Racial riots left a deep scar in this region. Four decades ago these events took place more frequently. Though less frequent , they crop up on many occasions whenever vested quarter adds fuel to situation getting worse. This week I took a look at the racial riots in the South East Asia and Bangladesh and tried to do a correspondence analysis between type of racial riots and 5 countries of the region. I was more interested in Bangladesh, Singapore, Malaysia,Indonesia and Myanmar. I willfully discarded the events in Indian Subcontinents as there are no short of such events and to make the analysis concise.
For this analysis, I picked up 12 such events for Bangladesh, 6 events each for Myanmar and Singapore, 3 for Malaysia and 2 for Singapore.Citations or documentations of the events are the criteria for picking up these events. There are many such riots however I picked up the ones mentioned in Wikipedia. They all occurred between 1950 and 2017.
I further split the riots into three types: political, military and racial. Political events are those riots influenced or caused by political decision or as a fallout of a policy or judgment. Politically motivated riots fall in this category. Riots occurred in military administered area or shaped and influenced by military are dubbed as Military. Racial riots are those happened as a consequence to an accident or mob violence.
Correspondence analysis captures the similarities between the categories of the variables and association between the variables. In the present case, country is the the row variable ranked in terms of strength of the economy and Singapore tops the rank followed by Malaysia. Riot is the column variable.
I am more interested to see which type of riot is more tied to each of five countries.
Unfortunately , Pearson’s chi-squared test has a significance level P = 0.1475 that requires to accept the null hypothesis that there is no association between row and column variables. So there is no point to go beyond this. The following discussion is useless. However, this discussion could continue as a matter of interesting exercise . As correspondence analysis intends to present low dimensional representation describing how the row and column categories contribute to the inertia in a table , two dimensions----dimension1 86.06% and dimension2 13.04% ---appear to account for 100% of the inertia.
The biplot reveals that horizontal dimension differentiates between regimented group backed riot and accidental or mob led riot. Since orientations from origin to Military and from origin to Myanmar are so close , it can lead to say that regimented group backed riots are more frequent in Myanmar. By the same token, it can be said that politically motivated riots are more common in Bangladesh.
Please note that Pearson test of independence finds no association between country and type of riots, making our above discussion redundant.
For this analysis, I picked up 12 such events for Bangladesh, 6 events each for Myanmar and Singapore, 3 for Malaysia and 2 for Singapore.Citations or documentations of the events are the criteria for picking up these events. There are many such riots however I picked up the ones mentioned in Wikipedia. They all occurred between 1950 and 2017.
I further split the riots into three types: political, military and racial. Political events are those riots influenced or caused by political decision or as a fallout of a policy or judgment. Politically motivated riots fall in this category. Riots occurred in military administered area or shaped and influenced by military are dubbed as Military. Racial riots are those happened as a consequence to an accident or mob violence.
Correspondence analysis captures the similarities between the categories of the variables and association between the variables. In the present case, country is the the row variable ranked in terms of strength of the economy and Singapore tops the rank followed by Malaysia. Riot is the column variable.
I am more interested to see which type of riot is more tied to each of five countries.
Unfortunately , Pearson’s chi-squared test has a significance level P = 0.1475 that requires to accept the null hypothesis that there is no association between row and column variables. So there is no point to go beyond this. The following discussion is useless. However, this discussion could continue as a matter of interesting exercise . As correspondence analysis intends to present low dimensional representation describing how the row and column categories contribute to the inertia in a table , two dimensions----dimension1 86.06% and dimension2 13.04% ---appear to account for 100% of the inertia.
The biplot reveals that horizontal dimension differentiates between regimented group backed riot and accidental or mob led riot. Since orientations from origin to Military and from origin to Myanmar are so close , it can lead to say that regimented group backed riots are more frequent in Myanmar. By the same token, it can be said that politically motivated riots are more common in Bangladesh.
Please note that Pearson test of independence finds no association between country and type of riots, making our above discussion redundant.
No comments:
Post a Comment