Imputation of missing categorical data by maximizing internal consistency

This paper suggests a method to supplant missing categorical data by reasonable replacements. These replacements will maximize the consistency of the completed data as measured by Guttman's squared correlation ratio. The text outlines a solution of the optimization problem, describes relationships with the relevant psychometric theory, and studies some properties of the method in detail. The main result is that the average correlation should be at least 0.50 before the method becomes practical. At that point, the technique gives reasonable results up to 10–15% missing data.We thank Anneke Bloemhoff of NIPG-TNO for compiling and making the Dutch Life Style Survey data available to use, and Chantal Houée and Thérèse Bardaine, IUT, Vannes, France, exchange students under the COMETT program of the EC, for computational assistance. We also thank Donald Rubin, the Editors and several anonymous reviewers for constructive suggestions.     

Slow alpha in the EEG power spectrum as an indicator for conceptual arousal]

Based on previous findings (B?sel et al., 1990) it was assumed that in concept learning tasks generating on hypotheses on a concept which has to be developed is accompanied by increases of the Alpha 1 power (7.5-10 Hz) in the spontaneous EEG activity. In this study 16 subjects performed five problem solving tasks with similar processing requirements. EEG data were analyzed by means of post hoc comparisons of subjects differing in performance quality. Additionally, four control tasks were employed in which, based on previous studies, variations in the Theta frequency range were expected. An effect in the Alpha 1 frequency band was observed in tasks requiring reconstructive recall or testing the usefulness of an mathematical algorithm. The creation of a rank order or mental map is accompanied by power increases in the lower portions of the Alpha 1 frequency band (7.5-8.5 Hz). Moreover a high amount of controlled variance (eta2 up to 34%) was obtained for this effect. Increases in EEG Theta power, which presumably indicate subjects' component analysis, were found before the subjects recognized parts of geometric figures or before relevant features in the "buddhist monk problem" were discriminated. The dynamics of EEG power over time is in examples of frequency/time plots in a figure, illustrated.     

EMG reactions of mimetic muscles after repeated presentation of pictures with positive, negative and neutral content]

Twenty male subjects (Ss) were repeatedly confronted for five seconds to pictures with positive, negative and neutral valence. During the picture presentation facial EMG-reactions of five mimic muscles (m. frontalis lateralis, m. corrugator supercilii, m. orbicularis oculi, m. zygomaticus major on the right and left side of the face) were recorded. In addition, heart rate was measured. It could be shown that during repeated presentation of pictures with positive valence the m. orbicularis oculi and the m. zygomaticus major on both sides of the face yielded enhanced average EMG-reactions as compared to the repeated presentation of pictures with negative and neutral valence. During presentation of pictures with negative valence the m. frontalis lateralis and the m. corrugator supercilii revealed enhanced EMG-reactions as compared to the repeated presentation of pictures with positive valence. Furthermore, it could be shown that during the repeated presentation of pictures with positive valence those facial muscles which mainly react during confrontation with pictures of positive valence showed a continuous decrease of EMG-activity. Facial muscles which mainly react during the presentation of pictures with negative valence do not show a continuous decrease in EMG-activity during the course of the repeated presentation of pictures with negative valence. Additionally, the EMG-activity of the m. zygomaticus major on the right and lift side of the face do not differ with respect to the mean activity and the time course during confrontation of pictures with positive valence. The m. orbicularis oculi yielded enhanced EMG-reactions during the presentation of pictures with positive valence as compared to both m. zygomaticus major muscles. The present results are discussed with respect to psychobiological theories concerning facial expression.     

Discrimination learning of rotated faces: a Markov analysis of coding and association processes]

30 subjects participated in a discrimination experiment learning face-letter associations under four rotation conditions (45 degrees, 90 degrees, 135 degrees, 180 degrees). Under each condition two thirds of the faces were presented twice, upright and rotated away from the vertical; the remaining faces were presented once, upright or rotated. Learning is described by a joint Markov model: For faces that are presented twice it assumes a separate association and encoding process (two-stage-model), for faces that are presented once it assumes an association process (all-or-none-model). The Markov model fits the data for all four rotation conditions. The angle of rotation does not affect learning for faces that are presented once. For faces that are presented twice it influences both the association and the encoding process. For the angles employed, the effect of rotation can be approximated linearly. The results suggest that the encoding of a rotated face differs increasingly from an upright face as a function of these angles of rotation. This confirms analogous conclusions from mental rotation experiments.     

Sentence encoding and implicitly activated memories

Some words have fewer direct associates than others, and, when words varying in set size are studied in a list-learning task, those with smaller sets are more likely to be recalled. This set-size effect is found in cued recall when the words are studied in the absence of related words, but not when studied in the presence of related words. Related words provide context and theoretically inhibit irrelevant associates. The present research determined that set-size effects are found when words are encoded in sentence contexts. In contrast to list-learning experiments, the results of three experiments found such effects even when lexically related words were present in the sentences. Other findings indicated that target-set-size effects were determined by the proximity of related words in the sentence and the nature of the test cue. The results are discussed in relation to a model for explaining set-size effects and to selective findings from the sentence-comprehension literature.     

The role of syllabic and orthographic properties of letter cues in solving word fragments

The present research examined the role of phonological and orthographic properties of cues in mediating the retrieval of words from the mental lexicon. The task required subjects to resolve fragmented words when provided with semantically related cues (e.g., spiteful:---DIC----). Phonological properties of the letter cues were manipulated such that the letters either corresponded to the syllables (e.g., DIC in vindictive) or nonsyllables (NDI) in the word. Orthographic properties of the letter cues were manipulated by selecting letter groups that either co-occurred frequently in the language or did not. In two experiments, results revealed little or no effect of the phonological variable (syllables) but a reliable effect of the orthographic variable (letter-cue frequency). Letter cues with a low frequency of co-occurrence in the language led to better completion of the fragmented words. We interpret these findings as support for models of lexical representation that are based on orthographic properties (e.g., Seidenberg & McClelland, 1989) rather than those based on phonological constraints.     

A psychophysiological study of the use of partial information in stimulus-response translation.

Ss performed a hybrid go/no-go reaction task in which colored letters were assigned in various ways to 4 finger responses, 2 on each hand. In addition to reaction time, psychophysiological measures were used to assess the duration of stimulus identification and the onset of central and peripheral motor activity. The results suggest that response selection can begin on the basis of 1 stimulus dimension (e.g., color), while the other dimension (e.g., letter form) has not yet been identified. Other results are discussed with regard to "selection for action" (Allport, 1987) and the importance of stimulus-response translation strategies in the use of partial information.     

Rubbing your stomach while tapping your fingers: interference between motor planning and semantic judgments.

Previous research (Klatzky et al., 1989) shows that the time required to make sensible/nonsensible judgments about an action-object phrase (e.g., "rub your stomach") is reliably faster when the phrase is preceded by a cure representing a specific prototypical hand shape (vs. a neutral cue). The current experiments investigated the effects of preparing for an alternate task (finger tapping vs. syllable vocalization) on facilitatory priming of sensibility judgments. Preparation for finger tapping reduced the magnitude of the priming effect more than preparation for vocalization, suggesting that resources accessed during semantic processing of action-object phrases are also used during manual response preparation. The results support the existence of a system representing manual actions that is limited in the number of activities that can be represented at one time and that is not so general that it represents manual and vocal tract movements.     

Temporary suppression of visual processing in an RSVP task: an attentional blink? .

Through rapid serial visual presentation (RSVP), we asked Ss to identify a partially specified letter (target) and then to detect the presence or absence of a fully specified letter (probe). Whereas targets are accurately identified, probes are poorly detected when they are presented during a 270-ms interval beginning 180 ms after the target. Probes presented immediately after the target or later in the RSVP stream are accurately detected. This temporary reduction in probe detection was not found in conditions in which a brief blank interval followed the target or Ss were not required to identify the target. The data suggest that the presentation of stimuli after the target but before target-identification processes are complete produces interference at a letter-recognition stage. This interference may cause the temporary suppression of visual attention mechanisms observed in the present study.