Cognition and Communication in Longitudinal Datacollection

The psychology of autobiographical memory examines the units, structures and processes that lie beneath the long-term storage of autobiographical information and, on this basis, explains memory phenomena such as remembering, forgetting, misremembering and misdating. Within this framework, autobiographical remembering is a task, that depends on a sequence of cognitive processes of encoding, representation and adaptive re-construction of autobiographical information.

Empirically, the quality of this task varies with the length of the retrospective interval: facts and events from the distant past are more likely to be forgotten, misremembered or incorrectly dated. This effect, however, is moderated by factors associated with the person, the information to be remembered and the recall situation - sometimes beyond recognition and in a way that we can relate to the cognitive, motivational and social-communicative functions and conditions of autobiographical remembering.

When collecting life history data as retrospective self reports, we are dealing with the products of autobiographical memory processes. There is, therefore, widespread concern, that the amount of forgetting, misremembering and misdating will increase with the length of the retrospective interval and, above all, that the reports will systematically deviate from reality. Since no suitable external validation criteria are available for most life history surveys, these concerns are not easily alleviated.

The GLHS Data do offer two attractive empirical paths for an assessment of whether and how much the regularities of autobiographical memory become manifest in the retrospectively collected and dated event sequences:

  • More than 1300 participants from the East German Life History study were re-interviewed approximately five years after the first survey; the reference periods overlap for 1.5 to 2.5 years. By comparing the reports of the same people for the same reference period, we can estimate the reliability or temporal consistency as an indicator for validity
  • About 80% of the participants of the West German Life History Study (cohorts 1964 and 1971) agreed to have their reports linked to officially collected German social security data on an individual level. By comparing retrospective self reports with process produced data from the Beschäftigtenstatistik, we can assess the validity of the life history data with an external validation criterion.

Moreover, the tape-recorded interviews and the complete documentation of the data-editing process allows an in-depth analysis of where an inconsistency has come about in the data generation process. Thus, we have the opportunity to look at memory errors as measurement errors within a framework of 'Total Survey Error', that various error sources produce during the process of data generation, to disentangle these error sources and to estimate their relative contribution to the Total Survey Error.

Project Team

Maike Reimer

Further Information