The million-dollar question


To be more accurate, it is a one million thirty-seven six hundred sixty-four-dollar question. This projection is based on research from Khateera.com. This represents the cumulative projection of monetary value a woman would accrue; had she been paid for the labor of home-care duty (modeled per child).

Image source and number projection proxy via Khateera.com


With the onset of COVID-19 in early 2020, and with a new home-centered lockdown reality, it could be foreseen that the burden of home-care duty on women would increase and drive a deeper wedge in global development goals towards gender equity.

The UNDP 2020 HDP COVID-19 and Human Development: Assessing the Crisis, Envisioning the recovery report stated that “Under such conditions, women will see their burden increase and their effective labor participation and productivity—enhanced capabilities—constrained, limiting their opportunities to live at their full potential at work and in their households.” The report goes on to estimate that unpaid care work needs would rise by 20 hours a week during the pandemic.

The brunt of these hours, almost one full day, is dictated by the “social norms” of a nuclear home; this remains largely governed by majority group/ cultural behaviors.

The Hypothesis: experiment goals and design
The case of home-care duty is transversal, as there are many contributing factors besides social norms and behaviors. However, in our mission as an Accelerator Lab to “build the world’s largest and fastest learning network around development challenges” we can concede to the variability of causes, while we seek to address one priority area in a well-designed testing space so we can remove ambiguity and produce focused evidence that can grow our collective understanding on a complex issue.

According to UNDP’s 2020 HDP on Tackling Social Norms: A game changer for gender inequalitiesDiscriminatory social norms and stereotypes reinforce gendered identities and determine power relations that constrain women’s and men’s behavior in ways that lead to inequality”.

To combat the gendered impacts of COVID-19, UNDP and UN Women Lebanon ventured early 2020 into a pre-emptive awareness campaign to address the “social norms” of the nuclear home.  As a lab, we found that a communication campaign on the topic would present a critical learning opportunity, particularly if we (a) observe and measure the way different demographics react and (b) use this data to conduct experiments that produce behavioral evidence on the topic.

In collaboration with UN Women, our learning experiment became focused on identifying the possible behavioral drivers that, if surfaced, could inform insight interventions around social norm changes in hopes of relieving women and strengthening gender equity at home.

Following the OECD Behavioral Insights Toolkit & Ethical Guidelines we used the awareness campaign that both UNDP and UN Women Lebanon launched in March 2020 as a discovery stepping stone. The OECD guidelines directed us to isolate one “target behavior”, in this case it was the (non)norm of men performing home-care duties.

Using the guidelines, we continued to illustrate the recommended approach of considering variable recorded data sources about the target behavior in an effort to build a testable hypothesis.

Data source 1: Manual sentiment analysis and observations

The awareness campaign consisted of two messaging modes. The first was a series of posts that highlighted facts and figures on the uneven distribution of home care duty to invoke empathy for unjust burdens on women. The second employed an influencer “social media challenge” that solicited male influencers to post videos as they perform home-care duties to normalize and encourage other men.


A manual sentiment analysis helped us reflect on the campaign results in ways
where we could see three clear reception patterns, amongst men in particular. We observed a very slim portion of men exhibiting a welcoming attitude towards home-care duties as part of any adult individual’s routine, while a larger segment expressed acceptance of homecare duty as “help” offered to variable women in their life and regarded as “virtue”, leaving the remainder of expressions (across genders) taking shape in outcry over a perceived “abnormal role reversal”.

An archetypal representation of the common patterns of campaign reception by men


Data Source 2: Discrimination Science Literature

While women’s reactions to the March campaign were also observed, our readings into Michelle “Mikki” Hebl’s discrimination science literature helped us make the decision to target only men in our experiment. From a system thinking perspective, “Lebanese Men/ Men residing in Lebanon'' as an actor, or in this case non-actor, represented the larger lever to learn about.

“A social norm will be stickiest when individuals have the most to gain from complying with it and the most to lose from challenging it.” UNDP’s 2020 Human Development Perspectives, Tackling Social Norms

“Social norms” are dictated and followed by any given group’s “larger majority” or “ones who have the most to gain from complying” and accordingly, to change norms, we may want to investigate what motivates that majority. Discrimination science tells us that men respond better to other men, as opposed to women, especially on topics related to gender disparity.

“Majority members who advocate on behalf of members are more impactful than targets themselves, if you are not a target and you are helping, you are setting the norms...” - Mikki Hebl, PHD in Psychology, Center for Teaching Excellence Faculty Fellow Rice University

This implies that in any behavioral intervention that aims to target men as a larger majority benefitting from an instilled social norm, we would benefit from recruiting male voices or likeness to engender messaging.

Data source 3: UN Women International men and gender equality survey (IMAGES) Lebanon based findings

Our collaboration with UN Women Lebanon also meant that we look into IMAGES MENA (Understanding Masculinities) where 1,050 men and 1,136 women between the ages of 18 and 59, representing both the Lebanese and Syrian populations living in Lebanon answered a survey on several topics around masculinity, with substantial attention given to perceptions on home-care duty. The research outlines what men ‘say’ about:

  • Home care duties they are least likely to take part in at home (e.g., childcare, bathroom cleaning, elderly care, dishes…)
  • Duties that they report actively engaging with (e.g., cooking, repairs around the house, groceries…)
  • Their self-reported motives and barriers to engage in home-care duty (e.g., being a modern family man, public perception of them, independence…)


Our circumstances

If we are to target men around home-care duty with the goal of uncovering levers of behavioral change we needed at first to test their self-stated motives and barriers. Self-professed motives or perceptions is what we call “Attitudinal data”. This type of data is valuable, however if we were to quote the mother of anthropology herself “What people say, what people do, and what they say they do are entirely different things.” - Margaret Mead. Accordingly, attitudinal data from IMAGES research was used to draw the testing parameters for an experiment that aims to uncover “Behavioral data”.

“Self-reported memories, beliefs, preferences, intentions and experiences are not mental facts fetched from our inner libraries, but rather seem to be constructions assembled when the circumstances calls for them...” - the OECD Behavioural Insights Toolkit & Ethical Guidelines

“When the circumstances call for them” is the key phrase here. To uncover any meaningful insights around men’s behaviors we must construct a simulated circumstantial environment of choices, in an effort to figure out what would be the real motives behind picking up homecare duties for a 21 - 65+ year old man residing within Beirut and Mount Lebanon.


The test: construct a simulated circumstantial environment of (measurable) choices
While different combinations and tests can be drawn from the data triangulated above, our interest mainly took shape in testing for four key “assumed drivers”:

Behavioral driver prompts to test included:

  • Modern family man as an identity driver:
    Identifying oneself as a modern family man and contributing to one’s family’s happiness as a driver (source: UN Women IMAGES research and UNDP 2020 HDP COVID-19 and Human Development report)
  • Self-independence as a behavioral driver:
    The desire for self-independence as a driver (source: UN Women IMAGES research and manual sentiment analysis)
  • Home-care duty as monetary investment vs. “free care” expectation mental frame:
    Linking home-care duty to monetary and investment language as a driver (source: desk research)
  • Performing home-care duty as a virtue driver:
    Reputational motivation or “self-signaling” oneself as virtuous for partaking in home-care duty as a driver (source: desk research + sentiment analysis)

Having to roll out this experiment under COVID constraints made us think of the online social media as the lived circumstantial space for a large swathe of men in Lebanon. Accordingly, and with Facebook’s targeting and geofencing capacity we were able to promote combinations of images and captions, each of which representing one of the behavioral drivers we wanted to test. The men, naturally, scrolling their Facebook feed would see a variety of prompts and only take the ‘quiz’ to find out the monetary value of their ‘investment’ in home-care duty by clicking on the behavioral prompt that resonated with them. The quiz results submitted would not be a reliable data source for us; but the quiz was a good gamification technique to draw interest into the topic, the main source of data was the “choices” men made when presented with different prompts on Facebook.

The capacity to measure which prompt combinations drove men to engage enough to take the quiz acted as the “observed built environment of choice”. In this sense by collecting the data of what choice an average man would make in this “natural environment” and not under the blistering light of interrogation, we were able to identify and debunk some of our assumptions on behavioral motives and barriers.

How did we get there?

a) Men in Beirut and Mount Lebanon would see variable combination prompt(s) on their Facebook feed.

b) The call to action on all prompt variations is to “Measure their investment in homecare duty via a quiz”, upon clicking they would engage with the online quiz site to measure their home-care investment in monetary and hourly value.

c) The results page shows their score in $ value and time, while also displaying a series of “how-to” videos around key home-care duties that the research suggests men take little to no interest in.

d) If any of the “how-to” videos are clicked they would be directed to a “how-to” play-list on UNDP’s YouTube channel.

The different engagement metrics across this journey are given different weights and consider the different variables attached to each click and accordingly we could discern between behavioral drivers or insight, interest, and valuable potential for further experimentation.

So, what gets men in Lebanon interested in home-care duty?

Graph that illustrates men’s engagement metrics across the steps of the experiment journey


We were able to use this quasi-natural environment to test our hypothesis with 25,000 Lebanese men online, 6,000 of whom answered our quiz, and 700+ watched the “how to” content.

The experiment ran for 2 weeks at the end of December 2020. The target was men aged 20 to 65 in Beirut and Mount Lebanon; our original goal was to drive a maximum of 6,000 men to just see the quiz, as that would ensure that we have statistically significant data on what prompt combination i.e. behavioral driver drove most men to show interest in home care duty. However, as the actual result was 4 times this number, we can imply that gamification and monetary value linked to home care duty are successful in creating interest for men.

The Facebook prompts test was the main source of behavioral insight. While our target was men aged 21 to 65 years old, in the second week of the experiment we noticed two trends in overall interest. The montage of celebrity influencers we reused from the earlier campaign garnered little to no interest at all; the second interest trend was that there was a clear divide in how men 21-35 years old reacted in contrast to their older counterparts (35 – 65+). The younger group gravitated in droves to one of the four prompts, and it was “Self-independence as a behavioral driver”. However, it appeared that older men would be more inclined towards “Performing home-care duty as a virtue driver/ signal”.

Our findings suggest that older men were moved by messaging that amplifies their reputational virtue; we can infer that perhaps this age bracket can be externally motivated if more of their majority groups showed off homecare duty as a source of pride or if more prestige was attached to such labor. As the younger group, on the other hand, was moved with Self-independence as a behavioral driver, the landslide appeal of this prompt tells us that self-independence is a deeply moving goal for younger men, and if we are to attach home care duties to this value we can tap into a more intrinsic goal for young men. In contrast to looking good, this driver is centered more around lifestyle values and goals, thus presenting a better investment for future exploration.

The modern family man as an identity driver did not resonate across any of the age groups, in fact the family prompt used was the least popular amongst them all. This could go back to the adage shared earlier on the “intent and action gap” between what people say and what they, in fact, can be observed choosing in their day to day lives. This is also the reason why we disregarded all of the quiz results submitted, as the data from the quiz is self-reported and had a sole functional goal of incentivizing men to engage on the topic. However, the large number of quiz completion confirms that there is high interest value in linking “home care duty” with the language of pay, investment, and earned value.

While the experiment offered many other micro insights that would make for further testing potential, it is important that we share the findings that were statistically significant. This means that if we were to repeat this, we should get similar results. For this reason, we are continuing to experiment on this topic in 2021.

As we continue learning, we hope to keep sharing openly and at large so that we hear back on what our findings open-up in terms of spaces for communications and programmes for application. If self-independence presented as a key behavioral driver for most men, in what ways can we use this behavioral insight to drive measurable behavioral change?

If you or any of your teams are working on this topic and would want help in applying these behavioral findings in testable and measurable spaces, please get in touch!

Icon of SDG 03 Icon of SDG 05

UNDP Around the world

You are at UNDP Arab States 
Go to UNDP Global