Text Analysis of Puberty Experiences From Retrospective and Concurrent Perspectives

• In Event: Computational Text Analysis Methods for Developmental Research

Abstract

Background: Recent popularity of shows like Netflix’s Big Mouth, which follows a group of 7th graders navigating puberty and middle school, highlight public interest in the experience of puberty. Popular media outlets like The Guardian and Vox have praised the show for being a “sharp, surprisingly joyful look at the gross time that is puberty” and for being a “sex education comedy” (Bernstein, 2018; Framke, 2017). However, the show is based on adults’ retrospective narratives and often includes mature depiction of dating and sex during puberty that may not accurately reflect the concurrent lived experiences of youth. Accordingly, these shows may be misleading to both youth and adults in their representation of the experience of puberty. This is important because the experiences and self-perceptions that individuals have during puberty and when recollecting about puberty may have important psychological consequences (e.g., Moore, McKone, & Mendle, 2016).

Objective: The present study leverages data from youth writing concurrently about puberty and data from Big Mouth in order to compare puberty experiences from retrospective and concurrent perspectives. In particular, the present study focused on how concurrent and retrospective accounts compared in terms of physical and social changes salient to pubertal development.

Methods: The concurrent dataset was comprised of textual data generated by 181 youth participants (87% female, M age= 11.73 years) who engaged in four consecutive days of journal-style expressive writing focused on puberty. The retrospective dataset was comprised of all available Big Mouth scripts (N = 30 episodes) over the course of three seasons released from 2017-2019. We used a range of computational text analysis tools to compare the two datasets, including heatmaps and word embeddings.

Results: The relationships between multiple keywords salient to puberty were represented in a heat map of the expressive writing vector space and the Big Mouth vector space. Notably, “puberty” was related to feeling words like “different”, “hate”, “stupid”, “hurt” in the Big Mouth vector space, but not other feeling words. However, “puberty” was related to a range of feeling words (from “sucks” to “normal”) in the expressive writing vector space. In addition, the closest words to salient indicators of physical changes were examined with cosine similarity. Results highlighted a difference in how physical change words were referenced between the corpora. In particular, the contrast in nearest words to “breasts” between the two corpora demonstrate a tendency toward objectifying female physical changes in the Big Mouth corpus as compared to the expressive writing corpus (e.g., Figure 1).

Conclusion:
The retrospective perspective of Big Mouth does not consistently represent concurrent experiences of puberty despite media praise of the show’s portrayal. Big Mouth tended to characterize puberty in limited, negative ways as compared to youth concurrently experiencing puberty. Further, Big Mouth’s representations of female related pubertal changes may reinforce trends of objectification and peer sexual harassment during puberty. These results highlight the potential for analyzing representation of developmental phenomena with text analysis methods.

Authors

• Mary Kate Koch, University of Florida
  Presenting Author

• Jane E Mendle, Cornell University
  Non-Presenting Author