Resilience of pre-vocational secondary school students and university students during COVID-19: a qualitative study

Purpose This study aims to get insight into how pre-vocational secondary school and university students experienced resilience during the pandemic and what risk and protective factors could be targeted during crisis. Methods ● Data Collection - Type of data: qualitative data. - Dates collected: 2020-2022. - Method: secondary analysis on qualitative interviews. - Location: Rotterdam, the Netherlands. - Number of participants: 18. ● Participants There were two groups of students studied.

The first group consisted of 8 participants who were pre-vocational secondary school students. Most of the pre-vocational students’ parents held low-income jobs. They were on average 16 years old (range 12-24).

More than half had parents with a migrant background (63%), although most were born in the Netherlands, and less than half were female (38%). The second group were 10 university students. Most of these participants had parents with high-income jobs, the participants themselves often had side-jobs, and some lived alone and independently.

They were on average 24 years old (range 21-29), and most were born in the Netherlands (80%). Half of the participants had parents with a migrant background or were born abroad (50%), and more than half were female (60%). ● Interview Procedure For the first dataset, interviews were conducted in person and online. The interviews were 30-45 minutes long.

They were conducted in the language of choice of the pre-vocational students, which was mostly Dutch and for some English. Parents of the pre-vocational students signed an informed consent form, and verbal informed consent was asked of the pre-vocational students before the interviews commenced. The interview guide was semi-structured.

Questions about high school and going to university were asked. For anonymity we used audio recordings only, then transcribed them using pseudonyms to identify the transcripts. Any names or places mentioned in the transcripts were edited or changed.

No data saturation was sought, as the main aim of these interviews was to explore the experiences and perceptions of pre-vocational education students. No member check, i.e. checking the transcribed text with the interviewees, was performed due to the lockdown immediately after the interviews. For the second dataset, interviews were conducted online for approximately an hour per interview.

The interview guide included the topics COVID-19, preventive behavior and lifestyle. Data saturation was reached around interview 7, after which three more interviews were conducted. A member check was done after transcription of the interviews for credibility of the data.

For both datasets, questions have been posed open-endedly and neutrally to avoid bias as much as possible. Interviews were audio recorded and transcribed, and pseudonyms were used for anonymity. The reason why two different protocols are being described is the fact that these two datasets originated from two different studies.

However, to get insight into the disruptions during the pandemic, cross-over discussions between the researchers resulted in a post-hoc decision to compare the datasets instead of keeping them as distinct studies. Comparability is ensured by using a social constructivist epistemology with a thematic analysis approach. This ensures that the same deductive themes are being used in both datasets with a focus on their perceived resilience, protective and risk factors. ● Ethics The participants gave consent for their data to be used for scientific purposes.

Ethics approval has been obtained for this study from the department of Ethics of the Erasmus School of Social and Behavioural Sciences under reference number #ETH2324-0333. ● Anonymization Names of participants were anonymized, then the anonymized names were replaced by "Participant 1", "Participant 2" etc. for the scientific publication. ● Analysis All qualitative interviews were analyzed by means of a thematic analysis using Atlas.ti (v23).

Author JK created a codebook based on the topics resilience, COVID-19, and health. This was used to code the interviews deductively with pre-existing codes. Moreover, open coding was used to further develop the codebook based on the participants’ answers, coding the interviews inductively as well.

This was used, because the COVID-19 situation was a novel situation, which might have led to themes that could not be predetermined in the codebook. Before coding all the interviews in this way, two coders coded three interviews independently (JK and MH). After each interview was coded, an ICA (intercoder agreement) was calculated.

This is a measure used to calculate how similarly the coders interpret and use the codes to ensure validity and reproducibility. A discussion of the coding process ensued until consensus was reached. After the third interview was coded by both authors, a satisfactory ICA was reached of 0.64 (for the first dataset) and 0.52 (for the second dataset).

The range of an ICA can be between -1 and 1, so these ICAs provide substantial intercoder agreement for secondary data analyses as such. Author JK proceeded to code the rest of the interviews by using the deductive codes of the codebook and adding new inductive codes based on the interview responses. To further enhance reliability, the coding process was thoroughly discussed with the research team.

The coded interviews were then analyzed per theme. The researchers looked at the components of resilience, namely sense of mastery, sense of relatedness and emotional reactivity. They also looked at codes describing changes due to COVID-19, mental and physical health, and protective and risk factors.

These codes were studied per participant and captured in an Excel file to create an overview and see where possible patterns varied between participants’ background characteristics and between the two groups. Data Structure / Contents This repository contains the qualitative code tree used during the analysis of the interview data, as well as a README file describing the study design, data collection procedures, and analytic approach.

Value of the Data for Potential Reuse The shared materials provide transparency regarding the qualitative analytic process and coding procedures used in this study. They may support methodological understanding, facilitate reproducibility of the analytic approach, and assist other researchers conducting qualitative interview research on related topics. Conclusion Resilience seemed to be more process- and context-dependent rather than education or age-dependent.

Policy makers should focus on creating supportive environments and systems to foster youth resilience. ● Limitations One of the limitations of this study is that the interview guides included questions about different themes, as they were conducted in different studies. However, using a clear code book and consulting with researchers that were experienced in resilience and behavioral theories, we were still able to effectively code the interviews and extract the information that we needed for this study.

Another limitation is that the interviews of the two datasets took place at different times during the pandemic. However, they still took place during the pandemic around lockdowns and with restrictions, which provided the sought-after insight into how youth experienced the pandemic and how resilient they were during this time.