The Crime in India report for the year 2016 is finally released by the NCRB. The report raises questions about the reliability of government data. The National Crime Records Bureau (NCRB) has finally released the ‘Crime in India’ report for the year 2016. The data on communal incidents/riots from this report once again raises questions about the reliability of government’s crime data. Factly was one of the first ones to report that the National Crime Records Bureau (NCRB) data on communal incidents/riots was not matching with the data provided by the Home Ministry for the year 2014. The trend continued in the year 2015 as well as 2016. A deeper analysis of state wise data from these two sources raises serious questions about the reliability of government data. NCRB Data Collection on Riots The National Crime Records Bureau (NCRB) is an attached office in the Ministry of Home Affairs (MHA). NCRB is tasked with collection and dissemination of a secure national database on crimes, criminals and law enforcement agencies etc. Every year, the NCRB collects data from Police departments at the state level and compiles it into a Crime in India report. The report has detailed information on various types of crimes. In 2014, the NCRB has made significant changes to the method of its collection and dissemination of crime data. The template has been undergoing changes every year. The present template for collection of crime related data, has a section on riots as defined under Sections 147,148,149,150, 151 and 153A of the Indian Penal Code (IPC).There is also separate section to collect data on offences that promote enmity between different groups as defined under Sections 153A & 153B of the IPC.The template makes it clear that cases reported under Section 153A with other sections of riots (Section 147 to 151) should be shown in the category of riots whereas cases reported specifically under Sections 153A & 153B should be shown separately. Before 2014, all the data related to riots used to be collected under one head called riots. From 2014, the data on riots is categorized into many different sub heads namely communal riots, Industrial riots and riots for political reasons etc. There is also a separate section on ‘Offences Promoting Enmity Between Different Groups’. One has to remember that NCRB gets all the data from state police departments. And this data is strictly based on the sections mentioned in the First Information Report (FIR) of the cases filed. The NCRB also follows the ‘Principal Offence Rule’ for counting of crime. Hence among many offences registered in a single case, only most heinous crime (maximum punishment) will be considered as counting unit, thereby representing one case. Sometimes it is also possible that multiple cases will be filed for a single incident. NCRB Vs MHA – The discrepancy in data continues The data on communal riots in the NCRB report does not match with the data shared by the Ministry of Home Affairs (MHA) in the Lok Sabha. For the year 2014, NCRB reported almost twice the number of incidents compared to the number released by MHA. The gap closed in 2015 and again widened in 2016. In 2016, the NCRB reported 869 cases relating to communal riots compared to 703 communal incidents reported by the MHA.In addition, NCRB reported 336 cases relating to ‘Offences Promoting Enmity Between Different Groups’ in 2014 out of which 323 are offences for promoting enmity between groups based on religion, race etc. In the year 2015, the corresponding number of cases was 424 and 378. In 2016, a total of 478 cases relating to ‘Offences Promoting Enmity Between Different Groups’ were reported by NCRB out of which 447 were offences for promoting enmity between groups based on religion, race etc. If the number of cases relating to communal riots and cases relating to promoting enmity between groups from the NCRB report are added, then the difference between the NCRB data & MHA data widens further. What is the Government’s explanation? The government on multiple occasions maintained that (NCRB) collects police registered communal riots cases (FIRs) from States and there could be multiple FIRs in a single incident of communal violence and hence NCRB data may be different compared to MHA data which is communal incidents based but not FIR based. Why Government’s explanation is not plausible As per the government’s explanation, the NCRB numbers should be higher for each state since they are based on FIRs in police stations and that there could be multiple FIRs for a single incident. Even if this is believed to be true, the difference or the trend should be uniform across states.