The purpose of this project is to analyze content from temporary job postings. Overall research questions: what are the most common words and topics that appear across temporary job postings? What might this tell us about the employer's "ideal" temporary worker?
This study analyzes the frequency of words and topics within a sample of temporary job postings to understand how employers frame and promote the "ideal" temporary worker. This project will contribute to the literature on mechanisms of workplace control. Questions of workplace control have been central to studies in the sociology of work. Control scholars ask: How do employers extract labor from workers? How do they get them to consent to work, to being controlled on the job? How do they get them to work hard?.
Burawoy (1979) and Edwards (1980) are two of the most influential theorists in this area. Burawoy (1979) argued that employers controlled workers through getting them to "play the game" of "making out" or trying to achieve higher piece-rates. Edwards (1980) argued control took place through the structure of the workplace, either through direct supervision of workers, allowing technology to pace work (e.g. an assembly line), or through bureaucratic rules and rewards. Both theorists argue control takes place on the job, through the labor process.
The data obtained for this project will help to build theory on workplace control occurring prior to work on the job - that is before hiring begins during the advertising and recruiting phase. This area is under-theorized but audit studies and interviews with employers tell us that employers not only try to obtain worker consent to control on the job but during the hiring process. Employers actively seek the "ideal" worker who they think will perform diligently and with little resistance often discriminating on the basis of race, gender, and criminal record (Pager 2003; Waldinger and Lichter 2003).
By using text-mining techniques I can analyze the most frequent words and topics within the job posting data. I will better be able to understand how employers frame the "ideal" temporary worker and thus attempt to control on certain qualities before workers step foot on the job or for an interview. Based on my knowledge of temporary and other forms of unstable work I expect to find that "flexibility" will be an important topic. Additionally, I expect the temporary job postings to describe certain qualities that an "ideal" temporary work might appreciate (or enjoy) about temporary work (perhaps an opportunity to learn, autonomy, "flexible" hours).
I plan to collect data from the University of California, Davis Human Resources Department. The HR department will pull their posting data from 2016, 2017, and 2018 for all temporary, contract, and limited positions and provide me with an excel or .csv file containing all the text data that appears on a job posting.
At this point I do not have plans to have a test set. I am not sure how many cases the UC Davis HR dataset will contain. However, after receiving the first data file from UC Davis HR I may be able to obtain data from years prior to 2016.
All files are on my PC and backed-up weekly to google drive. This github repository mirrors the structure of the files and provides information on cleaning steps taken. In this repository you will find:
- Data management plan file (explains how data are to be organized, stored and shared)
- raw_data folder (will contain link to the storage of the directly downloaded data with no cleaning, contains information on cleaning steps undertaken)
- analyze_data folder (scripts for analysis of the cleaned data and links to the resulting processed dataset will be made available)