How do we recall the story of our lives? Evidence for a temporal order in the recall of important life story events

ABSTRACT

Based on the chronological nature of extensive life narratives, in two studies, we investigated whether a temporal order can also be found for the retrieval of important life story events. The data of Study 1 come from 52 older adults (M_age = 70.35, SD_age = 6.09) who reported seven important events from their life. We used multilevel analysis to take into account the hierarchical structure of the data but also to investigate individual differences. Results showed a significant temporal order effect which considerably differed between persons. In further exploratory analyses using group-based modelling, we identified three groups that differed in the direction (i.e., forward and backward) and in the size of the temporal order effect. The use of life scripts was not associated with a forward temporal order. Study 2 was conceptualised as a replication study and included 88 participants (M_age = 68.60, SD_age = 6.63) who recalled up to 15 important life events. The temporal order effect, as well as the three trajectories of temporarily ordering important life story events, were replicated. Again, the use of life script was not associated with a forward ordered recall. Potential explanations for a forward and backward temporal recall order are discussed.

KEYWORDS:

• Autobiographical memory
• life script
• life story
• temporal order
• multilevel analysis

Acknowledgments

We thank Tilmann Habermas and Annette Bohn for sharing their coding system for life script and life story event categories.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Note that the data were analysed as if they stemmed from independent observations, although each participant generated seven life scripts, which leads to dependent data.

2 Model 3 showed the best fit according to AIC and BIC. Model 4 showed the best fit according to –2LL and sample size adjusted BIC, but the differences were small, which is why we preferred the more parsimonious Model 3.

3 An anonymous reviewer suggested providing additional analysis to test the association between the use of the life script and a forward ordered recall. For this reason, we ran Model 1 again but only included scripted life story events. If the use of the life script was associated with a forward ordered recall, we would expect the effect of position to be more pronounced, respectively, to explain a higher amount of variance of age-at-event. However, results remained basically the same; that is, the effect of position did not change. Moreover, the amount of variance explained by position was 24.50%. Compared to the 26.50% of variance that was accounted for in the analysis, including all life story events, the difference seems negligible.

4 Two raters scored all life scripts for event categories, agreeing on 90.1% of all events. Disagreements were resolved by the first author. There was a high degree of overlap in the German and the Danish life script events: of our 28 categories, 21 events overlapped with categories from the Danish life script collected by Bohn (2010) (see Appendix, Table A4). As opposed to the Danish life script, the German life script included the categories “professionalism”, “professional success/career”, “relationship”, “disappointment”, “mobbing/discrimination”, “buy apartment/house”, “health (overcome serious diseases)” and “sports/hobbies”. In sum, the differences are minor and only refer to life script events that are low in frequency. Moreover, nine of the ten most frequently reported scripted life story events are identical across Study 1 and Study 2. This indicates that in both studies, life story events are mostly categorized according to the same life script events.

5 Model 3 showed the best fit according to BIC and Entropy. Model 4 showed the best fit according to –2LL and AIC. However, the differences were small. Moreover, compared to the model that included three groups, the explained variance of the model that included four groups improved only slightly. For this reason, we preferred the more parsimonious Model 3.

6 Again, we ran Model 1 for the scripted life story events only to test the association between the use of the life script and a forward ordered recall. Similar to Study 1, the effect of position remained mainly the same, and a comparable amount of variance was explained by position (19.40% vs. 24.30%).

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.