Abstract—Rater plays an important role in awarding fair judgment to students. However, the difficulty to consider fairness to the student applies, especially for the assessment of competency in design electronic circuit. Therefore, the use of an analytic scoring rubric as a guide can reduce the error due to the nature of rubrics. This present research employs Many Facet Rasch Measurement (MFRM) to explore rater error focusing on central tendency effect. Participants comprised of a sample of nine experienced teachers who were employed to assess 68 students in their competency of Electronic Circuit Design process in Vocational College in Malaysia. Students were observed using four-point analytic rating scale. The data were collected and analyzed using FACET, a MFRM computer software program. The results were presented in two ways: at the group level and at the individual level. At the group level, information from the scale category statistics indicated central tendency effect; however, none of the separation statistics indicated such an effect. At the individual level, there are two raters that exhibit centrality Index Terms—Central tendency, many facet Rasch measurement (MFRM) I. INTRODUCTION Practical work is one of the examples of performance based assessment that relies heavily on human judgment. However, this kind of assessment raises a variety of problems, especially in terms of validity and reliability since human judgment is subjective and uses a scale that is not dichotomized [1], [2]. In order to reduce human measurement errors and increase the validity and reliability to the decisions made in measuring students' abilities during practical work, a standard measurement process that includes a set of specific criteria and procedures that is valid and fair to all students should be conducted [3], [4]. An evaluation method that is claimed to be effective for fair judgment to students is by developing a standard rubric. Rubric refers to a scoring guide used to evaluate the quality of students’ constructed response. A rubric has four essential features: task description, scale, dimensions and descriptions of the dimensions. Task description is almost originally framed by the instructor and involves a performance of some sort by the student. The task can also apply to behavior, Manuscript received November 20, 2015; revised June 2, 2016. Azmanirah Ab Rahman and Nurfirdawati Muhamad Hanafi are with Universiti Tun Hussein Onn Malaysia, 86400 Parit Raja Batu Pahat Johor, Malaysia (e-mail: azmanira@uthmedu.my, nurfirda@uthmedu.my). Jamil Ahmad and Ruhizan Mohammad Yasin are with Universiti Kebangsaan Malaysia, 43650 Bangi Selangor Darul Ehsan, Malaysia (e-mail: jamil3191@yahoo.com, ruhizan@ukm.my). participation, use of proper lab protocols and behavioral expectations in the classroom. The scale explains how well or how poorly any given task has been performed. Terms used to describe the level of performance should be tactful and clear. The dimensions of the rubric present parts of the task simply and completely. Description of the dimensions means the quality definition represent of the quality of the performance for each rating scale [5]. Rating scale is a measurement instrument used by the rater to assigns ratees to positions along the continuum, denoting their relative ordering with respect to the trait being measured [6]. While rating scale provides information on the students’ performance, they are unfortunately subject to various sources of bias and error. Irrelevant factors can influence human judgment process, such as individual raters and certain tendencies that may be exhibited [6]. The four most common rater errors are severity/leniency, halo effect, central tendency effect and restriction of range [7]. This research is focused on Central Tendency Effect. II. CENTRAL TENDENCY EFFECT There are several definitions of central tendency noted by many researchers. Central tendency is defined as the tendency to interpret and apply the scoring scale for category idiosyncratic. Raters tend to use the mid-scale category without properly assessing student performance [8]. DeCottis defined central tendency as a rater’s unwillingness to go out on the proverbial limb in either direction, characterized by central ratings with little variability [9]. It is a special case of restriction of range [6], [7]. In the context of the MFRM, a rater is said to exhibit central tendency effect when he or she overuses the middle category, or middle categories of a rating scale while assigning fewer score at both the high and low ends of the scale [10]. The result is less variation in performance among students [11], [10]. According to [11], there are several factors that contribute to measurement error such as, 1) the rater failed to discriminate against students who were in a range between the lowest and highest. Therefore, the rater tends to leave marks on the mid-range 2) the examiner failed to distinguish between students' performance along the continuum, which does not understand the difference between each scale category. Thus, the examiner will allocate the same score to each student; 3) The rater did not have a strong background on the field or no training on the use of the rating scale during evaluation; 4) The examiner wants to be in a "play it safe strategy" for fear of being too soft or too lenient in awarding Investigating Central Tendency in Competency Assessment of Design Electronic Circuit: Analysis Using Many Facet Rasch Measurement (MFRM) Azmanirah Ab Rahman, Jamil Ahmad, Ruhizan Mohammad Yasin, and Nurfirdawati Muhamad Hanafi International Journal of Information and Education Technology, Vol. 7, No. 7, July 2017 525 doi: 10.18178/ijiet.2017.7.7.923