Johannes Naumann

Attitudes toward the computer : construct validation of an instrument with scales differentiated by content

In the present study, the construct validity of an instrument for the content-differentiated assessment of attitudes toward the computer is examined. The instrument is designed primarily for university students of the humanities and the social sciences. It contains eight scales related to different content aspects which are based on three dichotomous distinctions: (1) the computer as an instrument for learning and working versus the computer as an instrument for entertainment and communication; (2) personal experience with the computer versus consequences of computer technology for society; and (3) the computer as a ‘beneficial tool’ versus the computer as an ‘autonomous entity’ (cf. Brock & Sulsky, 1994. Attitudes toward computers: construct validation and relations to computer use. Journal of Organizational Behavior, 15, 17–35). For validation purposes, confirmatory factor analyses were conducted (n=232). The measurement model of content-differentiated scales was contrasted with five more restrictive (nested) models assuming bidimensionality (‘beneficial tool’ vs. ‘autonomous entity’), and estimates for the reliability and discriminant validity of the eight attitude scales were computed. The results fully corroborate the measurement model of content-differentiated scales. Consequences for the assessment of attitudes toward the computer in general are discussed.

Lexical Quality and Reading Comprehension in Primary School Children

In a cross-sectional study, we examined the relationship between the quality of lexical representations and text comprehension skill in German primary school children (Grades 1–4). We measured the efficiency and accuracy of orthographical, phonological, and meaning representations by means of computerized tests. Text comprehension skill was assessed with a standardized reading test with questions requiring recognition of text information and inferencing. Both the accuracy of and the efficiency of access to the three types of lexical representations contributed to explaining interindividual variation in text comprehension skill. Results from a path-analytic model suggest a specific causal order of the three components of lexical quality with the quality of meaning representations partly mediating the effects of form representations.

Metalinguistic awareness mediates effects of full biliteracy on third-language reading proficiency in Turkish–German bilinguals:

Bilingualism has been reported to be positively associated with both metalinguistic awareness and third-language (L3) acquisition. In the present research, the assumptions were tested that literacy in both first (L1) and second (L2) language (full biliteracy) is needed for bilingualism to be positively associated with L3 reading proficiency, and that positive effects of full biliteracy on L3 reading proficiency are mediated through metalinguistic awareness. L1, L2 and L3 reading proficiency and metalinguistic awareness were measured in 299 German and Turkish-German secondary school students. Overall, fully biliterate students outperformed monolingual and partially biliterate students in both L3 (ΔR2 = .07) and metalinguistic awareness (ΔR2 = .06). An effect of full biliteracy on L3 reading proficiency persisted when SES, gender, general cognitive ability and school track were controlled for. In addition, within the group of biliterate students, positive effects of the individual degree of biliteracy on L3 reading proficiency could be shown to be an indirect effect that was mediated through metalinguistic awareness.

LOGPAT: A semi-automatic way to analyze hypertext navigation behavior

In hypertext research, log files represent a useful source of information about users’ navigational behavior. Since log files can contain enormous amounts of data, methods for data reduction with a minimum loss of information are needed. In this paper, LOGPAT (Log file Pattern Analysis) is presented, a Web-based tool for analyzing log files. With LOGPAT, single-unit, sequential, and graph-theoretic measures (including distance matrices) for the description of user navigation can be computed. The paper gives an overview of these methods and discusses their value for psychological research on hypertext. Components and analysis options of LOGPAT are described in detail. The program’s basic options are illustrated by data from a study on learning with hypertext.

Strategische Verarbeitung beim Lernen mit Text und Hypertext

Zusammenfassung: Beim Lernen mit Hypertext mussen Lernende ihren Lernweg selbststandig zusammenstellen und auf verschiedene Knoten verteilte Lerninhalte zu einer koharenten Gegenstandsreprasentation integrieren. Diese Anforderungen sollten durch den Einsatz kognitiver und metakognitiver Lernstrategien besser bewaltigt werden konnen. In diesem Beitrag werden Ergebnisse aus einer experimentell-korrelativen Untersuchung zur Rolle von Lernstrategien beim Lernen mit linearem Text und Hypertext berichtet. Psychologiestudierenden (N = 30) wurden studiumsnahe Lernaufgaben gestellt, die anhand eines Hypertexts oder eines inhaltsgleichen Lineartexts zu bearbeiten waren. Lernerfolgsmase wurden uber eine Inhaltsanalyse von Lernskripten ermittelt. Der Strategieneinsatz wahrend der Textrezeption wurde uber Protokolle lauten Denkens erfasst. Zusatzlich wurde das Navigationsverhalten in Logfiles aufgezeichnet. Der Einsatz kognitiver Lernstrategien erwies sich unabhangig von der Textstrukturvariante als lernforderlich. Zu...

How adolescents navigate Wikipedia to answer questions / ¿Cómo navegan los adolescentes en Wikipedia para contestar preguntas?

Abstract In one experiment, we explored how high school students use hyperlink relevance cues while they navigate to answer questions from hypertexts. Current evidence has shown that students may navigate by either performing a deep semantic analysis of the relationship between the question and the existing hyperlinks, or by matching words in the question to words in the hyperlink labels. We focused on how students combine both cues during navigation, and how comprehension skills relate to the use of such cues. Our study revealed that 14 year old students (N = 53) selected hyperlinks by relying to a similar degree on both word matching and semantic overlap. Furthermore, when there was a conflict between an irrelevant link cued via word matching and a relevant link only cued through semantic overlap, students’ comprehension skills facilitated their initial selection of an informative relevant link. To conclude, we discuss the implications of these results for current models of hypertext navigation.

Scanning and Deep Processing of Information in Hypertext: An Eye Tracking and Cued Retrospective Think-Aloud Study.

When students solve problems on the Internet, they have to find a balance between quickly scanning large sections of information in web pages and deeply processing those that are relevant for the task. We studied how high school students articulate scanning and deeper processing of information while answering questions using a Wikipedia document, and how their reading comprehension skills and the question type interact with these processes. By analyzing retrospective think-aloud protocols and eye-tracking measures, we found that scanning of information led to poor hypertext comprehension, while deep processing of information produced better performance, especially in location questions. This relationship between scanning, deep processing, and performance was qualified by reading comprehension skills in an unexpected way: Scanning led to lower performance especially for good comprehenders, while the positive effect of deep processing was independent of reading comprehension skills. We discussed the results in light of our current knowledge of Internet problem solving.

Mehrebenenanalysen mit hierarchisch-linearen Modellen

Zusammenfassung. Dieser Beitrag gibt eine kurze Einfuhrung in hierarchisch-lineare Modelle mit Zufallskoeffizienten, eine Klasse von Verfahren zur Auswertung von Daten mit hierarchisch geschachtelten Analyseebenen (z.B. Individuen in sozialen Gruppen, Messzeitpunkte in Personen). Ausgehend von den Unterschieden zu herkommlichen linearen Modellen mit festen Effekten erlautern wir das Grundprinzip hierarchisch-linearer Modelle an einem Beispiel. Dann skizzieren wir das Vorgehen bei Parameterschatzung und Hypothesentestung, die Bedeutung unterschiedlicher Zentrierungsoptionen sowie Anwendungsvoraussetzungen hierarchisch-linearer Modelle.

Strategic Decisions in Task-Oriented Reading

Answering questions from texts are assessment and instructional activities that are frequently used in schools. Nevertheless, little is known about the strategic processes that students take while performing these tasks. We explored the amount and frequency that students initially read of a text before they answered questions pertaining to the material. In a procedure similar to the one used in the PISA (Program for International Students Assessment), one-hundred-seventy students between 7th and 9th grade read and answered several questions designed to assess task-oriented reading in three specific texts. We recorded on-line indexes that evaluated student behavior (e.g., the amount of text that students read before answering questions raised within a given text), performance, and comprehension skill. The results revealed that students skilled in comprehension initially read a high proportion of the texts, which in turn improved their overall performance in two of the three texts read (text 1: CI95%: 0.01 to 0.09; text 2: CI95%: -0.01 to 0.05; text 3: CI95%: 0.04 to 0.20). Therefore, we conclude that this strategic behavior should be considered during the assessment and instruction of reading literacy.

Computer Literacy, computerbezogene Einstellungen und Computernutzung bei männlichen und weiblichen Studierenden.

In einer Reihe von Studien haben sich Unterschiede zwischen mannlichen und weiblichen Studierenden hinsichtlich Computer Literacy-Variablen, computerbezogenen Einstellungen und tatsachlicher Computernutzung gezeigt. Die Befunde sind jedoch teilweise inkonsistent und die resultierenden Effektgrosen variieren. In dieser Untersuchung wurden 451 Studierende wirtschaftswissenschaftlicher Facher mit dem Inventar zur Computerbildung (INCOBI) befragt, das Computer Literacy-Variablen, computerbezogene Einstellungen und Aspekte der Computer- und Internetnutzung erfasst. Fur alle Variablen ergaben sich Geschlechtsunterschiede, wenn auch in unterschiedlicher Hohe. Manner wiesen umfassenderes theoretisches und praktisches Computerwissen, eine hohere Vertrautheit mit dem Computer, mehr Sicherheit im Umgang mit dem Computer und positivere Einstellungen auf als Frauen und nutzten den Computer bereits langer und intensiver. Die Geschlechtsunterschiede in den psychologischen Variablen liesen sich nur partiell auf die Dauer der bisherigen Computernutzung zuruckfuhren. Sowohl bei Mannern als auch bei Frauen ergaben sich selbst bei Berucksichtigung der Dauer der bisherigen Computernutzung Zusammenhange zwischen psychologischen Variablen und Masen der aktuellen Computernutzung. Die Geschlechtsunterschiede in der aktuellen Computernutzung konnten nur unvollstandig durch Sicherheit im Umgang und computerbezogene Einstellungen erklart werden. Die Ergebnisse werden im Hinblick auf ihre praktischen Implikationen diskutiert.