Document Type : Original Article
Authors
1 1.Medical Physics and Engineering Department, Medical School, Isfahan University of Medical Sciences, Isfahan, IRAN 2.School of Optometry and Visual Science, University of Waterloo, Waterloo, Canada.
2 3.Department of Physics, Babol University of Technology, Babol, Iran.
Abstract
Keywords
In most countries university admissions are based on the previous applicant performance. But in some countries such as India there is a special exam called GATE, or in China there is Goa Koa,the most competitive in the world. In Iran also, there are several competitive entrance exams (1,2).
This exam has been generalized for the admission of almost all of university branches . This area seems to have involved a vast majority of Iran's population, but there are not any distinct published research works in this area. The possible reason for this shortage could be the access restriction to the consensus and results of these exams.
Fortunately, in the case of this study applicant performance worksheets were available on the official site of Ministry of Health and Medical Education for the proposed year of study. The author of this report has been involved in one of the exam committees of the board of exam for almost twelve years. Therefore, it was realised that published researches in this area are few and it was worth investigating worksheets thoroughly.
One of the major issues in an exam could be the subject permutation in the exam sheet, or the taxonomy order(3-6). In competitive exams there are different effective factors which challenge applicant mentality. One of the mentality issues is the coverage of background study experience of the applicant with exam materials. Another issue could be the consistency of exam references all over the country. The other issue could be the board of exams' familiarization with new references. The tough competition and former issues result in big competitive market for private institutes. There are huge social consequences laid down behind it. The matter of either acceptance or rejection causes mental conflicts for the rejects and suspicious judgment follows.
One of the anxieties in the exam is the exam time plan for applicants. Irregular question taxonomy permutation could suppress moderate prepared applicant performance. Entangled taxonomy in all questions reduces exam result contrast for applicant discrimination. Therefore the exam output bend toward inconsistency. The unveiled procedure and analysis of the data could satisfy rejects and avoid their rumor for examiners and committees of the exam. In this study it is supposed to investigate the statistical parameters of applicants in some branches of study which have more common subjects and moderate number of attendance. It is proposed to finde inter correlation between different subjects in each branch of study. Also the discriminate factor of each subject is to be considered.
Methods
By the investigation of all 29 branches of study in the proposed Iranian educational year of study ( 86-87) in Ministry of Health and Higher Education, it was found out that three branches of study had more common subjects and applicant attendance. Targeted branches were bacteriology, parasitology and epidemiology. Mark sheets of all applicants in these three areas were downloaded and sorted for each subject (7,8). Subjects in bacteriology were bacteriology(M1), virology(M2), parasitology and mycology(M3), biochemistry(M4), molecular biology(M5) and immunology(M6). In parasitology, the subjects were cetology(M1), helminthology(M2), medical insectology(M3), hematology(M4), immunology(M5), bacteriology and mycology(M6) and biochemistry. Subjects in epidemiology were biostatistics(M1), epidemiology(M2), parasitology and mycology(M3),bacteriology(M4), virology(M5) and immunology. The max score level in all subjects was 100 in percentage. Statistical investigations were carried out through Minitab 15 statistical tool box(9).
Results
Most of the subject mark distributions are close to normal distribution therefore mean value and all of all of subject marks are listed in table 1.
Table 1: Statistical parameters of different subject scores in different branches of study
|
(M7) |
(M6) |
(M5) |
(M4) |
(M3) |
(M2) |
(M1) |
(score) |
|
|
------ |
42.8±27 |
18.3±16.7 |
41±23.5 |
27±19.1 |
41±22.3 |
31±17.7 |
146±71 |
Bacteri- |
|
39.5±23 |
25.7±10 |
32.7±33 |
33±20.3 |
62±17.2 |
43±17 |
53.7±21 |
290±96 |
Parasit- |
|
------- |
21.5±22 |
31±22 |
-0.4±8.4 |
19±14.2 |
29±15.8 |
57±20.5 |
218±92 |
Epidem- |
Total earned score of written exams are evaluated versus mark subjects through linear regression analyzer. Their equations are as follow.
Scores for bacteriology: - 0.00038 + 2.24 M1 + 0.450 M2 + 0.450 M3 + 0.450 M4 + 0.450 M5 + 0.450 M6
This equation was extracted from work sheets of 68 bacteriology candidates.
Scores for parasitology: 0.680 +1.79 M1 +1.81 M2 +0.592 M3 +0.605 M4 +0.591 M5 +0.574 M6 +0.605 M7
This equation was extracted from work sheets of 50 parasitology candidates and the following equation is extracted from 38 worksheets of epidemiology.
Score for epidemiology: - 0.00038 + 2.24 M1 + 0.450 M2 + 0.450 M3 + 0.450 M4 + 0.450 M5 + 0.450 M6
By observing the extracted equation for bacteriology it is obvious that the score weight of the first subject(M1) is five times of the other subjects in this branch of study. Therefore the most important and discriminative subject could be bacteriology in this branch. By looking at the regression equation of parasitology we will find out the weight score of the first(M1) and second(M2) subjects are triple times of other subjects in this branch of study. Therefore, these two subjects could have more impact than the other on the acceptance discrimination. Also in this respect, by taking a look at epidemiology equation it is realized that the first(M1) and second(M2) subjects are weighted twice the others. In comparison to other two former branches of study, in epidemiology the subjects are more balanced.
With the correlation tool the correlation between subjects and total score are extracted. Subject importance is extracted with respect to their P-Value.
Table 2: Pearson correlation coefficients of each subject with total score in bacteriology.
|
|
M1 |
M2 |
M3 |
M4 |
M5 |
M6 |
|
Corr(score) |
0.94 |
0.73 |
0.55 |
0.79 |
0.53 |
0.78 |
|
P-Value |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
Table 3: Linear regression of score versus single subjects.
|
Epidemiology |
Parasitology |
Bacteriology |
|
Score= -0.1+3.82M1 = 78.4+4.86M2 = 137+4.3M3 = 219+2.67M4 =131+2.8M5 =160+2.68M6 |
Score= 179+2.8M1 = 72.3+5.0M2 = 149+2.27M3 = 244+1.37M4 =219+2.16M5 =124+6.44M6 =178+2.8M7 |
Score= 28+3.79M1 = 50+2.33M2 = 90+2.1M3 = 47+2.4M4 =104+2.27M5 =61+2.0M6 |
Table 4: Pearson correlation coefficients of each subject with total score in parasitology.
|
|
M1 |
M2 |
M3 |
M4 |
M5 |
M6 |
M7 |
|
Corr(score) |
0.86 |
0.90 |
0.40 |
0.29 |
0.74 |
0.66 |
0.67 |
|
P-Value |
0.00 |
0.00 |
0.00 |
0.04 |
0.00 |
0.00 |
0.00 |
Table 5: Pearson correlation coefficients of each subject with total score in epidemiology.
|
|
M1 |
M2 |
M3 |
M4 |
M5 |
M6 |
|
Corr(score) |
0.86 |
0.84 |
0.66 |
0.25 |
0.68 |
0.64 |
|
P-Value |
0.00 |
0.00 |
0.00 |
0.14 |
0.00 |
0.00 |
Table 6: Pearson correlation M4 with other subjects in epidemiology
|
|
Total(score) |
M1 |
M2 |
M3 |
M5 |
M6 |
|
Corr(M4) |
0.25 |
0.14 |
0.40 |
0.08 |
0.12 |
0.27 |
|
P-Value |
0.14 |
0.39 |
0.81 |
0.63 |
0.47 |
0.10 |
Box plots are plotted for all the three exam branches. These box plots are showing decade distribution of different subjects against each other.
Discussion
As correlation coefficients of different subjects with total score in table 2 and table 3 and their statistical distribution in figure 1 show, exam subjects in bacteriology are reasonably balanced. Also small P-Values of correlation coefficients are in agreement with reasonable balancing. In bacteriology the correlations within subjects are comparably high with small P-Values. Therefore, there could be some information redundancy between different exam material subjects.
In parasitology as it is given in table 3 and shown in figure 2, the bacteriology and mycology subjects are(M6) highly weighted in one directional linear regression. Its coefficient is 6.44, which in comparison to other subjects is very large. If we look at its correlation with other subjects in parasitology as given in table 4, we will find that it is highly correlated with cetology(M1) and helminthology(M2) which have weighted by three. It could be concluded that the exam material of this subject should be repeated and redundant to these two subjects. As the overall score highly correlated with this subject(M6), there should be a close look at question paper or the material overlap in these subjects.
In epidemiology branch of study, bacteriology as given in table 1earned minimum regression coefficient value(-0.4) and lowest standard deviation. It means that this subject has the lowest contribution in total score and lowest discrimination impact. On the other hand it has the lowest correlation with other subjects in this branch as it is shown in figure 3 and given in tables 5 and 6. Therefore, this subject either is strange to applicants or its taxonomy is ill organized.
Figure 1:Box plot of different subjects in bacteriology
Figure2: Box plot of different subjects in parasitology
Figure 3: Box plot of different subjects in Epidemiology
Conclusion
Inter subjects overlap and redundancy could be guessed by correlation coefficient investigation between different subjects. The high redundancy between subjects could be effective and less discriminative. A subject is while discriminative which could enhance the overall score variation. In this study we found out that bacteriology and mycology in parasitology against its low score weight has earned the highest correlation with total score. Therefore it is either the core subject for this branch or wrongly designed in this exam. Also we found that the bacteriology subject had the lowest correlation with the total score and other subjects. Therefore this subject is either strange to applicants or its taxonomy is ill organized. In the case of unfamiliarity of applicants it should be omitted or in case of improper taxonomy, concerned board of exam should be informed. In the end it is suggested that for the generalization of this study, previous years and next year exams data be considered.
)