Where are brown people short?

In the comments to my post, Why brown people are midgets, a reader pointed me to this paper, which tabulates and analyzes some data from the 1960s for males. There isn’t anything too surprising in the data set; Punjabis are tall compared to non-Punjabis, higher castes are taller than lower castes. There is a lot of unaccounted for variation. This was before Indian Shining, and the Green Revolution probably hadn’t sunk in yet (I wouldn’t be surprised if the between-state differences increased, while the between-caste differences decreased, in the past 40 years). So appropriate caveats. The one surprising thing is that being a Muslim was a variable that predicted more height! This is correcting for regional distribution. The reason it’s surprising is that Muslims are genetically no different from non-Muslims (e.g., in Uttar Pradesh the highest proportion of recent West Asian ancestry is 5% across the whole population) and tend to socially disadvantaged in India (i.e., they are a Backward Class). The only explanation I have is that Muslims are less likely to be vegetarian or have dietary taboos which constrain their nutritional choices. Tables below the fold….

* p-value at 0.05
** p-value at 0.01

Table 4 – Anthnropometric Survey of North India: Male Height Least Squares Regression
Independent variables		Regression coefficient	t ratio
Intercept		63.38	983.2**
Age group:	20-40	0.37	6.6**
	41-50	0.19	3.1**
	51-60	-006	0.8
	Age factor chi-square		125.6**
Social group:	High caste	1.21	30.2**
	Superior sudras	0.93	20.4**
	Inferior sudras	0.24	4.8**
	Tribals	0.10	1.9
	Muslims	1.05	20.1**
	Others	0.43	5.9**
	Social group factor chi-square		1434.7**
State:	Assam	-0.52	8.2**
	Bihar	-0.41	9.0**
	Gujarat	0.11	2.1*
	Haryana	1.59	17.7**
	Himachal Pradesh	-0.41	5.5**
	Jammu and Kashmir	0.00	0.0
	Orissa	-0.46	9.6**
	Punjab	1.55	22.8**
	Uttar Pradesh	0.69	17.1**
	State factor chi-square		1793.2**
R2	0.107
s	2.37

Table 5 – Anthropometric Survey of South India: Male Height Least Squares Regression
Independent variables		Regression coefficient	t ratio
Intercept		63.50	122.8**
Age group:	20-40	0.18	0.4
	41-50	0.12	0.2
	51-60	0.19	0.3
			1.53
Social group:	High caste	0.87	9.6**
	Superior Sudras	1.08	16.6**
	Inferior sudras	0.48	7.9**
	Muslims	0.88	10.7**
	Others	0.08	1.2
	Social group factor chi-square		394.7**
State:	Andhra Pradesh	0.13	2.1**
	Karnataka	0.47	7.2**
	Tamil Nadu	0.02	0.3
	State factor chi-square		79.2**
R2	0.032
s	2.40