J. Philippe Rushton and Richard Lynn are two psychology professors working mostly on race and IQ relationship. Though they interpret the data bit strangely that could be attributed to their funding trust.
The studies that caught my eye were Rushton's imaginatively titled;
General mental ability in South Asians: Data from three Roma (Gypsy) communities in Serbia(Via Dienekes)
Skin color and intelligence in African Americans
According to Rushton;
Roma's average lower than Europeans on g, like South Asians.
According to Lynn;
There is a positive correlation between skin color and general intelligence among African Americans. Lighter the skin colour; higher the IQ. This could be attributed to European genes.
Now, let us consider the genetic make up of Roma and their founding population in India. Romas are believed to have originated in present day South-East Indian region(Bengal, Orissa and Andhra Pradesh). Their language shows close affinity to Sinhala. Sinhalese are mostly descendants of East Indian communities. Also, like East-South Indians they also lack Y-Haplogroup L.
I am not sure of Serbian Roma(the population used in the study) genetic make up. If they resemble non-Sinti Romas then they must have European male lineages like R1b and I and female lineages like H, X in appreciable frequency.
If there were no cultural differences as stated by Rushton then we should go by Lynn's skin color and intelligence measurement. In India, the erstwhile outcastes(to which Romas probably belonged to, they are culturally similar to Doms or Dombas) and tribals still do not enjoy similar cultural advantages as other castes we have to compare other non-Brahmin castes with Romas.
At least from photos on the web, Romas look lighter than most of the castes in South-East Indian region which we can attribute to their higher European contribution. By Lynn's analysis their g should be higher than (and not equal to) Indians which could be around 80-85. However, Rushton's study does not observe that. When it has been observed that lighter skin and IQ positively correlated among African Americans but not among Romas it might show that only cultural inputs are relevent. Gene inputs might be positive/negative if they are associated with cultural inputs but on their own they are actually neutral.