NEW UPDATE: Completed trials may now upload and register supplementary documents (e.g. null results reports, populated pre-analysis plans, or post-trial results reports) in the Post Trial section under Reports, Papers, & Other Materials.
Demand for Nontraditional Cookstoves in Bangladesh
Initial registration date
August 24, 2016
August 24, 2016 7:43 PM EDT
Stanford Medical School
Other Primary Investigator(s)
Department of Civil and Environmental Engineering, Stanford University, Stanford
School of Forestry & Environmental Studies, Yale University, New Haven
School of Forestry & Environmental Studies, Yale University, New Haven
Yale University and Jameel Poverty Action Lab (J-PAL)
Additional Trial Information
There are few marketing studies of social learning about new technologies in low income countries. This paper examines how learning through opinion leaders and social networks influences demand for non-traditional cookstoves – a technology with important health and environmental consequences for developing country populations. We conduct marketing interventions in rural Bangladesh to assess how
stove adoption decisions respond to (a) learning the adoption choices of locally identified ‘opinion leaders’ and (b) learning about stove attributes and performance through social networks. We find that households generally draw negative inferences about stoves through social learning, and that social learning is more important for stoves with less evident benefits. In an institutional environment in which consumers are distrustful of new products and brands, consumers appear to rely on their networks more to learn about negative product attributes. Overall, our findings imply that external information and marketing campaigns can induce initial adoption and experiential learning about unfamiliar technologies, but sustained use ultimately requires that new technologies match local preferences. Registration Citation
Despite much concerted global effort, demand for improved cookstoves remains low. The trial seeks to understand this issue by testing (i) price sensitivity for improved cookstoves (ii) how preference for the stoves varies by gender and (iii) whether social networks influence uptake decision.
The trial was conducted in in two rural sub-districts of Bangladesh where sample households were offered two variants of improved cookstoves - a chimney stove and an efficiency stove - at a variety of price points: free, half-priced and full-priced. Husband and wife were provided marketing information separately and in half the households only the men were allowed to place an order while in the other half only the women could place the order. Opinion leaders in the village were also given the same offers and their choices were announced in the village to see whether their adoption decision affects the adoption decision of other households. Finally, households within the social network of those who decided to use the cookstoves were also sampled to see whether they were convinced to make a purchase with information from those who had actually used the product.
Intervention Start Date
Intervention End Date
Primary Outcomes (end points)
1. Price sensitivity of demand for two non-traditional cookstoves technologies
2. How gender affects choice to adopt improved cookstoves.
3. The degree to which each gender is able to act on their preferences when prices go up.
4. How inter-spouse consultation changes original decision of each spouse.
5. Impact of opinion leaders in increasing adoption.
6. Impact of social network in increasing adoption
Primary Outcomes (explanation)
The variable 'Share of network with cooksstove' is a variable for estimating the outcome 'Probability of ordering stove'. This variable is constructed using stove discounts as an instrument. This experiment was conducted in two stages where in the first stage, the decision of opinion leaders was announced to the village and in the second stage only households who are within the social networks of the adopters were given the offer to purchase the cookstove. Since knowing someone who owns a non-traditional stove may be correlated with the strength of an individual’s preference for stoves (homophily), the price discount IV allows random variation in the proportion of one’s social network that own a stove and overcomes this issue.
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
In the gender experiment, villages were randomized to receive either free or highly subsidized improved cooking stoves and households were randomized to decide whether the husband or the wife shall receive the offer to purchase the stove. A team of two enumerators visited sample households and provided marketing information about the benefits of two types of improved cooking stoves – the chimney stove and the efficiency stove. After the information session, depending on the treatment group, either the husband or wife was offered an opportunity to pre-order either type of stove, but was not allowed to consult with his/her spouse before making the decision. Households were not required to pay at the pre-order stage nor was the pre-order binding. They would pay the price of the stove once the stove was delivered and were free to change their mind upto the point of delivery. This setup allows the study of unconstrained gender differences in prioritization of budget-saving and health-improving technologies. Alternatively, gender differences when stoves are free vs. when small positive prices are charged uncovers the degree to which each gender is able to act on their underlying differences in preferences in the presence of positive prices.
In the social network experiment, a two-stage intervention was made. In the first-stage, three village opinion leaders were identified through focus group discussions. The villages were then randomized to receive either the chimney or efficiency stove at full or half price. In half the villages, the opinion leaders were first offered the cookstoves and their adoption decision was made known to subsequent households who receive the offer. In the other half of the villages, opinion leaders were not engaged. Like the first experiment, households did not have to pay when placing the order and were free to cancel the order upto the point of delivery. In the second-stage intervention, households with social links to first-stage households that purchased an improved stove were offered to buy stoves themselves. This allows the trial to observe whether purchase decisions are affected when participants know their leaders making decisions with the same amount of information as themselves vs. when their social network has more knowledge than they do by virtue of actually adopting and using the cookstoves.
Experimental Design Details
Randomization done in office by computer.
Village, household, para (neighborhood: each village divided in into 3 paras) depending on experiment.
1. Price: randomized at the village level. 22 villages were chosen to receive stoves at full price, 20 remaining villages received stoves at half price. Total 42 villages
2. Gender: randomized at the household level for 800 households.
3. Opinion Leader: randomized at para level. Each village was divided into three 3 paras thus giving 66 full-priced paras and 60 half priced paras. In the full price paras, 30 were randomly assigned an opinion leader while the remaining 33 were not. Likewise in the 60 half- priced paras, half were randomly allocated opinion leaders while the other half were not. Total 126 paras.
Was the treatment clustered?
Sample size: planned number of clusters
Half-price stove: 8 villages
Free stove: 8 villages
Husband choice: 200 households
Wife choice: 200 households
Social Network clusters:
Full price opinion-leader: 30 paras
Half-price opinion-leader: 30 paras
Full price no opinion leader: 36 paras
Half-price opinion leader: 30 paras
Efficient cookstoves: 20 villages
Chimney cookstoves: 22 villages
Sample size: planned number of observations
Gender experiment: 800 households (50 per village x 16 villages) Social network experiment: 2268 households ( 54 per village x 42 villages)
Sample size (or number of clusters) by treatment arms
Free stove: 8 villages x 50 households per village = 400
Half-price stove: 8 village x 50 households per village =400
Social network study:
1620 treatment, 648 control
Full price opinion-leader: 10 villages x 54 per village = 540 households
Half-price opinion-leader: 10 villages x 54 per village = 540 households
Full price no opinion leader: 12 villages x 54 per village = 648 households (control)
Half-price opinion leader: 10 village x 54 per village = 540 households
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
INSTITUTIONAL REVIEW BOARDS (IRBs)
Stanford University IRB
IRB Approval Date
IRB Approval Number
Stanford IRB protocol 6454
Post Trial Information
Is the intervention completed?
Intervention Completion Date
February 28, 2011, 12:00 AM +00:00
Is data collection complete?
Data Collection Completion Date
February 28, 2011, 12:00 AM +00:00
Final Sample Size: Number of Clusters (Unit of Randomization)
Same as pre-trial
Was attrition correlated with treatment status?
Final Sample Size: Total Number of Observations
Gender study: 800 households, social network: 2100 hh
Final Sample Size (or Number of Clusters) by Treatment Arms
Price: 22 villages full price, 20 villages half price
Opinion Leader treatment: 60 treatment, 66 control
Gender treatment: In 400 households husband received offer, 400 households wife received offer.
Is public data available?
This section is unavailable to the public. Use the button below
to request access to this information.
Reports, Papers & Other Materials
PROMOTION OF IMPROVED COOKSTOVE IN RURAL BANGLADESH
This study aimed to explore the factors affecting the promotion of improved cookstove (ICS) to replace traditional stove and hence to combat indoor air pollution (IAP). The study was conducted in 58 randomly selected villages of Jamalpur sadar and Hatia upazilas (29 villages in each) in 2008. Both qualitative and quantitative methods were used. Focus group discussions were performed in each village to divide the villages in three equal clusters as well as Paras and listed the opinion leader of the villages. Fifty randomly selected households and nine households of the opinion leaders were surveyed in each village. Thus, a total of 3,080 households were selected for quantitative survey with pre-designed questionnaire. These households were also offered two types of ICSs – portable and with-chimney under different experimental conditions. Among those who adopted ICS as was offered usually chose portable ICS since they believed this would reduce fuel consumption while they chose ICS with chimney to reduce pollution. We found that households were usually aware of IAP but not so much so of the existence of ICS. But once they came to know about it through this survey, they would expect ICS to be better than traditional stoves in producing better tasting food, less smoke emissions, less cooking and fuel collection time, etc. When compared with those who did not know about ICS before, prior knowledge on ICS was found to be associated with greater share of people thinking ICS was better than traditional in terms of taste of food and smoke emission. In most cases financial constraints was stated as a reason for not to adopt an ICS. The adoption decision was also found to be highly responsive to price. On the other hand, opinion leaders appeared to have a stronger impact on households’ decisions when the leaders decided against ICS as opposed to when they decided in its favour. Although this is a very product specific study the results can provide a guideline to understand similar constraints for many other improved technologies that exist but are not generally adopted.
Arif, Tamid, Anik Ashraf, Grant Miller, Ahmed Mushfiqu Mobarak, Nasima Akter, ARM Mehrab Ali, MA Quaiyum Sarkar, Lynn Hildemann, Nepal C Dey, Mizanur Rahman, Puneet Dwivedi, and Paul Wise. "Promotion of Improved Cookstove in Rural Bangladesh." Working Paper No. 22, BRAC, May 2011.
LOW DEMAND FOR NONTRADITIONAL COOKSTOVE TECHNOLOGIES
Biomass combustion with traditional cookstoves causes substantial environmental and health harm. Nontraditional cookstove technologies can be efficacious in reducing this adverse impact, but they are adopted and used at puzzlingly low rates. This study analyzes the determinants of low demand for nontraditional cookstoves in rural Bangladesh by using both stated preference (from a nationally representative survey of rural women) and revealed preference (assessed by conducting a cluster-randomized trial of cookstove prices) approaches. We find consistent evidence across both analyses suggesting that the women in rural Bangladesh do not perceive indoor air pollution as a significant health hazard, prioritize other basic developmental needs over nontraditional cookstoves, and overwhelmingly rely on a free traditional cookstove technology and are therefore not willing to pay much for a new nontraditional cookstove. Efforts to improve health and abate environmental harm by promoting nontraditional cookstoves may be more successful by designing and disseminating nontraditional cookstoves with features valued more highly by users, such as reduction of operating costs, even when those features are not directly related to the cookstoves’ health and environmental impacts.
Mobarak, Ahmed Mushfiz, Puneet Dwivedi, Robert Bailis, Lynn Hildemann, and Grant Miller. "Low Demand for Nontraditional Cookstove Technologies." Proceedings of the National Academy of Sciences (PNAS) 109(27): 10815-20.
LEARNING ABOUT NEW TECHNOLOGIES THROUGH SOCIAL NETWORKS: EXPERIMENTAL EVIDENCE ON NON-TRADITIONAL STOVES IN BANGLADESH
There are few marketing studies of social learning about new technologies in low income countries. This paper examines how learning through opinion leaders and social networks influences demand for non-traditional cookstoves – a technology with important health and environmental consequences for developing country populations. We conduct marketing interventions in rural Bangladesh to assess how stove adoption decisions respond to (a) learning the adoption choices of locally identified ‘opinion leaders’ and (b) learning about stove attributes and performance through social networks. We find that households generally draw negative inferences about stoves through social learning, and that social learning is more important for stoves with less evident benefits. In an institutional environment in which consumers are distrustful of new products and brands, consumers appear to rely on their networks more to learn about negative product attributes. Overall, our findings imply that external information and marketing campaigns can induce initial adoption and experiential learning about unfamiliar technologies, but sustained use ultimately requires that new technologies match local preferences.
Miller, Grant, and A. Mushfiq Mobarak. "Learning About New Technologies Through Social Networks: Experimental Evidence on Non-Traditional Stoves in Bangladesh." Working Paper, September 2013.
GENDER DIFFERENCES IN PREFERENCES, INTRA-HOUSEHOLD EXTERNALITIES, AND THE LOW DEMAND FOR IMPROVED COOKSTOVES
This paper examines whether an intra-household externality prevents adoption of a technology with substantial implications for population health and the environment: improved cookstoves. Motivated by a model of intra-household decision-making, the experiment markets stoves to husbands or wives in turn at randomly varying prices. We find that women – who bear disproportionate cooking costs – have stronger preference for healthier stoves, but lack the authority to make purchases. Our findings suggest that if women cannot make independent choices about household resource use, public policy may not be able to exploit gender differences in preferences to promote technology adoption absent broader social change.
Miller, Grant, and A. Mushifq Mobarak. "Gender Differences in Preferences, Intra-Household Externalities, and the Low Demand for Improved Cookstoves." Working Paper, Stanford University, January 2013.
REPORTS & OTHER MATERIALS