Employment information cocoon and anxiety amplification under algorithmic recommendatio

Abstract

In the social media ecosystem dominated by algorithmic recommendation technology, users are prone to form information cocoons due to long - term  exposure to homogeneous  employment  content, which may  exacerbate  the occupational anxiety of the college student group. Existing research mainly focuses on the behavioral impacts of information cocoons, lacking quantitative analysis of their psychological effects and the underlying mechanisms of algorithmic technology. This study aims to reveal how the algorithmic recommendation mechanism affects occupational  anxiety  by  constructing  an  employment  information  cocoon  and  to verify  the  moderating  effect  of  individual  cognitive  reflection  ability,  providing theoretical support for the precise intervention of occupational anxiety in the digital age. This study adopts a quantitative research method. Through stratified sampling, 420  undergraduate  students  from  three  universities  in  Guangzhou,  Guangdong Province, were selected as the sample. The Career Anxiety Scale and multi - level linear regression model were used for data analysis, and the Bootstrap method (with 1000 repeated samplings) was introduced to test the mediating effect. The study found that  the  intensity  of the  information  cocoon  has  a  significant  positive  impact  on occupational anxiety (Cohen ’ s d = 2.13, p < 0.001). For every 30 - day exposure in the information cocoon, the anxiety level increases by 0.23 standard deviations. The homogenization of recommended content affects occupational anxiety through two paths: directly enhancing the perception of occupational competition (B = 0.35) and indirectly reducing occupational self - efficacy (B = 0. 19), with the total mediating effect  accounting  for  58.7%.  Cognitive  reflection  ability  plays  a  moderating  role. High cognitive reflection ability can reduce the growth rate of anxiety (β = - 0.17, p = 0.032),  while  those  with  low  cognitive  reflection  ability  are  prone  to  fall  into  a negative cycle. The study confirms that the algorithm - driven information cocoon is a technical inducement for occupational anxiety. Based on this, it is recommended to formulate intervention strategies from two aspects: optimizing the recommendation logic and enhancing  individual  cognitive  reflection  ability,  providing  an  empirical  basis  for constructing an algorithm governance framework and a college student occupational mental health service system.