The misreporting of administrative health data creates an inequitable distribution of scarce health resources and weakens transparency and accountability within health systems. In mid-2010s, an Indian state, introduced a district ranking system to monitor the monthly performance of health programs alongside a set of data quality initiatives. However, questions remain about the role of data manipulation in compromising the accuracy of data available for decision-making. We used qualitative approaches to examine the opportunities, pressures, and the rationalization of potential data manipulation. Using purposive sampling, we interviewed 48 district-level respondents from high, middle and low ranked districts, and 35 division- and state-level officials, all of whom had data-related or program monitoring responsibilities. Additionally, we observed 14 district-level meetings where administrative data were reviewed. District respondents reported that the quality of administrative data was sometimes compromised to achieve top district rankings. The pressure to exaggerate progress was a symptom of the broader system for assessing health performance that was often viewed as punitive, and where district- and state-level superiors were viewed as having limited ability to ensure accountability for data quality. However, district respondents described being held accountable for results despite lacking adequate capacity to deliver on them. Many rationalized data manipulation to cope with high pressures, to safeguard their jobs, and in some cases, for personal financial gain. Moreover, because data manipulation was viewed as a socially acceptable practice, ethical arguments against it were less effective. Potential entry points to mitigate data manipulation include: (i) changing the incentive structures to place equal emphasis on the quality of data informing the performance data (e.g., district rankings); (ii) strengthening checks and balances to reinforce the integrity of data-related processes within districts; and (iii) implementing policies to make data manipulation an unacceptable anomaly rather than a norm.