The reason given for reservations is that in the current socioeconomic milieu, different categories of people face different challenges in obtaining education and employment. The social and economic obstacles are hypothesized to be so large that, even if education assistance is substituted for reservations, the impact would not be sufficient to ensure sustainable overall equality.
Multiple Index Related Affirmative Action (MIRAA)
One question that arises often is, why only caste? It would appear that the optimal way to ensure equality would be to use a basket of indicators including caste, gender, economic status etc. One such basket, named Multiple Index Related Affirmative Action (MIRAA), has been suggested by Prof. Purushottam Agrawal. The argument for using caste alone is that caste is the biggest indicator of underdevelopment. Indices such as MIRAA could certainly be more effective in improving the condition of people than caste alone, since they would allow reservations to be effectively targeted at the people who need them the most.
To understand MIRAA, we first consider the basket of socioeconomic indicators it suggests. The primary considerations when choosing such indicators should be as follows:
- The indicators should be indicative of education and employment levels
- Information on the indicators should be readily available
- Caste/Tribe
- Gender
- Economic status
- Kind of schooling received
- Region where candidate spent formative years
- Educational status of parents/family
Debating MIRAA
This system is already being debated on Tehelka. Praise for the system includes the fact that is is proven: Jawaharlal Nehru University has used it successfully in the past. The system is a self-organizing score in the sense that it targets the right sections of society in a manageable way. The need to target the right people is mentioned by several readers on the Tehelka debate. The system also appears to balance the needs of the group and the rights of the individual. Put another way, 70% of the final admission score is "merit" based.
Criticisms include one from Amit Sen Gupta, another commentator on Tehelka, who believes that:
Targeting of affirmative action to a section within “backward” castes will be used as a powerful tool to deny the benefits to as many as possibleand that the system would be a non-starter on a nation-wide scale. Some readers on Tehelka also expressed concerns about the exact weights given to various indicators.
The points raised by Mr. Gupta bear thinking about. One strength of MIRAA is that it is a single transparently computable score. This is good for scalability. MIRAA also does not explicitly target a section within backward castes. It targets those who are suffering the most; as an implicit consequence, it will target backward castes. Within backward castes, it would target specific sections, but this is still implicit. The system as a whole remains simple, based on a single score, and thus not prone to overly high levels of manipulation.
The unstated but most contentious issue is likely the low overall weight given to caste/tribe - just 5 points out of 30, or in the bigger picture, just 5% of the total candidate score. Resolving this bone of contention is crucial; most of the difficulties with MIRAA are likely to be about the relative weighting of the indices and about the total percentage of the MIRAA score included in the total candidate score.
Objections by other readers serve to strengthen this assertion. The exact weights used to compute the score here were selected based on the individual reasoning, personal experience, or personal preferences of a person or some persons. The 30% number was arrived at the same way.
In a word, MIRAA as it stands today is a subjective system.
Making MIRAA Objective: the Modified MIRAA Score
Turning MIRAA into an objective system requires only a little tweaking of the system itself. It would, in addition, involve some survey sampling and statistical analysis.
To understand how to make MIRAA an objective system consider what we mean when we say that a person belonging to a certain category, say an SC candidate, is at a disadvantage compared to a forward caste (FC) candidate.
An objective way of defining the amount of disadvantage is the following. In an examination, suppose the average SC candidate scores 12% less than the average FC candidate. Then the SC candidate is at a disadvantage of 12 percentage points compared to the FC candidate. In the above situation, the SC candidate should get a MIRAA score of 12. This is the correct score because it neutralizes the real disadvantage the average SC candidate has relative to the FC candidate. It is objective because the score is completely data driven; personal opinions don't come into the picture. The data and methods used to establish the actual disadvantage would be a matter of public record.
The MIRAA set of indices can be used to refine the above further. For example, SC women may, on the average, score 18% less than FC men, while the difference for SC men may be 10%. SC women should get a MIRAA score of 18, while SC men should get a MIRAA score of 10.
The same system can be extended to include all 6 variables. In this modified MIRAA system, there is no artificial percentage attached to group needs, such as the 30% in the original MIRAA. Their modified MIRAA score is is simply added to their score in the entrance exam to determine the candidate's final score. This final score may add up to more than 100, but that is not a problem if rank (based on the final score) is used to determine admission.
The modified MIRAA system handles the balance of merit and group needs in a more correct way by restoring to each candidate exactly the amount of merit that he/she was deprived of by the socioeconomic system.
Potential Drawbacks of the Modified MIRAA Plan
This method appears to have a drawbacks as well.
First, it appears to reward poor performance. The strata that perform the worst would have the highest MIRAA score. Thus it could be argued that this system may actually encourage poor performance. This objection is not valid in reality however. A counterpoint is that, within each stratum, candidates are selected by fair competition according to merit. Thus there is a strong incentive for each candidate to perform higher, and those who perform lower within each stratum would fail to obtain seats.
The second objection is that there is no single standardized exam in India (analogous to the SAT in the USA) on which the difference in score between different strata could be evaluated. This objection can be resolved by using statistical methods (such as "grading on a curve") to normalize the scholastic achievements in different educational boards. Alternatively, if it is felt that there are fundamentally different categories of examinations and the score should be different in each, several categories of exams could be created, with a different table of modified MIRAA scores for each.
Discussion and Conclusions
The existing reservation system does not necessarily get resources to those who need them the most; however, some sort of assistance must be provided to those who have historically suffered from socioeconomic discrimination. Prof. Agrawal's suggestion of MIRAA takes 6 important indices, as well as merit, into account when computing a score, and is simple enough to be implemented transparently. If implemented, it would be instrumental in giving specific socioeconomic strata of people the assistance they need. However, MIRAA as it stands today faces some objections that can be traced back to the subjective origins of its scoring system.
A "Modified MIRAA" score is proposed that achieves the same objectives as the original MIRAA system but eliminates the subjectivity of the score, potentially increasing its acceptability. The Modified MIRAA is also perfectly fair: it compensates each socioeconomic stratum for exactly the loss in merit imposed by the socioeconomic system. As a consequence it also balances merit and socioeconomic status in a natural way. The price paid for the objectivity of the Modified MIRAA score is data collection and statistical analysis; however, this could also be done using simple and transparent protocols.
No comments:
Post a Comment