This Article Covers:
Understanding the assessment report
Overview: A comprehensive view of the assessment report with an explanation of each component and how each of it becomes instrumental in making hiring decisions.
Understanding the assessment report: the report consists of eleven sections that delve deep into the test taker’s skills, namely
- Absolute scorecard
- Summary card
- Performance insights
- Quality analysis
- Plagiarism check
- Detailed and expanded view of each solution
- Deep code analysis
- Technologies used in submission
- Proctoring analysis
- Additional information
- Social footprint
Let's begin by understanding each one:
- Absolute score: a quick snapshot on seven decisive parameters to base a high-level decision.
- Time taken for completion: This indicates the total time taken by a candidate to complete the test.
- Overall score: The most noticeable component is the overall score that a candidate receives. This is usually the best indicator of how well a candidate performed on the given tasks. You should be able to gauge a candidate's performance over a certain cut-off.
- Test verdict: To check whether a candidate was able to pass by meeting the overall cut-off percentage. Use this to shortlist or filter out candidates quickly.
- The number of problems attempted: This will help you understand the number of problems attempted, against the total number of questions assigned in the test.
A detailed explanation of it can be found in the ‘Quality analysis’ section of the report.
- Code plagiarism: It simply flags the similarity of code written by two different individuals and recommends a review for any code plagiarism detected by the engine. Detailed analysis of this can be found in the ‘Plagiarism’ section of the report.
- Proctoring engine: Once you enable proctoring for your test, this section will flag it as ‘Suspicious’ or ‘Not suspicious’. Suspicious if the candidate has indulged in any malpractice behavior. More details can be found in the ‘Proctoring’ section, which provides detailed activity logs and flagged images where suspicion is more likely.
- Performance summary: Divides the solutions into three statuses. Ie ‘solution accepted’ when the solution fully meets the criteria. ‘Solution partially accepted’ when only part of it is right and lastly ‘Solution rejected’ if it fails to meet the solution criteria.
- Performance insights: This section will give a birds-eye view of a candidate’s skills, their areas of strength and weaknesses on three levels namely Beginner, Intermediate and Expert. Having this data allows recruiters to make talent decisions more objectively, by structuring the interview and later process, based on the insights gathered.
- Quality analysis: This graph will highlight the types of code quality violations during the test, on seven stated parameters such as security, clarity, bug risk, complexity, duplication, performance, and style.
- Plagiarism: The evaluation engine flags the problems where it suspects any instances of plagiarism, where two different candidates have identical code submissions. However, it is at the discretion of the evaluator to review the highlighted code and make the decision accordingly. (Plagiarism detection is applicable for coding type problems only).
- Solutions: A listed view of candidate's solutions that will enable the recruiter to understand the score breakup per problem and the problem's status of acceptance. To check the actual submission, just click on any one of the problems.
- Deep code analysis: it is an in-depth evaluation of a candidate’s coding skills. This engine at DoSelect analyses the code for detecting patterns and characteristics pertaining to how ‘maintainable’ the code is. DCA is applicable in the coding problem-type only and is currently available for programming languages like Python2, Python3, Java7, Java8 and C.
There are three parameters to evaluate candidates on the deep code analysis:
- Coding approach: Analysis of complex control flow structures used in the code, which affects code’s readability and susceptibility to errors.
- Code Modularity: Analysis of modular design patterns used in the code, like high-order abstractions and conciseness.
- Code Extensibility: Analysis of how easy it is to add new functionality to the code without requiring major changes.
- Technologies used in submission: Most tests that are created on DoSelect allow you to solve a problem in multiple programming languages. A candidate can choose from the assigned programming languages while attempting a language-based problem.
- Proctoring analysis: A deeper insight can be collected using DoSelect's proctoring engine about candidates' usage behavior during the test. The metrics like the number of browsers used to give the test and the number of times a user was away from the test window, will hint at the possibility of malpractices.
- Additional Information: Gather additional information/credentials about the candidate, essential for the later evaluation process.
- Social footprint: Assess the candidate's technology proficiency based on contributions made on Github, Stackoverflow & DoSelect
Conclusion: DoSelect provides a comprehensive and data-driven analysis that provides insights into skills, areas of strengths, and improvements, thereby enabling recruiters to move forward conclusively. It’s also important to understand that while the report quantifies the scores objectively, the recruiter needs to interpret qualitatively at some places to finally arrive at a hiring decision.