Keystroke analysis and implications for fieldwork WP3.3 – task 2 Johanna Bristle (SHARE, MEA) Verena Halbherr (ESS, GESIS) DASISH Final Conference – Gothenburg, 28 th Nov 2014
Introduction 2 Tools produce paradata, e.g. Fieldwork Monitoring Application produces contact data (task 1 of WP3.3) Goal of task 2: Analysing one type of paradata, namely SHARE keystroke data and ESS time stamp data Inform cross-national fieldwork & derive lessons learned Definition: Paradata = data about the process of data collection Keystroke data = actions on a keyboard and time stamps for each action taken are recorded and extracted
Deliverable D3.7 Structure of deliverable Description of data and preparation Diagnostic of keystrokes Outliers and distribution Measuring interview length Using keystrokes in a survey`s life-cycle Questionnaire development Fieldwork monitoring Post-survey quality assessment Available on the website of DASISH beginning of 2015 Today: comparing the two surveys 3 Focus of poster Focus of this presentation
Interview length Concept of interview length Perspective of questionnaire development vs. interviewer and respondent 4
Measurement error in time stamps Paradata is far from error-free Technical errors or breaks produce outliers or missing data Reporting error if interviewer is involved (rounding) 5
Cross-national interview length 6 Notes: SHARE: own analysis on panel single households in wave 5 ESS: Analysis by Loosveldt & Beullen round 5, all respondents.
Conclusions 1.Computerized data collection avoids interviewer-induced measurement error Data quality of paradata data quality of survey data 2.Cross-national & cross-survey comparison Beyond survey-specific reasons (topic) „Local survey culture“ (language) Potential of paradata Pretest: check quality of new items During fieldwork: real-time monitoring and interventions Post-survey analysis: survey methodology & enhancing survey items Harmonization in terms of documentation, dissemination and tool development produces useful para- and metadata 7
Further reading and references DASISH Deliverable 3.7 (forthcoming). Keystroke analysis and implications for field work. It will be available on Loosveldt, G. and K. Beullens (2013). "'How long will it take?' An analysis of interview length in the fifth round of the European Social Survey." Survey Research Methods 7(2): Kreuter, F. (2013). Improving Surveys with Paradata: Analytic Uses of Process Information. Hoboken, New Jersey, John Wiley & Sons. 8