Dataset details
Ref dataset: 828 (1.1.0)
Dataset title
Selects 2015 Panel / Rolling cross-section study
DOI 10.23662/FORS-DS-828-3
Previous versions 828-1:     No DOI is available for this version
828-2:     No DOI is available for this version
Ref dataset 828
Ref study 12447
Dataset language English
Additional information Changes to v1.1: - Added weighting variables for turnout and vote choice: weight_t, weight_p, weight_tot - Added combined turnout variable: W2_3_voted
Dataset version 1.1.0
Version notes
Publication date 28.06.2017
Nesstar link
Bibliographical citation Selects: Panel / Rolling cross-section study - 2015 [Dataset]. Distributed by FORS, Lausanne, 2016.
Data type Quantitative data
Consent of respondents
Time method Longitudinal
  • Digitalized data file
Data collector
Available document types
  • Collection instruments
Remarks about the documentation
Restriction Academic research and teaching
Special order
Special permission None
Status Published

Selects 2015 Panel / Rolling cross-section study data v1.1
Collection date - begin 13.06.2015
Collection date - end 11.01.2016
Analysis unit
  • Individual
Universe Swiss nationals aged 18 years or older
Sampling description Simple random sample of 29'500 persons from the whole of Switzerland with Swiss nationality aged 18 or older on the day of the elections. The sample is based on a individual register sampling frame maintained by the Swiss Federal Statistical Office (FSO).
Mode of data collection
  • Self-administered questionnaire: Web-based
Data collection instruments
  • Questionnaire
Data source Information on canton where the person has right to vote, marital status, country of birth, and household size come from the sampling frame delivered by the Swiss Federal Statistical Office (FSO). Municipalities are matched with information from the FSO such as the type and size of the municipality.
Weighting information Three weighting variables are provided: - weight_t: Turnout weight to correct for the overrepresentation of voters in the sample. This is meant to be used in conjunction with the combined W2&W3 turnout variable "W2_3_voted" / - weight_p: Party choice weight to correct the distribution of votes according to the official results. This is meant to be used in conjunction with the combined W2&W3 vote choice variable "W2_3_f11800main7" / - weight_tot: Combined turnout and party choice weight
Response rate Response rates per wave: First wave: 37% Second wave: 74% Third wave: 69% Fourth wave: 75% For the first wave, respondents were contacted by letter for a web interview. The first contact was a pre-notification letter with information on the survey and a flyer with past results. The second letter included the URL to participate in the survey, a personal login code, as well as a postal cheque of 10 Swiss francs that respondents could cash at any post office. Two reminders were also sent. At the end of the survey, respondents were asked to provide an email address for contact in subsequent waves. A raffle of 5 iPads, conditional on participating in the three first waves, was organized to boost enrollment in the panel as well as to reduce attrition. The second wave was conducted as a rolling cross-section survey (RCS) that started on August 17 and lasted for 60 days. The aim was to measure the evolution of opinions and vote intentions during the last phase of the campaign. Around 120 respondents participated each day. The invitations were sent by email with two reminders spaced close apart because to reduce the field time of each address. Pre-notification letters were sent weekly to inform about the email invitation to boost response rates. The third wave was conducted immediately after the federal elections. Those who had not provided an email address in the first wave were contacted by letter, the others by email. Additionally, a pre-notification letter was again sent for those invited by email. Finally, a fourth wave was conducted among all those who had participated in the third wave and for whom an email was available. This wave was timed to be just after the election of the Federal Council.
Data completeness 8 speeders were removed from W1 if the respondent took less than one third of the median duration to complete the questionnaire. 35 incomplete responses were also removed (if more than at least 15 variables out of 31 key non-filtered questions were not answered to).
Number of cases 11073
Number of variables 474