Intelligent Transportation Systems (ITSs) and their applications are attracting significant attention in research and industry. ITSs make use of various sensing and communication technologies to assist transportation authorities and vehicle drivers in making informative decisions and provide leisure and safe driving experience. Data collection and dispersion are of utmost importance for the proper operation of ITSs applications. Numerous standards, architectures and communication protocols have been anticipated for ITSs applications. In recent years, crowdsourcing methods have shown to provide important benefits to ITSs, where ubiquitous citizens, acting as mobile human sensors, help respond to signals and providing real-time information. In this paper, the problem of mitigating crowdsourced data bias and malicious activity is addressed, when no auxiliary information is available at the individual level, as a prerequisite for achieving better quality data output. To achieve this goal, an innovative algorithm is designed and tested on a crowdsourcing database of lane change evaluations. A three-month crowdsourcing campaign is performed with 70 participants, resulting in a large number of lane changes evaluations. The proposed algorithm can negate the noisy ground-truth of crowdsourced data and improve the overall quality.