|Home | About | Journals | Submit | Contact Us | Français|
Developing and testing the cognitive skills and abstract thinking of undergraduate medical students are the main objectives of problem based learning. Modified Essay Questions (MEQ) and Multiple Choice Questions (MCQ) may both be designed to test these skills. The objectives of this study were to assess the effectiveness of both forms of questions in testing the different levels of the cognitive skills of undergraduate medical students and to detect any item writing flaws in the questions.
A total of 50 MEQs and 50 MCQs were evaluated. These questions were chosen randomly from various examinations given to different batches of undergraduate medical students taking course MED 411–412 at the Department of Medicine, Qassim University from the years 2005 to 2009. The effectiveness of the questions was determined by two assessors and was defined by the question’s ability to measure higher cognitive skills, as determined by modified Bloom’s taxonomy, and its quality as determined by the presence of item writing flaws. ‘SPSS15’ and ‘Medcalc’ programs were used to tabulate and analyze the data.
The percentage of questions testing the level III (problem solving) cognitive skills of the students was 40% for MEQs and 60% for the MCQs; the remaining questions merely assessed the recall and comprehension. No significant difference was found between MEQ and MCQ in relation to the type of questions (recall; comprehension or problem solving x2 = 5.3, p = 0.07).The agreement between the two assessors was quite high in case of MCQ (kappa=0.609; SE 0.093; 95%CI 0.426 – 0.792) but lower in case of MEQ (kappa=0.195; SE 0.073; 95%CI 0.052 – 0.338). 16% of the MEQs and 12% of the MCQs had item writing flaws.
A well constructed MCQ is superior to MEQ in testing the higher cognitive skills of undergraduate medical students in a problem based learning setup. Constructing an MEQ for assessing the cognitive skills of a student is not a simple task and is more frequently associated with item writing flaws.
The evaluation of the competence of undergraduate medical students is a very critical task, as in the future, these ‘to be physicians’ have to cater with human lives. (1)
At undergraduate level there are three domains of skills to be evaluated i.e. Cognitive, Affective and Psychomotor. Cognitive domain can be evaluated (2) at different levels including Knowledge, Comprehension, Application, Analysis, Synthesis, and Evaluation. Modified bloom’s taxonomy (3) identified three levels of cognitive domain. In medical education, the major emphasis is on developing and evaluating the level III or problem solving skills, as most of the physician’s time is spent in analyzing patient’s problems.
Proper cognitive assessment tools reward the students for their higher cognitive skills and abstract thinking. (4) There are various methods to assess the knowledge domain which include Free response examinations (Long Essay Questions, Short answer Questions, Modified Essay questions), Multiple choice questions, Key feature questions, Self-assessment and peer-assessment. Each of these methods has its pros and cons and is addressed to assess different levels of bloom’s taxonomy. No single method of evaluation is superior to other and probably a reliable and valid evaluation requires a combination of these methods. (1, 4)
The Multiple choice questions are very popular in evaluation of undergraduate medical students. They are reliable and valid; moreover they are easy to administer to a large number of students. Well constructed MCQs have a greater ability to test knowledge and factual recall but they are less powerful in assessing the problem solving skills of the students. A large proportion of curriculum can be tested in a single sitting. The scoring is very easy and reliable using computer software, but the construction of good MCQs is difficult and needs expertise. Generally MCQs stimulate students to make a superficial and exam oriented study. (5, 6)
Modified essay questions are short clinical scenarios followed by series of questions with a structured format for scoring. They primarily assess the student’s factual recall but they also assess cognitive skills such as organization of knowledge, reasoning and problem solving. They also address the writing skills and even ethical, social and moral issues and attitudes. MEQs are more flexible and their value is somewhere between essay type questions and MCQ. But they need to be carefully constructed with provision of model answers and training to avoid inter-rater variability. (7, 8)
In a bird’s eye view, examination and evaluation are a source of anxiety and stress for the undergraduate medical students, but in reality it directs the students to study harder and improve their skills. Therefore, it is imperative not only to stimulate their cognitive skills during the teaching but also to frequently examine their higher mental and reasoning skills. (9, 10, 11)
Our objectives were
50 MCQs and 50 MEQs were chosen randomly from the written examinations delivered to fourth year medical students attending undergraduate internal medicine course at the Department of Medicine, Qassim University. In the 50 MEQs, there were 104 stems. Questions were analyzed individually by two independent assessors as per the ‘preformed criteria’ to label a question with Level I, II or III of the cognitive domain. Questions belonged to various final and midterm examination from 2005 to 2009.
No ethical approval was required as no human subjects were involved in the research
Each question was analyzed separately to
The analysis was done by each assessor as per the Performa shown in table (1).
Software ‘SPSS 15’ was used for data entry, analysis and interpretation. Kappa test was used to determine the agreement between the two assessors using ‘Med Calc’ software.
Total of 50 questions with 104 stems were included. Analysis is shown in table (2).
Total of 50 questions were evaluated. Analysis is shown in table (2).
Questions were representing different disciplines of Medicine (graph 1). No significant difference was found between MEQ and MCQ in relation to the type of questions (recall; comprehension or problem solving (x2 = 5.3, p = 0.07).
The analysis of item writing flaws is shown in table (3).
The results of this study show that Multiple choice question is a better test of cognitive skills than the Modified essay question as 60% of the MCQs tested the higher cognitive skills and only 40% of MEQs addressed the cognitive level III of modified Bloom’s taxonomy.
Although when the proportions of the two forms of questions addressing the different levels of cognitive domain were compared, there was no significant statistical difference among the level of cognition tested, as p-value was 0.07. It is also inferred that constructing an MEQ might be technically more difficult than an MCQ as item-writing flaws in MEQs were 16% as compared to MCQ only 12%.
The results of our study are consistent with Edward JP et al 12 who also found MCQ to be superior in testing the level 3 of cognitive domain. The results do not coincide with those of Irwin WG et al13 who found MEQ to be superior in testing the highest level cognitive skills but they did not choose the questions randomly rather they compared the whole examination papers of various years. They suggest that both MCQ and MEQ may be designed beforehand to test any particular level of Bloom’s taxonomy. Construction of MEQ needs expertise and training, model answers of such questions need meticulous consideration. (14, 15)
It also seems that some specialty examiners tend to put more stress on Level III testing or it is easier to design such questions for some specialties as most of level III questions in both MCQ and MEQ belonged to cardiology.
It is quite understandable that conduction of a proper assessment is not only dependent on the cognitive aspect of question, but there are so many factors which play a role like, reliability, content and construction validity, financial and human resources.
At the department of Medicine all the questions are constructed by the faculty members then these are put forward to a ‘question review committee’ and after its approval the questions are introduced into the question bank. It was observed during the period of the study that the content and the construction of the questions improved over the years but still individual errors occurred.
The two confounders that played a major role in deciding the level of cognitive domain addressed by each specific type of question were (i) The type of examination to which the question belonged i.e. Midterm or Final (ii) The specialty of the question like cardiology, neurology etc. It may be inferred that in final exams the examiners tend to put the type of questions that address the highest level of cognitive domain (level III) as 80 % of MCQs belonged to the final exams while all MEQs were taken from the Midterm exam papers. This may be one drawback of our study. The total number of questions from which the sample was taken was quite limited which accounts for the small sample size in this study but as more and more questions will be incorporated into the question bank of the department further validation studies will be required.
MCQs were found to be testing the level III of cognitive domain more frequently than MEQs. Training in formulating MCQs and MEQs, more so for MEQs is needed to assure achieving level III of cognitive domain and avoid item writing flaws.