Finding the best job applicants for a job posting: A comparison of human resources search strategies

Summary

The research paper "Finding the best job applicants for a job posting: A comparison of human resources search strategies" by C.G. Harris, published in 2017, investigates the effectiveness of various digital and crowd-based approaches in identifying suitable candidates for job openings. The core objective of the study was to evaluate and compare three distinct strategies designed to match job applicants with specific job skill requirements. These strategies represent different facets of human resources technology and artificial intelligence in recruitment. The first approach involved utilizing crowdworkers within a gamified environment, which leverages human intelligence on a large scale, often incentivized through game-like mechanics, to perform tasks such as candidate evaluation or resume screening. The second strategy employed information retrieval (IR)-based search methods, which typically rely on algorithms to search, rank, and retrieve relevant documents (like resumes or profiles) based on keywords and structural similarities to job descriptions. The third method was a text-mining approach, designed to extract meaningful patterns and insights from unstructured text data (such as resumes), often incorporating features and elements from IR-based search engines to enhance its analytical capabilities. Harris's methodology involved a comparative analysis of these three strategies against a benchmark of job postings and applicant data to determine their efficacy in matching candidates to job descriptions. The study differentiated between technical and non-technical job postings, understanding that the nature of required skills and their assessment might vary significantly across these categories. The findings revealed notable differences in performance among the strategies depending on the job type. Specifically, the crowdsourcing environment demonstrated superior performance when identifying the best applicants for technical job postings. This suggests that for roles requiring specialized and often complex technical skills, the nuanced understanding and qualitative judgment provided by human crowdworkers, even in a gamified setting, offered a distinct advantage. In contrast, for non-technical job postings, the study indicated that both the crowdsourcing environment and the text-mining approach yielded equally effective results. This finding implies that for positions where skill matching might be less about intricate technical validation and more about identifying keywords, experience, and general competencies from textual data, automated text-mining tools can be as proficient as human crowd efforts. The implications of this research are significant for modern human resources departments, highlighting that a one-size-fits-all approach to applicant search strategies may not be optimal. Instead, a tailored strategy, potentially combining the strengths of different digital tools and human-in-the-loop systems, could enhance recruitment efficiency and candidate quality, depending on whether the job posting is technical or non-technical.