Austrian Corona Panel Data: Method Report

Last updated: 18 September 2020

1          Introduction

The corona crisis has fundamentally changed everyday life in Austria as well as in many other countries. However, people are affected in very different ways. Against this background, the Austrian Corona Panel Project aims to provide an overview of various health, economic and social aspects of the Corona crisis. This report provides details on the methods and process of data collection of the Austrian Corona Panel Data gathered by the Austrian Corona Panel Project at the University of Vienna.

1.1       How to cite the data

The data from Austrian Corona Panel Project should be cited as follows:

  • Scientific Use File (SUF edition)
    Kittel, Bernhard; Kritzinger, Sylvia; Boomgaarden, Hajo; Prainsack, Barbara; Eberl, Jakob-Moritz; Kalleitner, Fabian; Lebernegg, Noëlle S.; Partheymüller, Julia; Plescia, Carolina; Schiestl, David W.; Schlogl, Lukas, 2020, "Austrian Corona Panel Project (SUF edition)", https://doi.org/10.11587/28KQNS, AUSSDA.
  • Open Access File (OA edition)
    Kittel, Bernhard; Kritzinger, Sylvia; Boomgaarden, Hajo; Prainsack, Barbara; Eberl, Jakob-Moritz; Kalleitner, Fabian; Lebernegg, Noëlle S.; Partheymüller, Julia; Plescia, Carolina; Schiestl, David W.; Schlogl, Lukas, 2020, "Austrian Corona Panel Project (OA edition)", https://doi.org/10.11587/P5YJ0O, AUSSDA.

Please also cite the data paper along with the data:

  • Kittel, Bernhard; Kritzinger, Sylvia; Boomgaarden, Hajo; Prainsack, Barbara; Eberl, Jakob-Moritz; Kalleitner, Fabian; Lebernegg, Noëlle S.; Partheymüller, Julia; Plescia, Carolina; Schiestl, David W.; Schlogl, Lukas, 2020, “The Austrian Corona Panel Project: Monitoring Individual and Societal Dynamics amidst the COVID-19 Crisis”, Working Paper. Available at SSRN: https://ssrn.com/abstract=3654139

2          Study Description

2.1       Title and version

Austrian Corona Panel Project (V20200918)

2.2       Principal Investigators

  •  Bernhard Kittel, Department of Economic Sociology, University of Vienna
  •  Sylvia Kritzinger, Department of Government, University of Vienna
  •  Hajo Boomgaarden, Department of Communication, University of Vienna
  •  Barbara Prainsack, Department of Political Science, University of Vienna

2.3       Funding / Acknowledgments

Data collection has been made possible by COVID-19 Rapid Response Grant EI-COV20-006 of the Wiener Wissenschafts- und Technologiefonds (WWTF) and financial support by the rectorate of the University of Vienna. Further funding by the Austrian Social Survey (SSÖ), the Vienna Chamber of Labour (Arbeiterkammer Wien) and the Federation of Austrian Industries (Industriellenvereinigung) is gratefully acknowledged.

2.4       Fieldwork agency

The fieldwork for the Austrian Corona Panel Project was conducted by “Marketagent.com online research GmbH” (Baden, Austria).

2.5       Topics / Keywords

  • Crisis perception: health, economy, crisis management;
  • Effects of the crisis: health, factual behaviour, day structure, psychological effects, emotional status, care, partnership and family;
  • Expectations of the crisis: social and individual norms, corporate feeling, societal mood;
  • Politics: political trust, democracy, relationship government - parliament, government performance, surveillance measures, privacy, party choice, left/right;
  • Work: job situation, change in job situation, income, state support;
  • Economy: perceptions inflation, unemployment, purchase behaviour, personal and national economic situation (prospective and retrospective);
  • Communication: media use (traditional and social media), interpersonal relationships, fake news, perception of reporting of crisis;
  • Psychological predispositions: risk behaviour, life satisfaction;
  • Sociodemographic information: age, gender, education, region, occupation, religion, household structure, migration background.

3          Study Design

3.1       Sampling

The Austrian Corona Panel Project surveyed respondents with access to the internet (via computer or mobile devices such as smartphones or tablets). Respondents were sampled from a pre-existing online access panel provided by Marketagent, Austria. Respondents were selected (quota sample) based on the following key demographics: age, gender, region (Bundesland), municipality size, and educational level based on official statistics, with the quota sample being structured to closely mirror the Austrian resident population. In order to participate in the study, the respondents had to be residents in Austria and at least 14 years old. Please note that the usual caution must be applied in interpreting these data as specific segments of the population continue to be hard to reach by online surveys.

3.2       Fieldwork

The data collection started on 27 March 2020 and is currently ongoing:

  • Wave 1:          27 March 2020 – 30 March 2020
  • Wave 2:          3 April 2020 – 8 April 2020
  • Wave 3:          10 April 2020 – 16 April 2020
  • Wave 4:          17 April 2020 – 21 April 2020
  • Wave 5:          24 April 2020 – 29 April 2020
  • Wave 6:          1 May 2020 – 6 May 2020
  • Wave 7:          8 May 2020 – 13 May 2020
  • Wave 8:          15 May 2020 – 20 May 2020
  • Wave 9:          23 May 2020 – 27 May 2020
  • Wave 10:        29 May 2020 – 3 June 2020
  • Wave 11:        12 June 2020 – 17 June 2020
  • Wave 12:        26 June 2020 – 1 July 2020
  • Wave 13:        10 July 2020 – 15 July 2020
  • Wave 14:        14 August 2020 – 19 August 2020
  • Wave 15:        11 September 2020 – 18 September 2020

The median interview duration was:

  • Wave 1:          16.4 min
  • Wave 2:          20.9 min
  • Wave 3:          16.7 min
  • Wave 4:          19.7 min
  • Wave 5:          22.6 min
  • Wave 6:          18.7 min
  • Wave 7:          16.9 min
  • Wave 8:          18.7 min
  • Wave 9:          21.4 min
  • Wave 10:        21.3 min
  • Wave 11:        17.3 min
  • Wave 12:        19.7 min
  • Wave 13:        17.9 min
  • Wave 14:        26.1 min
  • Wave 15:        16.8 min

 3.3      Panel participation

Respondents were asked and incentivized with 180 credit points to participate in each wave of the panel. The invitation for necessary fresh panelists (if the retention was not sufficiently large, n < 1.500) was sent out a few days after field-work started. All respondents who participated in the previous wave were also invited to participate in subsequent waves.

Net sample sizes were as follows:

  • Wave 1:          1541
  • Wave 2:          1559
  • Wave 3:          1500
  • Wave 4:          1528
  • Wave 5:          1515
  • Wave 6:          1551
  • Wave 7:          1517
  • Wave 8:          1501
  • Wave 9:          1502
  • Wave 10:        1504
  • Wave 11:        1510
  • Wave 12:        1522
  • Wave 13:        1532
  • Wave 14:        1540
  • Wave 15:        1581

A total of 494 respondents completed all waves (Wave 1 to 15).

Initial participation rate:

  • Wave 1:          35.2% (1541 interviews, 4381 invitations)

Retention rates for panelists were as follows:

  • Wave 2:          86.2% (1328 interviews, 1541 W1 invitations)
  • Wave 3:          81.3% (1437 interviews, 1768 W1-2 invitations)
  • Wave 4:          78.1% (1430 interviews, 1831 W1-3 invitations)
  • Wave 5:          73.8% (1425 interviews, 1931 W1-4 invitations)
  • Wave 6:          74.3% (1499 interviews, 2018 W1-5 invitations)
  • Wave 7:          70.8% (1464 interviews, 2067 W1-6 invitations)
  • Wave 8:          69.1% (1460 interviews, 2114 W1-7 invitations)
  • Wave 9:          67.6% (1456 interviews, 2153 W1-8 invitations)
  • Wave 10:        65.5% (1447 interviews, 2208 W1-9 invitations)
  • Wave 11:        65.7% (1484 interviews, 2258 W1-10 invitations)
  • Wave 12:        61.3% (1399 interviews, 2283 W1-11 invitations)
  • Wave 13:        62.5% (1505 interviews, 2409 W1-12 invitations)
  • Wave 14:        55.8% (1355 interviews, 2429 W1-13 invitations)
  • Wave 15:        59.0% (1573 interviews, 2664 W1-14 invitations)

Participation rates of replacements for drop-outs:

  • Wave 2:          5.2% (231 interviews, 4401 invitations)
  • Wave 3:          4.3% (63 interviews, 1469 invitations)
  • Wave 4:          5.6% (98 interviews, 1760 invitations)
  • Wave 5:          5.4% (90 interviews, 1670 invitations)
  • Wave 6:          8.1% (52 interviews, 640 invitations)
  • Wave 7:          4.3% (53 interviews, 1220 invitations)
  • Wave 8:          5.9% (41 interviews, 690 invitations)
  • Wave 9:          3.1% (46 interviews, 1502 invitations)
  • Wave 10:        3.2% (57 interviews, 1800 invitations)
  • Wave 11:        3.0% (26 interviews, 870 invitations)
  • Wave 12:        4.5% (123 interviews, 2720 invitations)
  • Wave 13:        2.4% (27 interviews, 1130 invitations)
  • Wave 14:        6.5% (185 interviews, 2830 invitations)
  • Wave 15:        4.0% (8 interviews, 200 invitations)

4          Data Cleaning

4.1       Data protection (all editions)

Timestamps were truncated to the date of interview and respondent identifiers were replaced by an anonymized random number for reasons of data protection.

4.2       Data protection: Scientific Use File (SUF)

Some variables were recoded or dropped from the data file for reasons of data protection (see variable list for documentation). 

4.3       Data protection: Open Access File (OA)

Some variables were recoded or dropped from the data file for reasons of data protection (see variable list for documentation).

4.4       Variable format and missing values

For most variables missing values were coded as follows:

  • 88 = don’t know
  • 99 = no answer

5          Weighting of data

The dataset includes two types of weights which can be used for post-stratification adjustment to known population distributions (for each wave separately):

  • Demographic weight (W*_WEIGHTD)
  • Demographic + Political weight (W*_WEIGHTP)

Two additional weights are included to weight respondents who participated in all ten waves according to the known target distributions (W1W10_WEIGHTD, W1W10_WEIGHTP).

Two additional weights are included to weight respondents who participated in all waves according to the known target distributions (ALLWAVES_WEIGHTD, ALLWAVES_WEIGHTP).

Target distributions are based on Micro Census data (Statistics Austria StatCube 2019) and official election results are provided by the Austrian Ministry of the Interior. The demographic weight adjusts the following demographics to the target distributions: gender, age, gender X age, education, region (Bundesland), employment status, household size, and migration background. The demographic + political weight, in addition, uses the vote recall to match the election result of the 2019 national elections. For the computation of weights, we used the STATA module “ipfweight”. Weighting variables were trimmed to a minimum value of 0.20 and a maximum value of 5.00. Missing values on the variables used were weighted neutrally.