A Sentiment Analysis of the PhD Experience Evidenced on Twitter
This article explores the content of PhD student tweets. It has three main aims: (a) to examine what is discussed regarding the PhD process, (b) to evaluate whether tweets express primarily positive or negative sentiments, and (c) to uncover the key themes discussed in both positive and negative tweets.
Recent surveys of PhD students have raised concerns about their wellbeing by pointing to high prevalence rates of adverse mental health conditions. However, our understanding of which factors pose the highest risks is still evolving. Self-selection into surveys also raises the possibility of discounting positive aspects of the PhD experience. We use a different data source (Twitter) to explore both these issues.
Using 16,928 tweets with the Twitter hashtags #phdlife and #phdchat, we first conduct dictionary-based sentiment analysis in R to determine whether tweets are dominated by positive or negative sentiment. We then hand-code the dominant sentiment of a randomly selected subset of 1,994 tweets and qualitatively analyse positive and negative tweets separately to uncover the key themes in each category.
This article contributes to the emerging literature on the wellbeing and mental health of PhD students by using a novel data source (tweets). It highlights both positive and negative aspects of the PhD student experience.
We find that most tweets express positive rather than negative sentiment, indicating that PhD students do enjoy many aspects of their experience. Negative tweets are dominated by mental health concerns. They also highlight problems with academic culture (especially the normalization of overwork) and the effects of the pandemic on students.
Our results indicate that there is a need to change the academic culture of normalizing overwork, ensure adequate institutional provision of mental health support and ability to spot signs of emotional distress, devise strategies to combat the imposter phenomenon, and respond to the particular challenges that the pandemic has created for PhD students.
The authors recommend that future research explore the specific challenges and opportunities faced by PhD students in different disciplines and geographical locations. As the data used here were collected during the pandemic, it would be useful to track post-pandemic sentiments to observe changes.
Improving the graduate experience of PhD students and providing them adequate mental health support will help to ensure their continued productivity and wellbeing.
Future research in this area should focus on the efficacy of different interventions to address key problems, such as the imposter phenomenon, stress, anxiety, depression, and isolation.