Assessment Methodology

1. The CEFR Framework

Our language proficiency assessments are strictly aligned with the Common European Framework of Reference for Languages (CEFR). The CEFR is the internationally recognized standard for describing language ability. It describes language proficiency on a six-point scale, from A1 for beginners, up to C2 for those who have mastered a language.

• Basic User (A1, A2): Can understand basic phrases and communicate in simple, routine tasks.
• Independent User (B1, B2): Can understand the main ideas of complex text and interact with a degree of fluency.
• Proficient User (C1, C2): Can understand a wide range of demanding, longer texts, and express ideas fluently and spontaneously.

Reference: Council of Europe (2001). Common European Framework of Reference for Languages: Learning, teaching, assessment. Cambridge University Press.

2. Adaptive Testing Architecture

Cubbbix employs a computerized fixed-item assessment approach designed to simulate Computerized Adaptive Testing (CAT) principles. By presenting a calibrated mixture of lexical, grammatical, and semantic items across varying difficulties, the assessment triangulates the user's proficiency level.

The assessment measures:

• Lexical Repertoire: Breadth and depth of vocabulary knowledge.
• Grammatical Accuracy: Understanding of syntax, morphology, and structural rules.
• Semantic Comprehension: Ability to infer meaning from context and understand nuances.

Reference: Wainer, H. (2000). Computerized Adaptive Testing: A Primer (2nd ed.). Lawrence Erlbaum Associates.

3. Question Generation and Calibration

The items in our assessment banks are developed by linguistic experts and refined through empirical data. Each item is tagged with a specific CEFR target difficulty (A1 through C2).

We use Item Response Theory (IRT) frameworks in the design phase to ensure that each question has a high discrimination index—meaning a C1-level question is highly likely to be answered correctly by a C1 learner and incorrectly by a B1 learner.

Reference: Lord, F. M. (1980). Applications of Item Response Theory to Practical Testing Problems. Routledge.

4. Scoring and Evaluation Mechanism

The final evaluation is not a simple percentage score. Instead, the algorithm analyzes the distribution of correct answers across the difficulty tiers. A user must demonstrate consistent mastery of lower-tier items (A1/A2) to be credited with higher-tier (B2/C1) capabilities.

For example, if a user correctly answers several C1 questions but fails numerous A2 questions, the algorithm penalizes the final CEFR rating to account for foundational gaps or guessing. This creates a highly accurate, normalized estimation of true language competence.

5. Typing Speed & Keystroke Assessment Methodology

The Cubbbix Typing Speed Assessment Engine evaluates raw speed, net precision, and character accuracy using microsecond-precision browser event hooks (performance.now()).

Key calculation principles include:

• Gross Words Per Minute (WPM): Calculated as (Total Characters Typed ÷ 5) ÷ Test Duration (Minutes). Standardizing 5 characters per word controls for word-length variance across different text passages.
• Net Words Per Minute (Net WPM): Evaluates true error-free speed by penalizing uncorrected errors: Gross WPM - (Uncorrected Errors ÷ Test Duration in Minutes).
• Anti-Cheat & Test Reliability: The timing engine monitors inter-keystroke arrival intervals (latency variance) and blocks synthetic paste/macro operations, ensuring every issued certificate reflects genuine human typing performance.
• Skill Classification Matrix: Scores map to standardized industry tiers: Novice (<35 WPM), Intermediate (35–49 WPM), Professional (50–69 WPM), Elite Typist (70–89 WPM), and Master Typist (90+ WPM).

Reference: Yamada, H. (1980). A Historical Study of Typewriters and Typing Methods: From the Position of Planning Japanese Parallel Processors. Journal of Information Processing.

6. Disclaimer