Logo

General Online Research 2012

March 5-7 2012, Baden-Wuerttemberg Cooperative State University Mannheim

Logo

Conference Agenda

Overview and details of the sessions of this conference. Please select a date or room to show only sessions at that day or location. Please select a single session for detailed view (with abstracts and downloads if available).

 
Session Overview
Session
A+C: Poster presentations
Time: Tuesday, 06/Mar/2012: 2:00pm - 3:30pm

Presentations

Mobile Phone Users Behavior: The Israeli Case

Yaron Ariel1, Eilat Chen Levy2, Sheizaf Rafaeli2

1Yezreel Valley College, Israel; 2Sagy Center for Internet Research & University of Haifa

In Israel surfing the Internet and particularly surfing by using the mobile phones has become a common habit. The mobile phone enabler of communication anytime anywhere, allowing users to be connected in a society typified by special security atmosphere as well as a strong family binding and the Israeli fascination for communication technological innovations (Cohen,Lemish & Schejter, 2008).

According to Jenkins (2004), media convergence changes the relationship between existing technologies, industries, markets genres and audiences. Using the mobile phone we can download information from the internet or send messages.

The Communications Minister of Israel data indicated that by the end of 2010 the number of mobile phone users reached 9.581 million while at the same year the total population rate was 7.695 million*. In 2011, the mobile phone penetration rate in Israel reached 130 per 100 inhabitants. In comparison, there are 2.1 million households in Israel. The fixed Internet connection (high speed) reached 1.7 million with penetration rate of 80% households at the same year.

This paper attempts to analyze the Israeli society surfing habits by mobile phone as well as seeks to shed light on users’ behavioral responses to text and multimedia messages.

Survey results are based on a representative sample of Israeli population, aged 14 and above (n=1502). 95% of the respondents surf the web from their home computer, 36% surf from mobile devices. Most of the respondents (70%) are still not accessing the internet to surf web pages through their mobile phone (n=931, users of '3rd generation' and smartphones). Text and multimedia messages are been sent a few times a day by 55% of the mobile users, and only 17% never used them. 31% use their phone to access their online social network at least one a day.

29% of mobile users defined themselves as "always connected with extensive usages" compare to others who define themselves as "regular" or "occasionally, only when needed" users. Chi square tests indicate that the "always connected" type of users also performs more frequently other online activities (e.g. in web surfing, online social network, and user generated content).

Ariel-Mobile Phone Users Behavior-126.pdf

Politicians' Publishing Behavior

Markus Baumann, Matthias Haber, Christian H. Wältermann

University of Mannheim, Germany

Relevance & Research Question: How political actors utilize communication tools to direct attention to certain issues has been intensively studied by the literature. However, relatively little is known about a politicians timing and emphasis choices in communicating an issue. Our study seeks to fill this gap by exploring the dynamics of issues raised by members of parliament and ministries.

Methods & Results: Using a recently developed unsupervised statistical learning model, we construct a time series data set resting on the political actors' press releases. The period of analysis is the 16th and 17th legislature of the German Bundestag (2005-2011).

Results: Results indicate that members of parliament and ministers communicate within the same set of issues. Furthermore, the actors make similar decisions on devoting emphasis to the topics.

Added Value: Our presentation demonstrates the usefullness of the data generating process and provides insights into the process of political communication and issue competition.


Management of close friendships by the means of Social Network Sites. Results of a longitudinal study.

Alexander Bohn, Nicola Döring

Ilmenau University of Technology, Germany

Relevance and Research Question

Online media, especially social network sites (SNS), are increasingly important for the management of social relationships. In recent research SNS are widely discussed as tools for accumulating social capital and the management of whole relationship networks, including various types of relationships like friends and acquaintances. In contrast the presented study focuses on a specific kind of relationship: the close friendship. Close friendships become more and more important in modern society, because families become smaller and the lifelong marriage is substituted by serial romantic-relationships. Relevant social-psychological theories concerning close relationships (e.g. Social Penetration Theory; Social Interdependence Theory) emphasize that friendships initiate, evolve, persist, deteriorate and end through interactions between the friends. Consequently the aim of the study was to explore interaction practices within close friendships with peculiar interest on SNS.

Methods & Data

Within the scope of a longitudinal study with two waves, students answered a standardized online-questionnaire. In the first wave n1=121 students (69% female, 31% male, average age 21 years) participated in the study. From the participants of the first wave study, n2=63 (70% female, 30% male, average age 22 years) took part in the second wave. The quantitative study was accompanied by a participatory observation of SNS-usage sessions of 8 students (4 male, 4 female) from the recruited sample.

Results

Results of the first study show that students maintain their close friendships most frequently via online communication (SNS, instant messenger, email), followed by mobile communication (SMS, mobile calls) and face-to-face-communication. Visiting online profiles, sending messages and commenting on photos are the most relevant interactions between close friends on social networking sites. The revealed media-related and non-media-related interactions in close friendships remained nearly constant in the second wave. Besides a increasing usage frequency of SNS and a change in the preferred SNS from studiVZ to facebook, no longitudinal effects appeared.

Added Value

The study indicates that SNS have become an established and important tool for students, to maintain and manage their close friendships. Thereby SNS do not replace but complement other ways of relationship management like mobile communication or face-to-face-communication.

Bohn-Management of close friendships by the means of Social Network Sites Results of a longitudinal study-74.pdf

A taxonomy of paradata for web surveys and computer assisted self interviewing (Casi)

Mario Callegaro

Google, United Kingdom

RRelevance and research question

The concept of paradata, also called process data, was first introduced in the survey research community by Mick Couper in 1998 and referred primarily to computer assisted interviewing in general. In 2001 Jeavons introduced the concept of paradata for web surveys at an Esomar event in Barcelona. Paradata are data about the process of data collection and come primarily from the system used in collecting the data.

Paradata have been collected since the beginning of computer administered interviewing and used primarily by survey managers and field supervisors to monitor the progress and quality of the data collection. Since then there have been many usages of paradata.

In this paper I create a new taxonomy of paradata for web surveys.

Methods and data

There are many possible types of paradata that can be collected for web surveys and computer assisted self interviewing (Casi).

Paradata for web surveys and Casi can be classified as

a) device type and

b) questionnaire navigation paradata.

In the first case it is possible to know what device have been used to fill out a survey (e.g.a smart phone or a desktop computer) the location of the device ( by using either the I.P. address or GPS information) and other parameters such as screen resolution and language of the operating system.

Questionnaire navigation paradata capture all sort of movements within the questionnaire, change of answers, time spent per question, mouse clicks and other non questionnaire related data such as activation of error messages, prompts, or clicks on help/FAQ within the survey.

Results

For each paradata types I provide examples of its usage in terms of quality improvements in the questionnaire or data collection, and in terms of understanding of the self administered response process.

Added value

This seems to be the first attempt to create a taxonomy of paradata for web surveys and computer assisted self-administered interviews (Casi). The paper will result in a book chapter already accepted for publication and presented as preview at GOR.

Callegaro-A taxonomy of paradata for web surveys and computer assisted self interviewing-14.pdf

Consumers Online Behavior in a Mature Digital Market

Simon van Duivenvoorde

Wakoopa, Netherlands, The

Relevance & Research Question:

To investigate the general online behavior of consumers in a mature digital market.


Methods & Data:

Passive online audience measurement using a panel of aprox. 10.000 Dutch Consumers. 


Results:

Complete overview of and detailed insights in the online behavior. This is broken down into different categories (e.g. social media, search, shopping, news) and the shares of different sites within those categories.

E.g. social networks have a 37,2% market share based on time spend, e-mail 12,1%, search 9,5%, shopping 9%, news 7,7%, games 6,4%… and much more.

Added Value:

Highly innovative research technique (user side passive online measurement) with a very interesting balance between high level trends and detailed insights.

van Duivenvoorde-Consumers Online Behavior in a Mature Digital Market-146.pdf

Health Anxiety and Internet

Christiane Eichenberg, Carolin Lea Wolters

University of Cologne, Germany

Relevance & Research Question: The term „cyberchondria“ defines the recently described phenomenon of unfounded anxiety or heightened attention to serious diseases based on the notice of web contents. Since the majority of internet users comes back to the internet to gather information about medical questions (Eichenberg & Brähler, in press), and two recent studies have shown correlations between health-related online research and health anxiety (Easton & Guinsler, 2006; White & Horvitz, 2009), further research in this field seems necessary. Therefore, an online-survey was conducted, examining the characteristics of users searching for health-related information on the internet as well as potential relationships with individual dispositions, like health anxiety.

Methods & Data: N = 471 participants were recruited by posting invitations on 182 messaging boards broaching the issues of various diseases and disorders. They were asked to fill out a self-constructed questionnaire as the German version of the „Illness Attitude Scales“ (Hiller & Rief, 2004) to measure clinically relevant health anxiety.

Results: Statistical analyses showed that the internet is more frequently used as health guide by women and older people, whereat the interexchange with other persons concerned on messaging boards seems to be the favored online service. Furthermore, the portion of persons suffering clinically relevant hypochondria was greater among users of health-related internet offers than among the general population. People experiencing health anxiety were found to utilize more different internet health services more frequently, considering them as more reliable and showing more probable behavioral patterns in reaction to the medical information obtained compared to internet users without hypochondria, e.g. ordering medicine via internet, intensifying their online research or consulting a doctor.

Added Value: The study presented hereby confirms previous findings indicating a relationship between hypochondria and the intensity of health-related internet research. User characteristics of those searching for health-related information on the internet were compared to general internet user characteristics described by a recent study (van Eimeren & Frees, 2010). Altogether, the findings emphasize the need for further research.

Eichenberg-Health Anxiety and Internet-136.ppt

Evidence-based information for patients on the Web: An investigation of the quality of internet-sourced information for patients with posttraumatic stress disorder

Christiane Eichenberg

University of Cologne, Germany

Relevance & Research Question: Broad segments of the population use the internet to get information about common health issues as well as specific questions about mental health. Therefore, quality assurance of health websites seems essential. Numerous international studies were conducted to identify lacks of evidence-based quality of health information on the internet, whereas corresponding studies of German speaking websites are to a great extent missing. For this reason, the evidence-based quality of information of German-speaking websites for patients with posttraumatic stress disorder was investigated.

Methods & Data: The research for offers of information about posttraumatic stress disorder was effected on Google’s German speaking websites by entering “posttraumatic stress disorder”. N= 20 of those, containing relevant information for professionals as well as persons concerned, kinsmen and interested laymen, were selected to be examined. These websites were rated analytically according to their content on a 17 dimensional set of categories (with three main categories: the website’s background information, content and design & extras). Each website was evaluated by two independent coders. Interrater-reliability of nominal variables was determined using Kappa coefficients (average k= 0.90) and of interval variables using intra-class-correlation (average ICC (3,1) = 0.84) and turned out to be good to very good.

Results: Altogether, quality of the 20 examined German-speaking websites can be classified to be in a medium range despite differing groups of providers (media, clinics, those concerned, etc.). Websites therefore seem to partly take into account criteria of evidence-based patient information. That evaluation by trend exceeds average evaluation measurements of Anglo-American studies. Misinformation that turned out to be nonhazardous could be identified on 20% of websites. Other studies report higher rates in regard to that issue. The findings show that different kinds of treatments (psychotherapy, psycho-drug therapy, self-help) are equally existent. However, information about psychodynamic therapies is underrepresented.

Added Value: This study is one of the first ones to explore the evidence-based quality of health information on German-speaking websites. The resulting findings are being discussed and implications for the future are made by depicting models and practices for quality assurance.

Eichenberg-Evidence-based information for patients on the Web-10.pdf

Better low-tech than sorry: How technophile questionnaires may affect psychological representativeness

Frederik Funke1,2, Ulf-Dietrich Reips3,4

1LINK Institut, Germany; 2http://frederikfunke.net; 3Universidad de Deusto, Spain; 4IKERBASKQUE Basque Foundation for Science, Spain

Relevance & Research Question: Optimizing online questionnaires for certain browsers not only raises the risk of loosing respondents but also may bias the sample composition. In their GOR paper on the low-tech principle, Buchanan and Reips (2001) found that more educated users were more likely to turn JavaScript off. Furthermore, respondents using Mac OS scored higher on the personality trait Openness to Experience than users of Windows OS. This study focuses on the question if these differences still hold, ten years later and with a sample of experienced Internet users.

Methods & Data: The questionnaire was a Big Five personality inventory. Following the low-tech principle, participation was possible with any Web browser.

Results: Overall, 2.6% of all participants (N = 358) had JavaScript disabled. Male respondents were more likely to have JavaScript disabled than female respondents, chi^2(1, N = 344) = 11.64, p = .001, odds ratio = 15.6. Mac users scored higher on Openness (e.g., “I enjoy hearing new ideas”) than Windows users, F(1, 294) = 12.14, p = .001, eta^2 = .040. Furthermore, respondents using Macs scored lower on Agreeableness (e.g., “I am interested in other people”) than users of Windows, F(1, 294) = 9,02, p = .003, eta^2 = .030. Within respondents running Windows Extraversion (e.g., “I am the life of the party”) was slightly higher for users of the Internet Explorer in comparison to users of Firefox, F(1, 277) = 4.13, p = .043, eta^2 = .015. Finally, there was a tendency that respondents with JavaScript deactivated scored lower on Openness than respondents with this technology activated, F(1, 312) = 3.05, p = .082, eta^2 = .010.

Added Value: The present study confirms and extends the results from Buchanan and Reips (2001). A questionnaire exclusively optimized for certain OSs or browsers can seriously bias the psychological and demographical sample composition. In the present study restricting participation to respondents with JavaScript enabled would have reduced the number of male participants. Overall, it is recommended either to refrain from using complex technologies or to implement alternative low-tech versions of questionnaires as fallback.


High potential for mobile Web surveys: Findings from a survey representative for German Internet users

Frederik Funke, Alexandra Wachenfeld

LINK Institut, Germany

Relevance & Research Question: Mobile Web surveys – i.e., online questionnaires that are accessed with mobile devices like smartphones – offer new possibilities for both behavioral research and consumer research. The aim of our study is to gain more insight into the potential of mobile Web surveys like speed, contacting hard-to-reach populations, or using location data.

Methods & Data: Participants (N = 1501) were recruited from the LINK online panel where all members are actively recruited subsequent to representative telephone interviews. Thus, there is no heavy-user bias, which often can be observed in access panels (see Wachenfeld, 2010). Overall, the panel aims to be representative for the general German Internet population.

Results: Overall 43% of German Internet users have a smartphone at their disposal. Most users (60%) would use a mobile device to participate in a Web survey. However, in the present study only 11% participated using a cell phone. The main reasons were time considerations (24%), not feeling like it (17%), and not knowing that it was possible (12%). As expected, participants using cell phones were more likely to be abroad instead of at home. A further insight is that nearly half of the smartphone owners would allow the read-out of their actual location (e.g., using GPS).

Added Value: Findings are representative for the German Internet users. Overall, there is a considerable potential for mobile Web surveys. New possibilities for mobile data collection (e.g., point of sale studies or mixed-mode designs) as well as the impact on data quality are discussed.


New Media and it´s Role for Political Voice: Public Reasoning in the Field of Labour Market Policy

Simone Antje Gerwert

Universität Bielefeld, Germany

Introduction: In the last two decades, the labour market policy and in particular the fight against unemployment received a “facelift”: Passive measures were substituted by activating programmes, in which “appropriate behaviour” of the beneficiary is rewarded with financial benefits whereas “inappropriate behaviour” is sanctioned. The strict procedures of the activating policy restrict people´s agency and their political voice with regards to young recipients particularly.

Research Question: How do young benefit recipients (re)produce agency and political voice by means of New Media, with specific reference to internet forums?

Method and Data: To get an insight into the “black box” of the production of agency and political voice from the recipient’s point of view, an online ethnography of 89 threads of an internet forum is conducted. The internet forum deals with problems of unemployment, poverty and social exclusion. The data will be analysed by the use of Grounded Theory.

Ethical Aspect: The threads are not produced for scientific research. The information given in the posts is not confidential and specifically given to the researcher, so ethical questions arise on how to deal with this data:

1. I informed the operator of the forum about my research and asked him for (informed) consent

2. The forum announces officially on their front page that all posts and topics could be publicly accessed and published

Preliminary Results: Young people use the internet forum as a social instrument for the production of strategies to achieve agency and political voice. This happens through public reasoning and negotiation processes within the internet forum. The strategies produced online have a deep influence on the offline reality.

Added Value: This study gives an insight into the current problems of the activating labour market policy in Germany. It shows how the political treatment affects the (occupational) freedom of young recipients and how they fight against these restrictions.

Gerwert-New Media and it´s Role for Political Voice-150.pdf

Recommendations for implementing online surveys and simple experiments in social and behavioural research: A review and evaluation of existing online survey software packages.

Claire Marian Hewson

The Open University, United Kingdom

Relevance & Research Question: To date, there exist only a small number of articles which have set out to review and evaluate existing online survey software packages and their features (e.g. Gordon, 2002; Wright, 2005). Furthermore, many of the available reviews do not engage in detailed testing of specific packages within a social/behavioural research framework. Few reviews actually engage in any depth with the extant literature outlining guidelines and principles of good practice for online survey research (for exceptions see Crawford, 2002; Kaczmirek, 2008). The present research adds to and extends these existing review/evaluation studies by carrying out an in-depth testing and evaluation of three popular, low cost, packages, within a context of good design practice and principles in social and behavioural online research.

Methods & Data: First, a large number of online survey software packages were sourced (using resources such as: http://www.websm.org/), and their features assessed against an initial screening list of 'essential criteria', derived with reference to literature on guidelines for good design practice in online survey/research methods. Emphasis was placed on features which support procedures that help maximise data integrity and adherence to ethical guidelines, as well as evidence of longevity, affordability and ongoing development. Second, three packages were selected and each used to implement a) a simple survey design, and b) a simple experimental design. Extensive testing was carried out in order to provide a detailed list of features, functionalities, strengths and weaknesses of each package.

Results: It was concluded that each package reviewed had it's own strengths and weaknesses, research design contexts in which it may be more or less useful, and caveats concerning where it may prove most problematic. Conclusions on which of the packages would prove most suitable for different researcher's needs in different design contexts are offered.

Added Value: The present research fills a gap in the literature by presenting the first (to the best of the author's knowledge) attempt to situate evaluation of online survey software packages firmly within a context of social/behavioural research, by using three such packages to implement actual research study designs.

Hewson-Recommendations for implementing online surveys and simple experiments-95.pdf

User Experience Research with Ipsos Social Listening / Tabet Market Case Study: Samsung Galaxy Tab versus Apple iPad

Sandro Kaulartz, Alexander Kruse

Ipsos / Ipsos Open Thinking Exchange Germany



Relevance & Research Question:

As part of a user experience case study in the German tablet market, Ipsos wanted to show how Social Listening and the insightful field of user generated social web content can add value to research with a combination of technology and research methodologies. The aim of this research was to elaborate a deep understanding of satisfaction driving product characteristics and their interrelation to each other in the German tablet market focusing on the Apple iPad and Samsung Galaxy Tab.

Method:

Quantitative analytics and qualitative immersion techniques have been taken to generate new insights on dissatisfaction drivers or previously unknown product issues. Through the integration of social media monitoring technology, web analytic metrics conclusions on the influential power and impact to the decision making process to peers could be drawn.

Results:

Ipsos discovered the nine decisive theme fields (Weight, Device Size, Touchscreen, Compatibility, Battery Life, Display Quality, Operating System, App Marketplace and Processing) from the web discourse around the two tablet products. In contrast to the Apple iPad, the Samsung Galaxy Tab shows an inconsistent interplay of product attributes. A high disappointment of the app marketplace very often resulted in dissuasion to purchase a Samsung Galaxy Tab. A rather homogenous picture without outlying product characteristics revealed the iPad analysis of theme interrelations. Nevertheless, the in-depth analysis uncovered some disappointing product attributes like the compatibility to other hardware components or the “closed” nature of the Apple operating system and increasing issues display processing quality.

Added Value:

From our case study its clearly visible that Social Listening provides us the opportunity to enter into the new dimension of self-motivated, freely formulated user generated content of peers and an organic ecosystem of opinions.

Our case study proved that the sole buzz monitoring is not conducive to reach the desired analytic deepness to provide market research buyers with a profound understanding of consumer perceptions, beliefs and usage behaviors of the market-compliant target audience. It is indispensable to combine social media monitoring technology with qualitative and quantitative research expertise for a more in-depth market research approach of social media monitoring.

Kaulartz-User Experience Research with Ipsos Social Listening Tabet Market Case Study-51.pdf

TextLab – An Objective, Reliable and Efficient Online-Tool for Readability Measurement

Jan Kercher1, Frank Brettschneider1, Anikar Haseloff2

1Universität Hohenheim, Germany; 2H & H Communication Lab GmbH, Germany

Relevance & Research Question:

Since its formation in the 1920s, readability measurement developed numerous instruments for academic and applied research, the so-called readability formulas. They allow for the comparison of text comprehensibility in an objective, reliable and efficient way. Applications include research on school books, administrative documents and newspapers. Since the 1980s, automated readability tools have been developed. However, academic research has yet largely ignored the potentials of these devices. In this paper, we present TextLab, an automated readability tool which we developed together with an agency for text analysis and optimization.

Methods & Data:

TextLab is an online-tool to measure the readability of any kind of continuous text. For this, users can revert to 22 validated readability formulas and over 150 readability-related text parameters. This also includes two formulas and several parameters and algorithms that were newly developed and offer certain advantages in comparison to the existing ones. The measurement procedure is fairly simple: After some standard adjustments, texts can be pasted into an in input field and are then analyzed automatically by the software. Results can be exported to csv-format.

TextLab has already been used for several readability studies. For example, we analyzed German party manifestos for several elections since 1994, including federal elections, the last election for the European parliament and several elections on regional and local level. Furthermore, we used TextLab to explore the comprehensibility of websites, interviews, speeches, newspapers and banking documents.

Results:

Our results show that TextLab is not only an objective, reliable and efficient but also a valid tool for readability measurements. In a doctoral thesis, parameters that had been collected with TextLab showed strong explanatory power in complex path models that also included relevant recipient characteristics and reactions. We also use TextLab to continuously measure the comprehensibility and the word choice of the major German political parties' online communication. The results show distinct relations between political language use, communicators, issue topics and events.

Added Value:

TextLab is a unique device for the measurement of text comprehensibility with an objective, reliable and efficient online-tool that offers a vast number of validated readability indicators.

Kercher-TextLab – An Objective, Reliable and Efficient Online-Tool-103.pdf

Cyber-hate in the Ukrainian cyberspace

Alexandr Krasnovsky

Kiev International Institute of Sociology, Ukraine

RELEVANCE & RESEARCH QUESTION:

The Internet is the popular platforms of free speech, but it has put the problem of hate content into sharp relief. The scholarships notice a growing number of an open aggression to certain categories of population or events. The situation in Ukrainian cyberspace is aggravated by tough political struggle with "black PR".

This survey was focused on

• the situation with hate discourse in Ukrainian Internet media, especially cyber-hate vectors and objects;

• the attitude of Internet users to cyber-hate, perception of this phenomenon as a problem, the practice of commenting and motives for aggressive comments posting.

METHODS & DATA:

A combined approach was used in the design of the study:

• Content analysis of comments on articles on the 20 most popular and influential Ukrainian socio-political websites. The articles were selected with a probability, which was proportional to commenting activity. As a result, 300 articles and 13,077 comments on them were analyzed.

• Online survey was based on InPoll online panel. The sample included 623 respondents who corresponded to the Ukrainian Internet audience by gender, age, and settlement type.

RESULTS:

The problem of cyber-hate is important for Ukraine because 71% of discussions contain hate expressions. The most frequent objects of hate are politicians and political opponents. The main hate vectors come down to confrontation between the Ukrainians and the Russians and between the residents of Eastern and Western regions. The most Internet users (78%) react to manifestations of cyber-hate negatively, and about 64% of respondents consider this phenomenon socially dangerous. Only 3% of site-visitors reported that they reply to hate comments in similar fashion. The rest try to ignore them, avoid or shift the discussion to a constructive direction.

ADDED VALUE:

Now we are planning next wave of cyber-hate monitoring in the Ukrainian Internet with particular attention to a) forms of hate content in social media; b) most effective responding to cyber-hate; c) correlation between aggressiveness and hate comments posting practice. This work should be finalized and presented by March 2012.

Krasnovsky-Cyber-hate in the Ukrainian cyberspace-63.pdf

The SoSci Panel

Dominik J. Leiner

SoSci Survey, Germany

Relevance & Research Question: Whenever a survey sample shall not represent society but individual internal processes, convenience samples have proven fruitful in research. A significant portion of research history and current research, as well, is therefore based on the economic base of convenience samples. Since 2009 an online panel for the German speaking part goes one step further. By gathering participants from thousands of surveys, a super convenience sample is created to provide samples for non-commercial scientific research – for free. With more than 60.000 members (at a typical participation rate of 20%) the SoSci Panel affords some fine advantages: There is no extrinsic motivation (e.g., cash) for the members to participate in surveys, small and medium samples are quickly available, and the participants feature the heterogeneity that traditional convenience samples regularly lack. The poster presents the SoSci Panel and its demographic margins.

Methods & Data: Form data voluntarily provided by the panel's participants.

Results: The SoSci Panel is far from being representative of the German speaking population, for example comprising a large number of students. However there are more non-students than students, there is a substantial portion of 50+ participants, and the members' regional distribution is a prime example for the panel's heterogeneity.

Added Value: A super convenience sample can dispel some of traditional convenience samples' disadvantages while providing even more convenience for the researcher.

Leiner-The SoSci Panel-40.pdf

An Evaluation of Two Non-Reactive Web Questionnaire Pretesting Methods

Timo Lenzner

GESIS - Leibniz Institute for the Social Sciences, Germany

Relevance & Research Question: Response latency measurement and eye tracking are two computer-assisted pretesting methods that may be particularly useful for evaluating Web questionnaires. In contrast to other techniques (e.g., expert reviews, qualitative interviews), both methods produce nonreactive and objective measures of behavior that are neither affected by the researcher (and the ways in which she tests the questions) nor by the research context. While previous studies have shown that longer response latencies and fixation times are indicative of problematic questions (Lenzner et al., 2010, 2011), little is known about the utility of the two methods (or measures) in the practical pretesting context (e.g., in testing draft questions). This study examines whether response latencies and fixation times are discriminative features to distinguish flawed from improved survey questions.

Methods & Data: In a laboratory experiment, respondents’ eye movements and response latencies were recorded while they were answering two versions of a Web questionnaire. One group (n=22) received a questionnaire including poorly worded questions and the other group (n=22) received the same questionnaire with improved question wordings. Given that response latencies and fixation times are highly individual, we computed the baseline fixation rate (eye tracking) and baseline reading rate (response latency) for every respondent from seven additional questions asked in the same Web survey. In the analyses, whenever the response or fixation times for a question exceeded respondents’ baseline by more than 15%, the question was deemed problematic. (The analyses were repeated with 10%, 20%, and 25% thresholds, but all conclusions remained unchanged).

Results: Fixation rate (eye tracking) was consistently more accurate than reading rate (response latency) in classifying the questions as flawed or improved. The overall accuracy of the fixation rate ranged from 60% to 85%, the accuracy of the reading rate from 43% to 70%. Also, the eye tracking measure resulted in considerably fewer misses (failures to detect problems) and fewer false alarms.

Added Value: This study suggests that fixation times and response latencies are potentially useful methods for pretesting (draft) Web questionnaires, albeit the level of accuracy with which they identify problematic questions is not yet satisfactory.

Lenzner-An Evaluation of Two Non-Reactive Web Questionnaire Pretesting Methods-135.pdf

Inside news aggregation: Understanding algorithms as social systems

Martina Mahnke

Universität Erfurt, Germany

Relevance & Research Question:

As publishing tools get better and easier to use, more and more information is produced, published and shared online. As a result the information being offered on the Internet grows exponentially. In the debate over the Internet and its possible democratizing impact, the common understanding is, that the more information citizens have access to, the better (Balnaves/Willson 2011). On the downside unlimited data leads to information overload (Toffler 1970) and therefore has to be filtered (Shirky 2008). On the Internet, in addition to journalistic filter mechanisms (McQuail 1994, Bruns 2003), new automatic services have emerged. Google News, for example, uses algorithms to select content automatically from over 50.000 sources and personalize it by taking the users behavior into account. Within social science algorithms have been taken for granted as mathematical deterministic entities (Pariser 2011, Meckel 2011). This paper proposes a different approach and looks beyond the mathematical description of algorithms. Therefore the leading research question is: how can algorithms be understood from a social science point of view?

Methods & Data:

In-depth expert interviews with the founders of algorithmic news selection software give insight into the processes behind the mathematical outcome of algorithms.

Results:

This research-in-progress opens up the black box behind algorithms and shows a more comprehensive picture. It suggests, that in order to understand the social impact of algorithmic news selection, it has to be studied as a social system. The vocabulary of systems theory (Luhmann 1984) offers a valuable framework.

Key results in brief: Function: Information selection; Code: Linked/Not-Linked; Reference: Social Reality; Product: Selection; Output: Information

Added value:

This paper focuses on a social sciences point of view on algorithms and thereby offers a new theoretical perspective. By looking beyond the mathematical description this research develops a theoretical framework that helps to ground future research in this area.

Mahnke-Inside news aggregation-148.pdf

Is Pretesting Established Among Online Survey Tool Users?

Nataliya Moeser1, Carsten Schmitz2, Guido Moeser3

1masem training GmbH, Germany; 2limesurvey.org; 3masem research institute GmbH, Germany

Relevance & Research Question:

Pre-testing survey instruments is a widely accepted method to test a survey before fielding it to the full sample. There are several advantages of pretesting an online survey, e.g., to collect information regarding survey responses or technical problems, to name only a few.

This study addresses the following questions:

(1) Is pre-testing used before fielding the online survey instrument to the full sample?

(2) Does frequency influence the usage of online survey tools and pre-testing?

(3) Are there differences in usage behavior across different areas (e.g. government, academic, non-profit or for-profit)?

(4) How large is the pretest sample?

Methods & Data:

Data has been collected via a web survey among users of LimeSurvey from July 2009 to October 2011. Of 40,663 responders 14,622 answered the question if they ever ran a pre-test or not. Analyses were conducted using descriptive statistics, cross-tabulations and related statistical tests.

Results:

The core result is that pre-testing online surveys before fielding it to the full sample is now an established method:

(1) Around 65.9% of online survey tool users occasionally or always run a trial survey (pre-test) before fielding it to the full sample.

(2) There is a u-shaped significant relationship between usage frequency of online survey tools and conducting a pre-test (…).

(3) The amount of pre-tests is highest among academic users (62.9%) and lowest among governmental users (58.3%).

(4) Around 64.5 percent of online survey tool users conducted a pre-test using a pre-test sample size of between 1 and 100 cases.

Added Value:

The results of this study provide insights to pre-testing online survey tools. Despite the fact that nearly two-third of online survey tool users are running pre-tests with between 1 and 100 participants, there is potential for the remaining one third of customers to improve their quality of online surveys through pre-testing. Overall, this should lead to a higher acceptance of online surveys.


VisAWI-S – screening for visual web aesthetics

Morten Moshagen2, Meinald Thielsch1

1University of Muenster, Germany; 2University of Mannheim, Germany

Relevance & Research Question:

Visual aesthetics has been shown to critically affect a variety of constructs besides usability, such as satisfaction, pleasure, urge to buy impulsively, preference, customer loyalty, and intention to revisit. The Visual Aesthetics of Websites Inventory (VisAWI; Moshagen & Thielsch, 2010) is a sound instrument for assessing four main facets of visual aesthetics (simplicity, diversity, color, and craftsmanship). However, the VisAWI may be too lengthy for certain purposes, for example when needing a simple manipulation check, or when a global assessment of aesthetics suffices. The VisAWI-S addresses this need for a very brief but well-founded assessment.

Methods & Data:

Based on theoretical considerations and the factor loadings as obtained in the original studies used to construct the VisAWI, we choose one item from each of the four scales to represent the general aesthetics of websites factor. In a series of three studies with a total of 1.673 participants we tested the VisAWI-S with respect to dimensionality, reliability and validity. While doing so, 60 different real and 24 mock websites were assessed.

Results:

The VisAWI-S is a unidimensional and reliable measure (α = .81) that provides a good approximation to the full-length version (r = .91). Convergent validity was established by a strong relationship to overall appeal (r = .72). Evidence for divergent validity was obtained by weaker correlations to perceived usability, pragmatic quality, and quality of content (.34 ≤ r ≤ .54) as well as by absence of a significant correlation to participants’ mood before using the website in question (-.05 ≤ r ≤ -.07). In addition to this, the VisAWI-S was found to be substantially related to the intention to revisit a website (r = .52).

Added Value:

Overall, the VisAWI-S may gainfully be employed to measure perceived visual aesthetics of websites when assessment times must be kept to a minimum. Considering that the VisAWI-S comprises merely 4 items, the scale showed impressive reliability and validity. Therefore, the VisAWI-S is of practical benefit not only for online and market researcher but also for practitioners designing web user experience.

Moshagen-VisAWI-S – screening for visual web aesthetics-149.pdf

“Core oder Casual??”: A quantitative survey on usage and gaming history of social gamers

Arne Müller, Daniel Schultheiss

Ilmenau University of Technology, Germany

Relevance & Research Question: In recent years no other social network showed a similar growth like Facebook. 500 million users, the integration in any smartphone and on gaming platforms like the Xbox 360 shows how important this social network became for users and companies. Facebook as social network is as usual as the Google search in any browser. For millions of users Facebook is not only a channel of information and exchange, it is a channel of gaming. 98 Million monthly users in one game (City Ville) were not reached before by any other game in history. So we ask two research questions: 1. What usage behavior and usage history show social gamers? 2. Could the factors gaming-history, usage and demography used for a new typology of social gamers? To answer the research questions we focused theoretically on a model of computer game usage by Wolling, Quandt and Wimmer (2008).

Methods & Data: For this purpose, a quantitative self-selective online survey of (N=678) social game players has been carried out on the social networking site Facebook and several game fan sites. In addition to a descriptive evaluation of the collected usage and personal data a cluster analysis was used to generate a typology of social gamers.

Results: Core results show that social games are used more intensive than expected. We can show six different types of players (conventional casual gamers, passionate core gamers, converted minimal gamers, adapted casual gamers, consistent core gamers, and family-focused minimal gamers) which were identified using cluster analysis. It turned out that among all social gamers many hardcore players exist beside the typical casual gamers. Additionally, three groups of players could be identified which have developed into different clusters of our typology despite similarly extensive gaming histories in the past.

Added Value: We deliver an insight into a completely new field of online games, the users and their usage behavior. Additionally we deliver a new typology of gamers which improves the old two-parted (core vs. casual) typology.

Müller-“Core oder Casual”-140.pdf

Continuous large-scale volunteer web-surveys: The experience of Lohnspiegel and WageIndicator

Fikret Oez

Institut Arbeit und Technik in der FH Gelsenkirchen, Germany

Relevance & Research Question:

Online Surveys has been playing increasingly an important role in the academic research. The Institute of Social and Economic Research (WSI) has been conducting a continuous web survey on pay and working conditions since 2004. Data is collected via an online questionnaire, accessible at www.lohnspiegel.de, which any interested visitor can complete. Lohnspiegel is a German part of the WageIndicator survey network, which organises web-surveys with a similar structure and content in 50 countries. The topics relate to wages and working conditions. The target population is the labour force, including employees, self-employed, informal workers, job seekers and the unemployed. Volunteer web-surveys are primarily held in marketing and in voting research. Except for the WageIndicator they are hardly used in the field of work and employment. The presentation concentrates on the experiences of the Lohnspiegel and WageIndicator as a continuous web survey and on the recommendations for other academic researchers.

Methods & Data:

The respondents of the web-survey are volunteers. A continuous survey challenges the traditional perception of web-surveys. The survey have several advantages: the potential to generate large sample sizes, the possibility of exploring rare groups or small-scale units, the inclusion of follow-up questions, Additionally, such web surveys offer multiple client-side feedback systems.

Results:

Since 2004 over 200.000 participants have completed the online questionary in Germany, with an average of ca. 25.000 per year. The Lohnspiegel is a non-probability survey and the data therefore biased and not representative to a certain extent. Several adjustment procedures to solve this problem have been applied. The results show that application of different weights produces rather diverse results, and that there is no certainty as to whether the representativeness of such web-surveys can be improved.

Added Value:

The experience of a continuous large-scale web-surveys “Lohnspiegel” and “WageIndicator” in general, would give new insights for further academic research and methodology development. The surveys generate longitudinal, multi-country data on wages, which might develop into a worldwide database on wages, benefits, working hours, working conditions, and industrial relations and, hence, contribute to a better understanding of the labour market worldwide.

Oez-Continuous large-scale volunteer web-surveys-121.pdf

WAGE INDICATOR, web surveys methodology, WEBDATANET and Life satisfaction determinants as an example of content research.

Pablo de Pedraza

University of Salamanca, Spain

Relevance and research question. Social scientist need quick access to internationally comparable data which can be done using web surveys. A good example is the Wage Indicator (WI): a continuous voluntary web survey with an homogeneous questionnaire in 65 countries.

The poster is used to show WEBDATANET members the possibilities offered by the WI for both methodological and content research. Firstly, it shows research conducted using Wage Indicator data. Secondly, it uses as an example of content research a study of how work characteristics, labour situation and labour preferences determine life satisfaction of an on-line sample of Spanish workers. The paper obtains useful methodological conclusions and open new opportunities for Life Satisfaction research.

Methods and Data. The paper uses a sample obtained in the Wage Indicator. Online voluntary web surveys like the Wage Indicator are non-probability surveys and results obtained from their data cannot, in principle, be generalized to the whole population of interest (the labor force). There is a three-step selection process: internet access, interest and decision to take up the web-survey.

The poster shows methodological approaches implemented to increase WI data quality (bias, weighting techniques, marketing and targeting measures to address underrepresented groups and paradata analyses). Finally, it uses probit regressions to estimate nested models of Life Satisfaction Determinants.

Result. Obtains useful conclusions for web survey methodology and shows the huge possibilities of the Wage Indicator for methodological research. Regarding life satisfaction research, results obtained in simple models do not differ from literature. New conclusion are obtained regarding life satisfaction explanatory variables and new research lines are open.

Added value. It has implications for several research lines within social sciences. Firstly, workers' happiness determinants and the future possibility of making global, real-time comparisons. Secondly, although conclusions are obtained from this online non probabilistic survey, they are in line with theory and literature. Thirdly, happiness determinants of these self selected workers are important per se (online participation, is becoming more and more important) although conclusion may not be applicable to the whole population.

de Pedraza-WAGE INDICATOR, web surveys methodology, WEBDATANET and Life satisfaction determinants as an example-48.pdf

Online research for general populations: whether and how results could be extrapolated?

Małgorzata Półtorak, Jarosław Kowalski

IIBR (Gemius Group), Poland

Relevance & Research Question:

According to ESOMAR, online survey is already the most popular quantitative market research technique in the world (regarding research spending and number of interviews). The results of studies conducted via the Internet are frequently extrapolated to general populations and/or compared to results obtained by other modes. And thus the questions appear: Whether and in what circumstances could such generalisations be accurate, given limited coverage and systematic nonresponse error in online samples? Is the sole correction of socio-demographic quota enough to ensure results validity? If not, what other sample characteristics shall be controlled? Do the comparisons of online and offline results always make sense? How big is the impact of sampling effect and technique effect?

Methods & Data:

To answer the above questions we conducted a set of online and offline studies in 2009 and 2010. We paid a special attention to keeping different modes of measurements equivalent, according to timing, wording and look of the questions etc. We considered various research phenomena and various features for correction. The final study, which we will focus on here, took place in May 2010, in Poland. We gathered 3006 questionnaires from personal interviews, 5485 from online on-site recruitment and 6399 from online AccessPanel were gathered. The comparisons comprised of raw data results as well as of data weighed in various ways.

Results:

Demographic data correction makes online results better estimate some offline results, for some populations. However, just including additionally non-demographic characteristics put the modes differences within the bounds of a random error. Not all straight comparisons of online and offline results make sense. A technique and interviewer effects are for a range of phenomena not removable. We are going to show all the above in details during our speech.

Added Value:

The presentation will provide the audience with valuable insights of what needs to be known and controlled in online research to ensure high quality and validity of the results and when offline results can be a real reference to online. The topic shall be of interest not only for market researchers but also for academics.

Półtorak-Online research for general populations-91.pdf

Thinking, Planning & Operationalizing Empirical Mixed Methods Research Design

Umar Ruhi

University of Ottawa, Canada

Relevance & Research Question: Mixed methods research designs comprising qualitative and quantitative data collection & analysis mechanisms offer many benefits to researchers. While triangulation of research findings is commonly considered to be the primary rationale for many mixed methods studies, researchers can also utilize a combination of methods to complement, elaborate, and add richness and detail to better understand research phenomena. This poster aims to exemplify the design of a mixed methods study by highlighting the process, procedures and products from a recent empirical investigation of member participation in online communities.

Methods & Data: As a model for mixed methods research design, the study demonstrated in this poster utilized an exploratory mixed methods research design to collect and analyze qualitative and quantitative data from members of various social, professional and commercial virtual communities. Toward this, the qualitative and quantitative approaches were utilized sequentially with the quantitative phase aiming to provide convergent validity for the findings from the qualitative phase. An adapted grounded theory methodology was used in the qualitative phase of investigation, while the quantitative analysis phase utilized structural equation modeling techniques. These phases of research drew from different sampling frames (qualitative: n = 68 ; quantitative: n = 232), and a combination of online focus groups, a qualitative web questionnaire, and a quantitative web survey instrument were used for data collection.

Results: In terms of outcomes, this poster exemplifies frameworks that can be used to convey research findings and results from qualitative and quantitative phases of investigation, and techniques that can be used for the integration of results from multiple phases. Qualitative analysis tools and techniques such as mind maps, concept maps, tag clouds, qualitative coding queries and coding matrices will be exemplified, along with their integration with quantitative exploratory factor analysis, and structural equation modeling procedures.

Added Value: Suggestions for operationalizing and performing mixed methods research are made.

Ruhi-Thinking, Planning & Operationalizing Empirical Mixed Methods Research Design-154.pdf

Automatic Forwarding on Web Surveys – Some Outlines and Remarks

Arto Selkälä

University of Lapland, Finland

It can be said that at one end of the design continuum of web surveys are form-based designs that present questionnaires as one long form in a scrollable window; at the other end are screen-by-screen questionnaires that present only a single item at a time. The automatic forwarding is a web survey function that leads a respondent immediately forward after answering. As a function it becomes close of the isolating procedure of single items. The automatic forwarding is rarely used but adopted in the web survey ZEF.

The ZEF shares the single questions into the separate files. The respondent is able to achieve the items of single file by clicking the file icon. After that the list of questions that the file include, become visible. The respondent is then able to perceive only the options for the one single question at the time, when the options for other questions in the file remain hidden. When answering, the automatic forwarding occurs as an immediate hiding of answer as soon as a radio button has been checked.

The procedure of automatic forwarding in ZEF leads to the lack of conventionally presented visual cues during answering process. It seems to removes the visual feedback and might be causes the loss of context. If a respondent makes a mistake - either because of misreading or because of problems with the mouse or the touch screen - it cannot be corrected. On the other hand, the benefits of individually presented questions on self- administrated questionnaires have been proved in many studies. The response times are longer for questions presented individually than for questions presented as a set. Answering the former takes more time, because the respondent must establish a cognitive orientation for each question separately. Clearly, the respondent must expend more cognitive effort to interpret individually presented questions. When trying to understand the possible consequences of automatic forwarding, the perspective of individually presented questions as well as the perspective of grouping questions should be taken account. Still, many questions remain open in automatic forwarding issue.

Selkälä-Automatic Forwarding on Web Surveys – Some Outlines and Remarks-151.pdf

Comparing Item-Non-Response and Open Questions within different Web Survey Types

Henning Silber1, Julia Lischewski2, Jürgen Leibold1

1Center of Methods in Social Sciences, Göttingen University, Germany; 2Institute of Sociology, Göttingen University, Germany

Relevance & Research Question:

Online research frequently employs a number of different types of Web Surveys. This research addresses the question of differences between these diverging types of Web Surveys in terms of Item-Non-Response and answers to Open Questions. Identical questionnaires are used, focusing on prejudice and sensitive questions, in particular questions concerning attitudes towards Jews and Muslims in Germany. The research includes a new Web Survey Type recruited by a random sample drawn from a systematically generated pool of email addresses.

Methods & Data:

Data was collected via three different types of Web Surveys. The field time of the three samples was from March to May 2011. The first Web Sample is an Online-Recruited Online-Access-Panel (n=500), the second sample an Offline-Recruited Online-Access-Panel (n=200) and the third is a Web Sample (n=100) drawn from a random sample of a systematically generated pool of email addresses of different providers. In all three different types of Web Survey the identical questionnaires were employed in order to attain comparability.

Results:

The primary result is that the outcome is heavily dependent on the type of survey employed. While the first and second survey types show almost the same result concerning Item-Non-Response and response rate to Open Questions, the third has significantly lower Item-Non-Response and fewer, but longer answers to Open Questions. In addition the overall interview time differs between the three different Web Sample Types, being significantly longer for type three.

Added Value:

The core implication for further web research is that it would be too simplistic to divide the world of surveys in online and non-online surveys. Which type of Online Survey is used has a significant impact on data quality. Sensitive questions are particularly unimmunised within the usage of specific types of Web Surveys. By deciding on any type of Online Survey, researchers should keep that in mind and choose carefully.

Silber-Comparing Item-Non-Response and Open Questions-138.pdf

Opinion Leadership 2.0 – A Quantitative Study on Opinion Leadership in Social Networks

Saskia Stäudtner, Alexander Bohn, Nicola Döring, Daniel Schultheiss

Ilmenau University of Technology, Germany

Relevance & Research Question: Online social networks as research objects are interesting in various senses. In comparison to other online media they are extremely frequently used. Also both genders from all social classes are represented equally (Roth and meadow, 2011). Among other things the relevance of identification of so-called opinion leaders in online social networks is based on the fact that they are multipliers. Thus they are decisive involved into the spread of information and opinion. They often occupy the roles of the model and the trend setter as well (I-cod ltd., 2009).

In one empirical study we want to examine which characteristics identify opinion leaders on the online social network Facebook. We also want to examine in what way the determined results agree with theoretical approaches of the classical opinion leader research. Besides socio-demographic differences our main focus is on information behavior of opinion leaders and non-opinion leaders.

Methods & Data: Instrument for gathering data was an online survey. The recruitment of our self-selected ad-hoc-sample was done through distributing the link of our survey in selected forums and groups on Facebook from the end of June until the middle of July 2011. To answer our research questions n=463 valid questionnaires (of a total sample of 988 participants) were used for a quantitative analysis.

Results: It could be determined that Facebook users form a relatively homogeneous group based on their socio-demographic and social characteristics.

The assumtions of previous theoretical work could be confirmed mainly for the information behavior of the users. Results also show significant differences regarding opinion leaders on Facebook. They are more informed and search clearly more active for information than non-opinion leaders.

Added Value: Opinion leadership as well as social networks are a relatively new research area. Because of the very broad usage of social media the combination of both subjects was obvious. Characteristics of opinion leaders in social networks could be identified in a first study, so we could deliver a first contribution in the area of the opinion leader research in new media and social networks.

Stäudtner-Opinion Leadership 20 – A Quantitative Study on Opinion Leadership-152.pdf

Little experience with technology as a cause of nonresponse in online surveys

Bella Struminskaya, Ines Schaurer, Lars Kaczmirek, Wolfgang Bandilla

GESIS - Leibniz Institute for the Social Sciences, Germany

Relevance and Research question: It is well acknowledged within online survey research that Internet surveys are prone to noncoverage error when panelists are not provided with Internet access. However, even Internet users may differ in their ability and comfort with the Internet technology. This may lead to differential willingness to participate in online surveys within the group of Internet users. If respondents differ in their levels of comfort with the Internet technology from nonrespondents, this may result in a biased sample. We study how experience with technology affects the willingness to participate in an online survey and whether respondents to the online questionnaire differ from nonrespondents in levels of technological sophistication.

Methods and Data: The data were collected in telephone recruitment interviews for a probability-based online panel. The landline and mobile telephone interviews were conducted from February to September 2011. In the course of the interview along with demographic information and some attitudinal questions, Internet-usage related questions were asked. Experience with technology, measured via several questions on Internet proficiency and mobile phone ownership and use, is studied as an addition to a framework for unit nonresponse in surveys (Groves & Couper 1998).

Results: It can be shown that those having lower levels of experience with technology are less likely to be willing to join an online panel. Some of the differences stay significant in the selection step of actual online participation. However, no effect of experience with technology can be found on data quality (marginal effect on breakoff and none for item nonresponse).

Added Value: This study implies the importance of respondents’ experiences with technology for nonresponse in various steps of the recruitment process for a panel. Taking this selectivity aspect into account may prove useful for improving recruitment protocols.

Struminskaya-Little experience with technology as a cause-137.pdf

From first impression to recommendation – users’ view on websites

Meinald Thielsch1, Rafael Jaron2

1University of Muenster, Germany; 2Nordlight Research GmbH

Relevance&Research-Question: For most enterprises and institutions are websites an essential part of marketing and communication. When we visit a website we build immediately a first impression, a personal opinion about the site, its content, usability and beauty. Although each of these aspects has been researched individually, little is known about their interplay. And what about our general impression or our intention to revisit or recommend a website? Which impact has each construct - content, usability, and aesthetics - to different stages of website use?

Methods&Data: In a first study with 330 participants we asked web user directly about their value for the different constructs in respect to the use stages. In two following studies with a total of 812 participants we tested 46 websites regarding users evaluation of content, usability and aesthetics. With regression analyses we determined the impact of each construct on users’ reactions to a website depending on the time and the focus of the evaluation. While doing this in two independent studies, we were able to cross-validate our findings.

Results: In our first study content was rated as most relevant for all four use stages. Usability was rated second, except for first impressions, for which aesthetics was second in importance. To sum the results of our two following website studies up: Aesthetics had the largest influence on first impressions (β = .47**), while all three constructs had an impact on first and overall impressions (.16** ≤ β ≤ .49**). The evaluation of content is very important for the intention to revisit or recommend a website (.60** ≤ β ≤ .64**), here was no influence of usability found (β < .05, ns). But aesthetics showed a small but significant impact at this point (.09* ≤ β ≤ .11*), especially if the website in question was unknown (.18** ≤ β ≤ .19**).

Added-Value: The understanding of processes behind our website use is of theoretical and practical benefit not only for online and market researcher but also for practitioners designing web user experience. Our results indicate, that aesthetics drive the perception of company website attributes in early stages of information processing.


Effects of number of response options in web surveys: The role of verbal labels

Fanney Thorsdottir1, Marek Fuchs2, Gudbjorg Jonsdottir3

1University of Iceland, Iceland; 2Technische Universität Darmstadt, Germany; 3Social Science Research Institute, University of Iceland, Iceland

Effects of number of response options in web surveys: The role of verbal labels

Relevance and research question: Fully labelled agree/disagree rating scales are often used to obtain data pertaining to subjective phenomena in web surveys. Although the link between the number of response options and the quality of information obtained with rating scales is well established, the role of the verbal labels attached to the options is less well studied. The verbal lables define the length and the fineness of a fully labelled rating scale and adding response options may therefore either lengthen the scale and/or make it finer. The aim of the present work it to examine the impact of altering the number and verbal labels of response options on the quality of the information obtained. Fully labelled rating scales with five, seven and nine response options are examined. Verbal labels are either used to lengthen the scales or to make them finer.

Method and data: University students in Germany and Iceland were presented with six questions measuring attitude towards university education in a web survey. Respondents were randomly assigned to six experimental conditions. 1) Five response options with the labels strongly, somewhat and neither/nor, 2) five response options with the lables fully, somewhat and neither nor, 3) seven options with the labels fully, strongly, somewhat and neither nor, 4) seven response options with the labels strongly, somewhat, slightly and neither/nor, 5) seven response options with the labels fully, somewhat, slightly and neither/nor and finally 6) nine response options with the labels fully, strongly, somewhat, slightly and neither/nor.

Results: The results show that the effect of the number of response options on the quality of the information obtained depends on whehter the added response options lengthen the scale or make it finer. A finer scale seems to reduce non-differentiation and extreme response style.

Added value: The findings show that the verbal labels attached to the response options must be taken into account in empirical research on the optimal number of response options.

Thorsdottir-Effects of number of response options in web surveys-143.pdf

Approaches for validating automatic Analytic Tool results on social networking data for its Exploitation within Politicians' everyday Workflow

Timo Wandhoefer

GESIS - Leibniz Institute for the Social Sciences, Germany

Relevance & Research Question:

This poster discusses criteria and approaches for validating results from automatic analytic tools on social networking data for its exploitation within politicians’ everyday workflow with respect to eParticipation. The outcome could be used to identify a borderline how many posts one politician is possible or willing to read and what is the count of comments an automatic analysis tool delivers similar or better results?

Methods & Data:

The idea is to use the politician’s experiences on social networking sites (Abbr. SNS) data to assess results from automatic analysis tools with the methodologies of the content analysis. The content analysis will be applied on the toolbox developed by the EU project WeGov - Where eGovernment meets the eSociety. This research project aims to connect citizens’ opinions on social networking sites like Facebook with the decision-making process of politicians’. In theory the WeGov software provides politicians’ summarized SNS content. The software uses automatic analytic tools like “topic opinion analysis” that is based on the Bayesian algorithm. What SNS users taking about the topic nuclear phase-out in Germany and what topics citizens are discussing within a politician’s constituency or in general are main functionalities of the toolbox. In practical the quality and exploitation of automatic analysis results within the politicians’ everyday workflow is not jet validated against “everyday” SNS data.

Results:

The idea for this poster takes advantage of the politicians’ experiences and their sensitization with respect to SNS data to design criteria and a content analysis on automatic analysis results. The development process of the WeGov toolbox with sixteen semi-structured interviews at different parliaments has confirmed that politicians using SNS for public relation and presswork purposes. Another outcome was that most of the interviewees are observing theme groups on Facebook and especially local Facebook groups within their constituency.

Added Value:

Consequentially automatic extracted topics from SNS, especially on the local level, might be an opportunity to launch a content analysis carried out with politicians. A positive outcome might be that politicians get informed earlier about upcoming topics in difference to their everyday workflow of presswork.

Wandhoefer-Approaches for validating automatic Analytic Tool results on social networking data for its Exploitation-141.pdf

Using SKOSified vocabularies to support scholarly research based on the upcoming Linked Data model of Europeana

Dov Winer1, Allison Kupietzky2, Kostas Pardalis3

1MAKASH - Israel, Judaica Europeana EAJC - UK; 2Israel Museum Jerusalem Israel; 3National Technical University of Athens

Relevance:

Linking Open Data (LOD) is extending the Web with a data commons with the publication of various open data sets as RDF on the Web and setting RDF links between data items from different data sources. The resulting structured Web can be queried through the SPARQL query language; crawled by RDF search engines, browsed by RDF enabled browsers. Such tools feed innovative applications like mashups using a universal API. In January 2012 there were 31,634,213,770 Triples and 503,998,829 Outlinks in the Linked Data cloud. Projects like LOD2 and ResearchSpace are applying such resources for online research. Europeana, the European digital library, museum and archive developed its new Europeana Data Model(EDM) based on a Linked Data approach using OAI-ORE for the representation of metadata and SKOS/RDF for vocabularies. Vocabularies are a critical tool in navigating and applying LOD data. Judaica Europeana supports the implementation of EDM in the Jewish thematic domain to provide answers to queries like What? Who? When? Where? The initial application of this program is demonstrated using the thesauri of the Israel Museum Jerusalem(IMJ).

Methods & Data:

The IMJ tables for Objects, Events, Places and Artists were expressed in RDF/SKOS. Alternative ways to access the Europeana collections were developed (1) Multilingual access: the URI identifier for concepts in the thesauri supported the translation of queries from Hebrew to English (2) The hierarchical tree of the thesauri serves as a browserfor accessing Europeana content with each concept leading to the relevant content.

Results:

Over 30,000 terms from the Hebrew and English thesauri were SKOSified from four tables: Artists, Places, Periods and Objects. Multilingual access to Europeana through the concepts of the IMJ vocabularies has been established. A browse tree application for the table for Objects has been concluded and possibly another one for Judaica Terms and Places will be developed as well. See: http://judaica-search.image.ntua.gr

Added value:

LOD enrich the scope of Web based research and should become part of the GOR agenda. The present work support the work of online scholarly research through access to a knowledge base of cultural heritage through SKOSified vocabularies.

Winer-Using SKOSified vocabularies to support scholarly research based on the upcoming Linked Data model-142.pdf


 
Imprint · Contact Address:
Conference: GOR12
Conference Software - ConfTool Pro 2.6.39
© 2001 - 2011 by H. Weinreich, Hamburg, Germany