Thank you, Elisabeth, for inviting me to participate in this.

Science articles, unfortunately, have to be very short (with strict guidelines from the journal), and have to written in a specific style (with a tough anonymous peer review process mediating the stiff competition for space on the journal), so it was not possible to answer all the relevant questions that have come up, in the text of the short article. The journal therefore allows us to link to an "online supplementary materials" document that provides more details. I have attached the supplementary materials doc here*, and this is much longer than the article itself. Many of your questions will get answered if you read this document.

As you read it, it is useful to keep in mind that the various interventions described were not dreamt up solely by the research team, but were co-designed with, and implemented by the inventors of CLTS: the NGO Village Education and Research Committee (VERC) and Wateraid, Bangladesh. Kamal Kar was a consultant at VERC, I believe, where the CLTS model started. All interventions described were implemented by these organizations with no interference from the research team (except to ensure scientific rigor), in order to maintain research and evaluation independence.

Finally, each community was randomly assigned to just ONE particular intervention arm, as is standard in any randomized controlled trial. VERC and Wateraid did a fantastic job implementing, as far as we (the research team) can tell using the five rounds of data we have collected from these 18,000+ households. So I don't think the interventions in any one village or community was "confused" in any way.

I have to teach today, and then present at a seminar, but I'll try to come back and answer questions.


Note by moderator:
* Or see here in the SuSanA library: www.susana.org/en/resources/library/details/2246

Just a little note to say that I will send an e-mail today to the lead author, Ahmed Mushfiq Mobarak, to alert him to this discussion. I actually e-mailed him already earlier today because he is a grantee with the Gates Foundation and I asked him if he could share more information about two of his sanitation grant in the grant database, i.e. these two:

Project: Does sanitation behavior migrate? Evidence from seasonal migrant workers in Bangladesh
www.susana.org/en/resources/projects?sea...ion+behavior+migrate

and
Project: Bolstering demand for improved sanitation when adoption decisions across households are inter-linked
www.susana.org/en/resources/projects?search=Bolstering+demand

Sounds like they are doing interesting research there at Yale University.

And I do think that the journal title was most likely chosen by the authors, not by the journal (unless Science is different to other academic journals, I am not sure). Joe, do you have the URL for the press release that you posted at the beginning of this thread?

I also read the Kaiser piece and couldn't really understand what it was supposed to add. It reads to me a bit like a news article (I have written news for Science before) but the way it is displayed makes it look like part of the journal. I really could not see what relevance describing the comment about Hindu religious feelings (which we have discussed here on the forum) have to a research study in Bangladesh.

I agree with Giacomo that latrine building is not necessarily a good measure of "sanitation", but I think it is a fair way to compare the treatments under discussion. What systems could we expect CLTS to construct which were better than latrines?

Finally, the title might have been decided by the journal, so I'm not sure we should blame the authors for a title which might claim wider relevance than it deserves. Clearly this only applies to a particular situation in Bangladesh, but the debate is whether it is illustrative of CLTS as a whole.

Dear Joe and others,
Overall, I find this an interesting article. However, it does leave some issues unanswered as Joe and Marijn already pointed out. The chosen language also lightly irritated me, especially given that Science is quite an influential journal.

First, I would like to help answer one of your questions Joe. In the the article the "supply treatment" is described as follows:

The Supply treatment was intended to improve the functioning of markets by providing technical assistance and information. In communities assigned to the Supply treatment, VERC selected a local resident with technical skills and trained him as a Latrine Supply Agent (LSA). The LSA received a fixed salary to provide information to neighborhood residents on (i)where to purchase a hygienic latrine; (ii) how to assess the quality of a latrine offered for sale; (iii) how to install and maintain a latrine.

(page 2)
My question here is, how comparable is this to 'typical' sanitation marketing approaches? Anyone with experience who can comment on this?

Now, on to some other issues that sort of bother me (other than the use of incomprehensible acronyms).

1) The article is named "Encouraging sanitation investment in the developing world: A cluster-randomized trial", while the experiment is clearly only taking place in Bangladesh. I know it's tempting to have one fix-it-all solution, but please let's refrain from making such generalist assumptions. Even in the accompanying commentary article it is stated

But others say an encouraging result in Bangladesh may not say much about India, where the hygiene problem is much bigger. A survey released last year by the Research Institute for Compassionate Economics (RICE) in New Delhi found that many Indians who have access to a hygienic latrine still prefer open defecation, particularly Hindus, who were not represented in the Bangladesh study. The reasons may involve Hindu caste system beliefs that allowing feces to accumulate in a latrine is impure and that only an “untouchable” can remove them, says RICE Executive Director Dean Spears. He adds: “I would just say we should be very careful about generalizing [the Science findings] to Hindu rural north India.”

(Kaiser, Jocelyn "For toilets, money matters" Science 17 April 2015:Vol. 348 no. 6232 p. 272)

2)This article is about increasing access and use of latrines by promoting construction of latrines. Is this 'sanitation'? To many, yes. To me (and others), no. As the quote above also makes clear, we all (should) know by now that sanitation is more than just building a latrine. There's a whole sanitation chain that needs to be taken into account if we want the actual service to keep running for more that 2 years. This also comes at a higher cost (as the WASHCost project of my former employer pointed out). Articles like this, published in influential journals, perpetuate in my view the idea that once a latrine is built all problems will be solved. Maybe we should really start to make the distinction between 'sanitation' and building latrines and make it into a point of importance for example when reporting or doing peer-review.

Best,
Giacomo

Hi Marijn - here is a brief synopsis of the paper, but it would be interesting to interview the authors to ask more detailed questions about this research. I might do that if I can find an outlet for it, or maybe see if we can have them here on the forum to explain more about it.

On the specific point you are asking: the paper says that 380 communities of 18,254 households in 107 villages in the Tanore region were randomised to three different treatments. First the Latrine Promotion Programme, a "community motivation and health information campaign", second the programme with subsidies, third something described as "a supply-side market access intervention linking villagers with suppliers and providing information on latrine quality and availability", fourth, the supply intervention with subsidies and finally a control group which received none of the interventions. The researchers followed up a year after the intervention to see the effects.

It says the LPP followed "closely" the principles of CLTS, with a course over several days to raise awareness of sanitation issues. But it does say that the difference to the standard model was that it focussed on the importance of hygienic latrines rather than just ending Open Defecation.

In the subsidy treatment, the poorest three-quarters were entered into a lottery to receive vouchers for subsidies which represented 75% saving on the instillation of facilities. The number of winners was randomised to 25%, 50% and 75% to see the impact of intensity.

It doesn't seem to say much about what the "supply" intervention involved.

In terms of results, it says that at the baseline 78% of households had access to a latrine, and the subsidy had a 9.4% increase on latrine access. The effect on ownership was higher, with a 20% increase over the baseline.

The LPP and supply treatments (ie without subsidy) had no effect on adult OD, but adding a subsidy to the LPP led to a 22% reduction in OD rates compared to the control group.

Voucher winners were more likely to own hygienic latrines than households in communities that only had LPP, or lottery losers in subsidy villages, and voucher winners were more likely to invest it in a latrine if more neighbours also won the lottery. A lottery winner in a high intensity village is 20% more likely to own a hygienic latrine than someone in the LPP only treatment.

On this point the authors say:

The move from subsidizing 25% to subsidizing 50% of the poor produces the largest demand spillovers in our context. Asking community members to make a joint investment commitment, as in CLTS, is a potentially useful intervention, but our results suggest that this should be accompanied by targeted subsidies.

Finally they say that the study has limitations because it might not give a general indication to other populations outside of the one studied in Bangladesh, only considered one level of subsidy, did not include a subsidy-only treatment (as considered bad practice to give subsidies without education) and based OD rates on self-reporting, which could be biased. They also did not consider health outcomes.

They also say that the study allows

us to document some of general equilibrium changes operating via a social influence mechanism, but our results remain silent on wider general equilibrium effects operating via price mechanisms.

My observations are:

• I wonder if the emphasis on latrines may have put off people who might have taken up other kinds of cheaper intervention if it had been offered
• I wonder how they are defining 'hygienic' latrines
• The complexity of the treatments seems to me to give quite a confusing story
• It is not clear to me whether the reduction in OD is related to the increase in latrines (are people using subsidised latrines of their neighbours? or finding some other non-latrine way to reduce OD?)
• I don't really understand what the 'supply' treatment is
• Other than stating that CLTS principles were closely followed, it does not seem to indicate exactly what was involved, who did it, the quality of the intervention or if there was any variability across the group
• I also wonder if the groups knew about each other, and whether there was any kind of disapproval effect. If the non-subsidised groups knew about the subsidies others had access to, did this mean that they were less likely to participate in the LPP programme?

If I was to write about this study, these are the kind of questions I would be asking the authors about. But, all of that said, it seems to be a properly (although rather oddly) randomised trial which compared CLTS behavioural techniques with controls and subsidies, and so perhaps addresses some of the complaints recently levelled at the World Bank report and the Indian Total Sanitation campaign study.

Joe,

Thanks for posting this. It says in the article, that the community motivation program (LLP) was modeled after CLTS. Is there more info in the document as to how close it resembles "pure CLTS"?

A new study in the journal Science, from Innovations for Poverty Action. This is the full press release:

April 16, 2015, NEW HAVEN, CT – With poor sanitation estimated to cause 280,000 deaths per year worldwide, improving sanitation is a key policy goal in many developing countries. Yet governments and major development institutions disagree over how to address the problem. A new study released in Science today found that in Bangladesh, a community-motivation model that has been used in over 60 countries to increase use of hygienic latrines had no effect, yet latrine coverage expands substantially when that model is combined with subsidies for hygienic latrines targeted to the poor.

The study, led by Raymond Guiteras of the University of Maryland and James Levinsohn and Mushfiq Mobarak of Yale University, and implemented by Innovations for Poverty Action, tested three different approaches that are commonly used in the development sector for increasing the use of hygienic latrines. Reducing open defection, which is still practiced by 15 percent of the world’s population, is a key policy goal for this sector. The study took place in northwest Bangladesh, in an area where 50 percent of the population had access to a hygienic latrine before the study began.

“While there is general agreement among development professionals and institutions about the importance of improving access to hygienic latrines, there is still vigorous debate about the most cost-effective ways to achieve this.” said Mobarak. “Is the problem a lack of cash, or is the problem an absence of strong community norms against open defection? Even when households are willing to pay for hygienic latrines, does lack of access to toilet components or lack of information about quality or installation methods impede adoption?”

Researchers randomly assigned 380 neighborhood communities, or 18,254 households total, to one of four groups. Villages either received a community motivation program, subsidy vouchers with the community motivation program, information and technical support, or none of the above. By comparing outcomes in latrine coverage, investment in hygienic latrines, and open defecation between the groups over time, researchers were able to compare the effect of the different approaches.

The subsidy vouchers, which were only provided to a random subset of households in the second group through a public lottery, could be redeemed for a 75 percent discount on available models of latrines, priced (after subsidy) from $5 to $12. The households were responsible for their own transportation and installation costs, and the richest 25 percent of households were not eligible for vouchers.

The community motivation program, called the Latrine Promotion Program (LPP), was modeled after “Community-Led Total Sanitation”, which focuses on behavioral change and community mobilization in eliminating open defection. Such programs have been implemented in over 60 countries worldwide.

Researchers found that the community motivation model alone did not significantly increase adoption of hygienic latrines or reduce open defection relative to the comparison group, nor did providing information and technical support to community members.
However, the subsidy had substantial effects when coupled with the community motivation program, increasing hygienic latrine coverage by 22 percentage points among subsidized households and 8.5 percentage points among their unsubsidized neighbors.
This suggests that latrine investment decisions are inter-linked across neighbors, and that there are positive effects on others of subsidizing even a few households. People were more likely to invest if more of their neighbors received vouchers, pointing to a virtuous cycle where adoption of improved latrines spurs further adoption.

Adding subsidies to the community motivation model also reduced open defection rates by 22 percent among adults in villages that received subsidies (including households that did not receive subsidies), relative to the comparison group.

These results counter the concern among many development practitioners that subsidies undermine intrinsic motivation. Rather, this research shows price is a primary barrier, which is consistent with a growing body of research on adoption of health products.

“These results have particularly important implications in densely populated developing countries, such as India and Bangladesh, where sanitation coverage is low and the public health consequences are high,” said Annie Duflo, Executive Director of Innovations for Poverty Action. “The study also teaches us about how to conduct ‘smart subsidy’ policy, allocating subsidies in a way that maximizes the chances of behavioral changes among neighbors. Given how widespread the community motivation model is, the results of this study can help the sector allocate funds more efficiently,” Duflo said.

The study is here: www.sciencemag.org/content/early/2015/04/15/science.aaa0491 but only the abstract is freely available:

Poor sanitation contributes to morbidity and mortality in the developing world, but there is disagreement on what policies can increase sanitation coverage. To measure the effects of alternative policies on investment in hygienic latrines, we assigned 380 communities in rural Bangladesh to different marketing treatments—community motivation and information; subsidies; a supply-side market access intervention; and a control—in a cluster-randomized trial. Community motivation alone did not increase hygienic latrine ownership (+1.6 percentage points, p=0.43), nor did the supply-side intervention (+0.3 percentage points, p=.90). Subsidies to the majority of the landless poor increased ownership among subsidized households (+22.0 percentage points, p<.001) and their unsubsidized neighbors (+8.5 percentage points, p=.001), which suggests investment decisions are interlinked across neighbors. Subsidies also reduced open defecation by 14 percentage points (p<.001).

+++++++
Note by moderator:
Title and author details:

SANITATION SUBSIDIES
Encouraging sanitation investment in the developing world: A cluster-randomized trial
Authors: Raymond Guiteras1, James Levinsohn2, Ahmed Mushfiq Mobarak2,*