Negative and normal self-schemata : content differences and processing influence

Social cognition suggests that self-schemata (cognitive structures of self-knowledge) guide the processing of personal information and facilitate recognition of self relevant evidence. Likewise, negative self-schemata are thought to bias recognition toward negative personal information in cognitive depression and performance anxiety. Research aims were to specify how negative and normal self-schemata differ, and to examine self-schema processing influence. Self-schemata constitute an active, conceptually-driven process which enhances the discriminability of favoured evidence. Bias can also arise from a passive, stimulusdriven process due to lower recognition thresholds for favoured evidence. A yes/no decision task was used for investigating schemata by determining sensitivity and response latency for personal descriptor content. Experiments 1 to 3 compared negative and normal schematics. Descriptors varied in pleasantness and relevance to personal competence. Self-ratings showed that negative schematics had superior discrimination for competence whereas normal schematics had superior discrimination for pleasantness. Negative self-percept reflected incompetence, not unpleasantness. Normal characterization reflected marked self-enhancement. Role-reversal ratings maintained the sensitivity differences observed for self-ratings. Response latencies consistently indicated effects resulting from stimulus driven processes and not schema-driven influence. Experiments 4 and 5 compared subjects with or without a self-schema for a normal personality dimension. Self-rating discrimination of test descriptors was greater for schematics yet this superiority vanished for vignette-rating which dissociated self-schema influence from discriminative ability. Response latencies largely indicated effects resulting from stimulus-driven processes but a slight residue of schema-driven influence was also obtained for self-judgements. Experiment 6 used a word naming task and presented variously degraded personal and impersonal descriptors. Schema-relevant items were named more rapidly and accurately than impersonal and schema-irrelevant items. Schema-driven processing was the more probable source of the effect. Choice decision latency data were reanalysed to determine how a word relevance bias related to a word frequency bias. A cross-over interaction resulted for self-rating schematic subjects whereby the normal frequency effect reversed for self-irrelevant words. Despite considering various processing models, this outcome defied explanation. Reversal of the frequency effect occurs in recognition memory performance and similar processes were possibly implicated in self-rating. Major findings were firstly that competence information has more salience for negative schematics. Secondly, judgement latencies show little schema-driven influence. Finally, a self-relevance bias probably results from schema-driven processing in which case, processing of relevant evidence is highly accurate.