that there are biases in consent to data ...
2005| Stephen P. Jenkins, Lorenzo Cappellari, Peter Lynn, Annette Jäckle, Emanuela Sala