Psychologisches Institut

This article summarizes expertise gleaned from the first years of Internet-based experimental research and presents recommendations on: (1) ideal circumstances for conducting a study on the Internet; (2)what precautions have to be undertaken in Web experimental design; (3) which techniques have proven useful in Web experimenting; (4) which frequent errors and misconceptions need to be avoided; and(5) what should be reported. Procedures and solutions for typical challenges in Web experimenting are discussed. Topics covered include randomization, recruitment of samples, generalizability, dropout, experimental control, identity checks, multiple submissions, configuration errors, control of motivational confounding, and pre-testing. Several techniques are explained, including “warm-up,” “high hurdle,” password methods, “multiple site entry, ”randomization, and the use of incentives. The article concludes by proposing sixteen standards for Internet-based experimenting.

The present article describes VAS Generator (www.vasgenerator.net), a free Web service for creating a wide
range of visual analogue scales that can be used as measurement devices in Web surveys and Web experimentation, as well as for local computerized assessment. A step-by-step example for creating and implementing a visual analogue scale with visual feedback is given. VAS Generator and the scales it generates work independently of platforms and use the underlying languages HTML and JavaScript. Results from a validation study with 355 participants are reported and show that the scales generated with VAS Generator approximate an interval-scale level. In light of previous research on visual analogue versus categorical (e.g., radio button) scales in Internet-based research, we conclude that categorical scales only reach ordinal-scale level, and thus visual analogue scales are to be preferred whenever possible.

In three experiments we investigated whether two procedures of acquiring knowledge about the same causal structure, predictive learning (from causes to effects) versus diagnostic learning (from effects to causes), would lead to different base rate use in diagnostic judgments. Results showed that learners are capable of incorporating base rate information in their judgments regardless of the direction in which the causal structure is learned. However, this only holds true for relatively simple scenarios. When complexity was increased, base rates were only used after diagnostic learning, but were largely neglected after predictive learning. It could be shown that this asymmetry is not due to a failure of encoding base rates in predictive learning because participants in all conditions were fairly good at reporting them. The findings present challenges for all theories of causal learning.

The Internet experiment is now a well-established and widely used method. The present paper describes guidelines for the proper conduct of Internet experiments, e.g. handling of dropout, unobtrusive naming of materials, and pre-testing. Several methods are presented that further increase the quality of Internet experiments and help to avoid frequent errors. These methods include the "seriousness check", "warm-up," "high hurdle," and "multiple site entry" techniques, control of multiple submissions, and control of motivational confounding. Finally, metadata from sites like WEXTOR (http://wextor.org) and the web experiment list (http://wexlist.net/) are reported that show the current state of Internet-based research in terms of the distribution of fields, topics, and research designs used.

While in an earlier commentary (Honing & Ladinig, 2008) we stressed the potential of Web-delivered experiments for music perception research, the ongoing discussion on Web-based versus lab-based studies seems to circle around issues of method and control (Mehler, 1999; Kendall, 2008). We agree with the importance of these issues from a methodological point of view. However, we continue to stress that these issues are not essentially different for Web-based as compared to lab-based studies.

People are increasingly required to disclose personal information to computer- and Internetbased systems in order to register, dentify themselves or simply for the system to work as designed. In the present paper, we outline two different methods to easily measure people’s behavioral self-disclosure to web-based forms. The first, the use of an ‘I prefer not to say’ option to sensitive questions is shown to be responsive to the manipulation of level of privacy concern by increasing the salience of privacy issues, and to experimental manipulations of privacy. The second, blurring or increased ambiguity was used primarily by males in response to an income question in a high privacy condition. Implications for the study of self-disclosure in human–computer interaction and web-based research are discussed.

Die fortgeschrittene, weiterhin fortschreitende und in immer mehr Bereichen alltäglich werdende Nutzung des Internets im Allgemeinen und des Word Wide Webs im Speziellen macht das Netz gleichermaßen für Wissenschaftler, Marktforscher und Demoskopen zu einem interessanten Medium. Für computergestützte selbstadministrierte Befragungen (CAWI, computer assisted web interviewing) und andere Datenerhebungsformen im Web steht eine Vielzahl von Möglichkeiten für den Einsatz qualitativer und quantitativer Verfahren zur Verfügung.

Ein Grund für diese Popularität liegt sicherlich darin, dass der selbstadministrierte Befragungsmodus, bei dem der Untersuchungsteilnehmer die Dateneingabe mit dem Beantworten selbst übernimmt, schnelle Untersuchungen großer Stichproben verspricht. Immer mehr (frei verfügbare) Software und Onlinetools für das Erstellen von Befragungen im Web und deren Auswertung, ermöglichen es auch technisch nicht versierten Personen, Daten im Netz zu erheben. Dabei wird scheinbar oftmals vergessen, dass sich Daten zwar leicht generieren lassen, sie jedoch über bestimmte Qualitäten verfügen müssen, sollen aus ihnen valide Schlüsse gezogen werden. Sowohl Qualität, als auch Quantität von Daten werden durch die gewählten Messmethoden beeinflusst.

Die beiden Haupteinflüsse auf die Qualität der Ergebnisse einer Untersuchung, Repräsentation und Messung (Überblick in Groves/Fowler/Couper 2004), haben für Datenerhebungen im Web ebenso ihre Gültigkeit, wie für andere Erhebungsarten. Während der Bereich der Repräsentation, also der Weg von der angestrebten Grundgesamtheit über die realisierte Stichprobe und weitere Einflüsse wie Antwortverweigerung, sich nicht grundsätzlich von anderen Arten der Datenerhebung unterscheidet, gibt es, durch das Medium Computer bedingt, einige Besonderheiten im Bereich der Messung. Während Konstruktvalidität und spätere Datenverarbeitung noch unabhängig von der Erhebungsart sind, gibt es spezifische Einflüsse auf den Fehler durch den Messvorgang, die einen Einfluss auf die Ausprägung des Wertes haben, der gemessen werden soll. Die hier vorgestellten Messinstrumente und Skalen werden vor allem hinsichtlich ihrer Wirkung auf den Messfehler dargestellt.

Die internetbasierte Datenerhebung ist zwar in der Lage, einige Schwierigkeiten papierbasierter Untersuchungen zu beheben (z. B. manuelle Dateneingabe; Bewertung von Stimuli, die sich nicht abdrucken lassen), manche Probleme treten jedoch in beiden Erhebungsmodi auf (Antwort nach sozialer Erwünschtheit; Effekte der graphischen Darstellung) und es kommen online neue Herausforderungen hinzu (Browserkompatibilität; Antworten durch automatische Skripte; Mehrfachteilnahmen). Aber es ergeben sich auch neue Möglichkeiten, die in Papier-und-Bleistift-Befragungen nicht oder nur mit einem erheblichen Aufwand (sehr große Teilnehmerzahl; Messung von Antwortzeiten; komplexe Filterführung; multimediale Stimuli; völlig freiwillige Teilnahme) zu realisieren sind.

Die folgenden zwei Abschnitte zur Nutzeridentifizierung und zum Kompatibilitätsproblem sind für die spätere Datenauswertung von entscheidender Bedeutung, da stets kontrollierbar sein sollte, welcher Nutzer welche Angaben unter welchen Bedingungen gemacht hat. Der anschließende Teil über Messmethoden fokussiert auf die technische Seite der Datenerhebung im Netz. Dort werden die Vielfältigkeit der computergestützten Messungen und der damit verbundenen Einfluss auf Datenqualität und -quantität vorgestellt. Der Abschnitt über webbasierte Skalen zeigt die große Bandbreite von reaktiven Erhebungsinstrumenten in selbstadministrierten Befragungen. Letztlich richten wir - exemplarisch für Innovationen der Datenerhebung im Internet - einen Blick auf dynamische Formulare, die neben neuen Möglichkeiten der Datenerhebung auch eine Fülle von methodischen Fragen mit sich bringen.

Three studies were conducted to examine the effect of
personalized salutation and sender power on signing up to an online survey panel, and subsequent survey response rates. In
the first study, significantly more people joined a panel if addressed by a personalized salutation.
In Study 2, this effect was replicated using an invitation to leave a second panel. In the final study, a significant salutation effect was found when power of the sender was high, and not when power of the sender was neutral. It is argued that for this sample, power of audience and participant identifiability linked to create a compliance-based motivation to join and maintain membership of an online panel. Implications for the maintenance of online panels, survey response rates, and the collection of sensitive personal information, are discussed.

Two studies are presented that examine disclosure of sensitive information and personalization in Internet-based surveys. In the first study, the impact of a personalized salutation on two forms of non-disclosure to a sensitive personal question (salary level) is tested. The results
revealed that a personalized salutation tends to increase levels of active non-disclosure (measured through use of an I prefer not to answer option), but not passive non-disclosure (where the respondent selects no option). In the second study, participants are directed to the study via either a personalized URL (which incorporated an encoded identifier not obvious as such to the responder) or via a secure log-on page that required the user to type in identifying information.
Non-disclosure to a sensitive question (salary) was significantly higher when participants went through a log-on procedure. We suggest that this pattern of non-disclosure to
sensitive questions reflects increases in identifiability when a personalized salutation or logon procedure is used. We further suggest that the provision of an active non-disclosure option
to a sensitive question is particularly appropriate in contexts in which anonymity may be compromised, since it enables participants to both protect their privacy and respond appropriately to the survey.

One major advantage of Web-based research lies in its ability to reach and study people who have rare conditions of interest. Another advantage is that, due to the anonymity of the survey situation, the Internet is particularly suited for surveys on sensitive topics. Sexsomnia is a newly identified medical condition whose sufferers engage in sexual behavior during their sleep. Problematic cases are highly distressing and have forensic implications. The consensus among opinion leaders in sleep medicine is that sexsomnia may be quite common but
that it often goes unreported because of shame and embarrassment. Thus, little is known about this condition’s
demographics and clinical features. This article reports findings from a sample analysis of 20 years of research on sexsomnia and discusses the results, strengths, and weaknesses of a recent Web-based survey conducted on the difficult-to-reach clinical population that suffers from sexsomnia.

A consistent finding reported in online privacy research is that an overwhelming majority of people are ‘concerned’ about their privacy when they use the Internet. Therefore, it is important to understand the discourse of Internet users’ privacy concerns, and any actions they take to guard against these concerns. A Dynamic Interviewing Programme (DIP) was employed in order to survey users of an instant messaging ICQ (‘I seek you’) client using both closed and open question formats. Analysis of 530 respondents’ data illustrates the importance of establishing users’ privacy concerns and the reasoning behind these concerns. Results indicate that Internet users are concerned about a wider range of privacy issues than surveys have typically covered. The results do not provide final definitions for the areas of online privacy, but provide information that is useful to gain a better understanding of privacy concerns and actions.