Neoliberal policies have changed the recruitment practices in higher education institutes. A decline in full-time and tenured jobs in the academy is evident as market and competition-oriented neoliberal ideas have penetrated higher education. In Turkey, 28.6% of academic staff are research assistants (Council of Higher Education, 2021), most of whom are temporarily employed and experience job insecurity (Durmaz, 2017). In the Turkish higher education system, research assistants are regarded as supportive staff responsible for assisting research and other duties, which are not explicitly stated in the Turkish Higher Education Law. This law introduces two distinct positions in research assistantship: 33/a, which secures a permanent job for research assistants; 50/d, which terminates research assistants’ contracts after obtaining their doctorate. Therefore, these two types of assistantship positions differentiate mainly in job security.

The existing literature on research assistants in Turkish universities is limited in scope and focuses mainly on revealing current problems and how differences in conditions of employment influence research assistants’ attitudes and emotions. Thus far, previous studies have reported that research assistants employed in 50/d positions are more likely to experience higher job insecurity than the research assistants from different statutes (Karcıoğlu & Balkaya, 2018), which negatively influences their attitudes toward the job (Altay & Tekin Epik, 2016), results in future anxiety, low productivity, and a feeling of exclusion (Durmaz, 2017), and also leads to lower job satisfaction and organizational commitment (Arı, 2007; Şengül-Avşar & Barış-Pekmezci, 2019; Tekin & Birincioğlu, 2017). Furthermore, the research assistants with 50/d cadre have been found concerned that their terms and conditions of employment pave the way for favoritism and mobbing in the workplace (Aydın & Özeren, 2019; Çolak, 2015). In an investigation into the significance level of research assistants’ work-life problems in terms of appointment type, Uysal, Ertuna, and Anıl (2015) found that the three topmost problems of research assistants with 50/d positions are the lack of guaranteed position, mobbing, and financial problems, respectively, while research assistants with 33/a positions complained of the heavy workload, financial problems, and mobbing at most. However, previous studies have failed to address how academics develop a collective attitude towards increasing job insecurity in higher education.

When employees believe that their voices are not heard through organizational channels, they may choose to perform activism in social media to “exert pressure on organizations to change policies, practices, or conditions” (Smith, 2005, p. 5). Similarly, research assistants employed in 50/d positions frequently attempt to become a trending topic on Twitter by sharing tweets using hashtags they determined to attract public and political attention to the issues related to their employment status. These tweets can reveal research assistants’ and other actors’ perspectives on job insecurity issues in the Turkish higher education system and the flow and directionality of relationships between actors that posted tweets or were mentioned in these messages. Therefore, this research aims to explore the themes in tweets posted under hashtags related to the 50/d research assistantship cadre and discover the patterns of Twitter users’ activities under these hashtags. This study aimed to address the following research questions:

1. What are the common topics of tweets sent about 50/d research assistantship?
2. How are 50/d research assistants linked together and other actors in Twitter?

This study uses Twitter data to explore the topics of the tweets shared under hashtags related to 50/d assistantship position by thematic analysis (TA) and reveal the meaning of the ties between actors sharing messages or are mentioned under these hashtags by social network analysis (SNA). Twitter data was mined using the academictwitteR package of R (Barrie & Ho, 2021) through a query including 16 hashtags, such as #Akademide50dSorunu (#50dProblemInAcademy) and #50d33aOlsun (Let50dTurnInto33a), that have frequently been used by 50/d research assistants to become trending topic in Twitter. The query with predetermined hashtags captured more than 48 thousand tweets posted in the last three years. Because it was observed during data screening that the data set included unrelated spam tweets, duplicate tweets, non-Turkish tweets, tweets that comprised of multiple hashtags but without other content, data were cleaned before analyzing the data using the dplyr package (Wickham et al., 2019) in R. The resulting data set included about 40 thousand tweets. Following Marwick’s (2014) framework for qualitative analysis of social media textual data, two thousand of the tweets (5% of the main corpus) were randomly sampled because the amount of data was not manageable to perform qualitative TA (Braun & Clarke, 2006). Initially, all sampled tweets will be read to familiarize with the data. Secondly, initial codes will be created inductively through MAXQDA. Later, codes will be analyzed to develop encompassing themes. Finally, created themes will be reviewed for consistency and relevance. Then, another rater will apply the themes to a randomly selected 20% of the data used in the TA. Inter-rater agreement will be assessed, and the researcher and the rater will reach a consensus about the themes. SNA will be conducted using the visNetwork package of R (Almende et al., 2019) to explore the actors found under hashtags about 50/d research assistantship and the ties between these actors. The term centralization in SNA for this study refers to a Twitter user’s influence in the network that can be assessed by in-degree and out-degree influence. In-degree influence is related to how many times the other users mention a particular Twitter user. In contrast, out-degree influence is related to the number of users a Twitter user mentions. This study uses centralization measures to explore the nature of the network among the actors. To visualize the network, Fruchterman and Reingold layout algorithm will be used.

Initial data screening showed that 50/d research assistants actively use Twitter to make their voice heard by labor union leaders, politicians, popular journalists, government agencies, and members of the higher education council. In the social network, there are 3470 nodes, with 11040 unique links among them. The links among research assistants in the social network are few in number when compared to the links between research assistants and union leaders and politicians. The vast majority of connections between RA and other actors in the network are unidirectional (from RA to the other actors). Even though many direct links exist between RA and politicians, it was observed that union leaders are key nodes in the network since they serve as a bridge between these actors. The central nodes include two union leaders, the president of Turkey, cabinet members and other members of executive power, and some anonymous Twitter accounts that were created presumably by some 50/d RA to disseminate tweets about their problems and demands. A quick look into textual data suggests that tweets posted under 50/d-related hashtags include proposed solutions toward academic insecurity resulting from different kinds of assistantship, information about 50/d RA problems and feelings, demands of 50/d RA from other actors. The results of this study provide insights into how various higher education stakeholders perceive academics' job insecurity and how academics take action towards their perceived job insecurity in a global academic labor market that has experienced radical changes in traditional academic positions due to pervasive neoliberal policies. Therefore, this study could be used by policymakers and university administrators to improve academics' working conditions and resolve security problems.

Altay, S. ve Tekin Epik, M. (2016). Perceptions of job insecurity and effects on business attitudes of 50 D research assistants. Atatürk University Journal of Economics & Administrative Sciences, 30(5), 1273-1287. Arı, A. (2007). Üniversite öğretim elemanlarının sorunları. Kırgızistan Türkiye Manas Üniversitesi Sosyal Bilimler Dergisi, 17, 66-74. Aydın, E. & Özeren, E. (2019). The phenomenon of work alienation in academia: a qualitative field study on research assistants. International Journal of Economic and Administrative Studies, BOR Special Issue, 159-178. https://doi.org/10.18092/ulikidince.518296 Barrie, C., & Ho, J. C. T. (2021). academictwitteR: an R package to access the Twitter Academic Research Product Track v2 API endpoint. Journal of Open Source Software, 6(62), 3272. https://doi.org/10.21105/joss.03272 Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77–101. https://doi.org/10.1191/1478088706qp063oa Council of Higher Education. (2021). Yüksekögretim bilgi yönetim sistemi. Ögretim elemanı istatistikleri: Uyruga göre ünvan bazlı akademisyen sayıları (Higher education information management system. Academic staff statistics: The titled-based number of academic staff by nationality. Retrieved from https://istatistik.yok.gov.tr/ Çolak, E. (2016). Precarious work in academy: Research assistants’ experiences. ViraVerita E-Dergi, 2 , 23-44 . https://dergipark.org.tr/en/pub/viraverita/issue/22434/240077 Durmaz, N. (2017). The prekariat in academy: The case of research assistants working under 50/d contract. Ankara Üniversitesi SBF Dergisi, 72(4) , 945-975 . https://doi.org/10.1501/SBFder_0000002474 Karcıoğlu, F., & Balkaya, E. (2018). Relationship between union membership, union satisfaction and job insecurity: An application on research assistants. Journal of Social Policy Conferences, 75, 307-326. Marwick, A. E. (2014). Ethnographic and qualitative research on Twitter. In Weller, K., Bruns, A., Burgess, J., Mahrt, M., & Puschmann, C. (Eds.) Twitter and society, (Vol. 89, pp. 109–121). Peter Lang. Uysal, İ. , Anıl, D. & Ertuna, L. (2016). Scaling of research assistants' professional problems in Turkey with paired-wise comparison method. Journal of Measurement and Evaluation in Education and Psychology , 6 (2) , 279-292. https://doi.org/10.21031/epod.14618 Smith, M. F. (2005). Activism. In R. E. Heath (Ed.), Encyclopedia of public relations (pp. 5–9). Sage. Şengül Avşar, A., & Barış Pekmezci, F. (2020). Research assistants’ views on the transition to uniform staffing in research assistantship. Marmara Üniversitesi Atatürk Eğitim Fakültesi Eğitim Bilimleri Dergisi, 51, 50-71. https://doi.org/10.15285/maruaebd.655075 Tekin, E. & Birincioğlu, N. (2017). Examining the level of organizational commitment of research assistants in terms of their type of staff and demographic characteristics. Uluslararası İktisadi ve İdari İncelemeler Dergisi, 19, 171-196. https://doi.org/10.18092/ulikidince.277858