Transcription and analysis

The analysis was conducted based on the six thematic analysis phases of Braun and Clarke [33], which included: (i) familiarizing yourself with the data, (ii) generating initial codes, (iii) searching for themes, reviewing themes, (v) defining and naming themes and (vi) reproducing the report [15]. This analysis was undertook by Four different authors. First, the recorded audios were transcribed verbatim and double-checked with the hand-written notes (R.Y. and J.Y.). After reading the transcriptions multiple times to promote detailed familiarity with their contents, primary coding was conducted using thematic analysis by NVivo 11. (W.L. and Z.J.). The initial coding summarized “psychological stress”, and codes were gathered together under umbrella codes such as “impact from the mass media”, “self-abasement caused by changes in appearance”, to ease navigation of the hundreds of codes. In the second phase of coding, researchers discussed and made decisions about which codes were the most important and useful by comparing and synthesizing the initial codes. The code were collated to generate potential themes and double-checked with the original data. Three themes were identified in this process and were reviewed by other authors (J.W. and J.M.). Finally, the whole Shanghai project team came together to discuss and decide on the final set of themes, including resolving any discrepancies identified in the initial rounds of thematic analysis. When all the themes, subthemes, and nodes were determined, they were discussed and reviewed by all authors to ensure the accuracy and were used to form the final themes [33]. As it is very essential to guarantee the accuracy of the participants’ quotes when translated from Mandarin to English, a researcher (C.H. and X.M.) who spoke both Mandarin and English checked the translation several times to ensure the greatest possible equivalence between the translation and the original text.