What is FWI’s current standard?
Version 1 of our welfare standard in India contains two core components: a stocking density ask and a water quality ask. As of this writing, 54 fish farms and 1 corporation have voluntarily committed to these standards.
When farmers sign up for the Alliance for Responsible Aquaculture (ARA), they agree to a stocking density cap of 3000 fish per acre. 
Lowering stocking densities often is important for farmed fish, both directly because it reduces ammonia production, disease rates, and biological oxygen demand, and indirectly because it prevents farmers from using our water quality improvements as motivation to increase the number of fish in the system, thereby sacrificing any welfare gains that have been achieved.
3000 fish per acre has, in the relatively sparse literature assessing Indian Major Carp, been shown to be beneficial for promoting fish health.  However, we plan to further test the correct balance between improved conditions for fish and economic viability from the farmer’s side.
We also have stocking caps for other types of systems:
Fish-Shrimp polyculture: 3000 fish/acre and 80,000 shrimp/acre
Juvenile fish: 50,000 fish/acre
We are currently considering a new stocking density for larger ponds (>20 acres)
As a member of the Alliance for Responsible Aquaculture (ARA), every 15 days a FWI representative visits a farmer’s pond and takes water quality measurements. The core measurements are:
Nikhil, one of our data collectors, using a water quality monitoring device at a fish farm in Andhra Pradesh, India.
FWI has created required and optimal ranges for each of these parameters. When we find that one of these parameters is outside of the required range, we tell the farmer corrective actions that we believe will resolve the problem (see our corrective actions). To remain in the ARA (and to thus gain benefits like regular water quality monitoring) farmers must consistently make a good faith effort to implement the corrective actions.
Typical corrective actions include turning on aerators, increasing water exchange, reducing feed quantity, and liming the pond.
What impact has our standard had?
Over 2021, Version 1 of our welfare standard potentially helped about 350,000 fish (our goal for the year was 300,000). This was around 40% of the total fish on ARA farms. We did not help fish on every ARA farm either because certain farms were already performing well (no corrective actions or stocking decrease needed) or because our corrective actions failed to improve conditions on certain farms. We believed our standard potentially helped a further ~1 million shrimp, although we are highly uncertain about this number and don’t think much weight should be put in it. For more information, see Our Impact.
These are promising results, however, we believe that it is important to assess both how many animals were affected, as well as to what degree we expect those animals to be living better lives.
Considering our stocking density cap, there is some evidence this has not had significant positive effects on fish welfare. Conducting a Pearson correlation coefficient on the relationship between stocking density and dissolved oxygen (the main parameter that excessive stocking densities jeopardize, and a closer proxy for welfare) showed no significant correlation (roughly -0.1). 
This would make sense, as 3000 fish/acre does not represent too large of a decrease in the total individuals from what farmers were already stocking (on average farmers show a ~325 fish per acre reduction, or about 10%, in their intended stocking before and after joining the ARA). We are currently assessing the viability of lower stocking densities (see What are the next steps below).
Note that this is not considering longer-term implications of keeping stocking densities at suitable limits, such as preventing more intensified production practices that damage fish welfare and the environment.
Considering water quality corrective actions, there seems to be good evidence that FWI has had positive effects on fish lives. Roughly 60% of our corrective actions are followed by an improvement in the parameter that was outside of the required range. Some of this will happen naturally through a regression to the mean. However, given our field staffs' extension interaction with and understanding of these farms, we feel intuitively confident that a significant portion of these improvements had causal links to FWI (and fish marked as potentially helped by FWI are only those that we feel ≥80% confident we are causally responsible for improvements in their pond).
However, corrective actions have appeared to only have short-term implications for water quality parameters. Across the entirety of our measurements, we have not seen a significant improvement in water quality being within the range that we would expect from our improvements. As such, we believe our standard currently reduces extreme suffering events (such as critically low oxygen events), but does not give long-term relief from the chronic welfare issues fish farming creates (such as consistently harmful oxygen levels). As such, we believe that pivoting to a standard that focuses on preventative measures may be able to increase our overall impact across the entirety of fishes’ lives.
What are the next steps?
Currently, FWI is continuing to implement Version 1 of our standard under the belief that this is bringing better lives to fish, both through direct water quality improvements and indirect benefits such as preventing future stocking density increases. We will continue to add more farmers to the ARA in order to proliferate Version 1 as well as fish welfare concepts more broadly.
Whilst this is happening, however, several of our team are working to define Version 2 of our standard that we hope can bring more long-term benefits to fish. We are currently finalizing the plans for a test of a new standard in Eluru, one of the regions in which FWI works in India, which will focus on further reductions in stocking density as well as trialing different methods for controlling phytoplankton populations (a key aspect of controlling dissolved oxygen in earthen pond systems). To learn more about our plan for 2022, see our 2022 scale up plan.
1 - We used fish per acre even though we believe fish per cubic meter is more accurate to the experience of the fish. This is because acre is the most common denominator used in India, and almost all farmers use uneven earthen ponds that they do not know a volume for.
2 - Feeding and feed management of Indian major carps in Andhra Pradesh, India. 3 - However, the Pearson correlation coefficient is a bivariate analysis that cannot take into account more complex relationships between multiple variables. As such, deeper assessment would be needed to draw strong conclusions (a rudimentary principal component analysis showed a larger negative correlation).