DASP provides you an implementation in Stata.

http://dasp.ecn.ulaval.ca/dmodules/m_md_inequality.htm

Also, you should look at the OPHI, e.g.,

https://ophi.org.uk/multidimensional-inequality-measurement/

I am not entirely sure what you want to do, but if the five levels are an independent variable, then you do not want to begin by separating the data according to level. Instead, you should calculate the index on the total sample, and then determine where there are differences across levels.

Dear Mr. David L Morgan

Thank you for your response. Basically I want to determine the differences betweent 5 level. Your answer gave an idea about how handling the index. But I do not know how to determine the difference according to these 5 categories after calculating the index.

You could use Analysis of Variance with a post-test comparison (Turkey test).

Dear Mahmuh Khaddi

Far from being a specialist on the subject. I would add to what José-Ignacio Antón and David L Morgan have said, an alternative way of calculating inequality in education among the various possibilities of doing so. Assuming that their statistical inputs were census data, the basis of which contains demographic information, together with education variables, among others.

You could choose to define a minimum age value - as an assumption - in order to concentrate a large part of the population that no longer attends an educational establishment. Generate a census base query on the maximum level of education attained in that population - for example, less than complete primary level, maximum complete primary level, maximum incomplete secondary level, maximum complete secondary level, etc. –. Then weight each of these education level categories covering all years of schooling passed by the population with a value between zero and one.

Then multiply the relative weight of each population group by this weighting. And finally add the values of the partial indices corresponding to the categories of maximum level achieved. The result is a single value for the total of that population that will be between zero and one - closer to one the higher the average educational level, and closer to zero when the majority of the population has few years of schooling. In addition, you can calculate the variability values for each level of education or between these categories, in order to compare the results with those of another population.

I am sorry if I misunderstood the query or if the answer I wanted to give was unclear. I remain at your disposal. Regards

Fernando (from Argentina)

Estimado Mahmuh

Lejos de ser un especialista en el tema. Agregaría a lo expresado por José-Ignacio y David, una alternativa de cálculo de la desigualdad en la educación entre las diversas posibilidades de realizarlo. Suponiendo que sus insumos estadísticos fueran datos censales, en cuya base contiene información demográfica, junto a variable de educación, entre otras.

Podría optar por definir un valor de edad mínimo–como supuesto-, para concentrar a gran parte de la población que ya no asiste a un establecimiento educativo. Generar la consulta en la base censal, sobre nivel educativo máximo alcanzado en dicha población – por ejemplo, inferior a nivel primario completo, nivel máximo primario completo, nivel máximo secundario incompleto, nivel máximo secundario completo, etc. –. Luego ponderar cada una de esas categorías de nivel educativo que abarca todos los años de estudios aprobados por la población, con un valor entre cero y uno.

Seguidamente multiplicar el peso relativo de cada grupo poblacional por ese ponderador. Y finalmente sumar los valores de los índices parciales correspondiente a las categorías de nivel máximo alcanzado. El resultado es un único valor para el total de esa población que estará entre cero y uno – más cerca de uno cuanto mayor sea nivel educativo en promedio, y acercándose a cero, cuando la mayoría de la población posee pocos años de estudios aprobados–. Adicionalmente puede calcular los valores de variabilidad para cada nivel de educación o entre las dichas categorías, con el objetivo de comparar los resultados con los de otra población.

Perdón si no comprendí bien la consulta o la respuesta que quise expresar fue poco clara. Quedo a su disposición. Saludos

Fernando (desde argentina)

Dear Fernando Ariel Manzano

Thank you so much for all the effort and your very detailed answer. Actually, the aproach you've proposed is exactly what I'm looking for. Could be tricky but I think It could fit income variable instead of age. Because I want to measure inequality using income values. Even so, chosing income make me so desperate to formulazation.