MIT Research News' Journal
 
[Most Recent Entries] [Calendar View]

Tuesday, July 17th, 2018

    Time Event
    4:55p
    A new way to measure women’s and girls’ empowerment in impact evaluations

    Women make up half the world’s population, but just 12 percent of the world’s heads of state and government. This disparity underscores a persistent reality in the 21st century: Despite steady advances in women’s rights in recent decades, gender norms and biases continue to constrain human potential around the world.

    A growing number of policymakers believe that investing in women and girls’ empowerment can reduce these and other gender-based inequalities. The United Nations' Sustainable Development Goal 5, for example, seeks to achieve gender equality and empower all women and girls. Increasing empowerment is also seen as a promising strategy to unlock greater economic growth in low- and middle-income countries.

    In order to design effective policies and programs, however, researchers, policymakers, and practitioners must be able to accurately measure women’s and girls’ empowerment. A new research resource from MIT’s Abdul Latif Jameel Poverty Action Lab (J-PAL) addresses this challenge.

    Co-authored by Rachel Glennerster, chief economist at the UK Department for International Development and former executive director of J-PAL; and Lucia Diaz-Martin and Claire Walsh of J-PAL, the “Practical Guide to Measuring Women’s and Girls’ Empowerment” offers guidance on strategies to help navigate and overcome common challenges in effectively measuring empowerment in impact evaluations.

    Evaluating impact

    Researchers can rarely observe people’s decision-making in real-time, and survey questions that ask study participants about decision-making do not always lead to reliable responses, particularly when questions touch on sensitive topics. For these reasons, it can be hard for researchers and practitioners to identify whether a social program or intervention actually increased people’s decision-making power, a common measure of empowerment.

    Beyond survey challenges, questions arise around which outcomes best capture changes in empowerment. For example, should researchers focus on measuring educational attainment? Agency in household or community decision-making? Employment and control over income? Women and girls experience constraints that are deeply tied to their specific context. Because these constraints can vary so widely, what empowerment looks like for female students in Ghana might be very different from what it looks like for women living in a rural village in India. Researchers often grapple with how to generalize lessons learned from a particular program when what empowerment looks like can vary greatly around the world.

    J-PAL’s new guide draws on strategies from multiple academic disciplines to tackle these measurement issues. Rich with case studies and concrete examples, it outlines actionable steps to improve measurement.

    Determine local context

    Understanding the local context is key. Before trying to measure empowerment in an impact evaluation, researchers must have a nuanced picture of the local context and the specific barriers that women face when trying to make meaningful choices about their own lives. Qualitative research methods such as semi-structured interviews, needs assessments, direct observation, and focus groups can help by creating repeated opportunities to listen closely to the people living in a particular community. A measurement strategy to quantify empowerment is only as good as researchers and practitioners’ understanding of gender and power dynamics in the local context.

    In an evaluation in Bangladesh, for example, Glennerster and co-authors were interested in measuring adolescent girls’ mobility. After several focus groups with adolescent girls they learned that asking the generic question “How far away can you travel from home by yourself?” would not capture how a girl’s mobility was constrained, because the answer depended on what she was doing and for whom. Girls could travel to and from school alone, but they could not travel alone to do things that only had value to them, like going to local fairs. Since empowerment is about people’s abilities to make choices that matter to them, researchers added a question about mobility for activities that only had value to the adolescent girls in addition to the usual questions about going to school or visiting relatives.

    Develop a theory for how intervention generates impact

    Developing a clear theory of how an intervention generates impact can help researchers select accurate indicators of empowerment. To identify the outcomes of a women’s empowerment program (the change or impact we expect to see) and indicators (observable signals we use to measure that change), researchers and practitioners need a deep understanding of the pathways through which the program can affect people’s lives.

    Mapping these pathways, from program inputs (like funding and staff time) to long-term outcomes, is also known as a “theory of change” and results in documentation of a program’s logical chain of results. This mapping process helps clarify appropriate measurement indicators, and helps researchers identify which assumptions must hold true for the program to succeed.

    Develop a plan and carry out testing

    Once researchers decide what outcomes to measure, they should develop and pilot data collection instruments in communities similar to ones where the evaluation will take place. This is an important reality check to make sure surveys work in local contexts.

    For example, many commonly used survey questions to measure household decision-making are hard to ask and answer in practice. In Bangladesh, Glennerster and co-authors found that women gave very different answers to the general question, “Who usually makes decisions about healthcare for yourself: you, your husband, you and your husband jointly, or someone else?” and the more specific question, “If you ever need medicine, could you go buy it yourself?” Piloting different versions of a question can help researchers learn whether they are truly capturing the information they think they are. 

    Non-survey instruments can also be powerful for measuring things surveys can’t capture accurately — like gender bias. In a study on female leaders in India for example, researchers randomly assigned survey participants to hear one of two identical audio recordings of a short speech by a political leader, one spoken by a man and the other by a woman. They then asked participants to rate the leader’s effectiveness. Because the gender was the only difference between the two recordings, researchers could use this technique to measure bias against female leaders.

    After researchers conduct a comprehensive pilot and incorporate lessons learned, they should design a practical data collection plan. Although data collection can be full of unexpected challenges, finding reliable, culturally appropriate, and convenient methods and times to collect survey data can help overcome measurement errors.

    Why measure empowerment?

    “If measurement techniques are inaccurate, it can be difficult to understand whether programs are effective, and how to improve on existing approaches,” says J-PAL’s Claire Walsh. Ensuring that measurement tools are reliable and precise can help researchers avoid drawing inaccurate conclusions about the impact of a program.

    J-PAL recently announced new efforts to further expand the base of policy-relevant evidence related to gender and women’s empowerment. Alongside this research, J-PAL continues to create practical resources to support policymakers, practitioners, and researchers in effectively incorporating analysis of gender dynamics and impacts into their impact evaluations. For more information about this work, visit povertyactionlab.org/gender.

    11:59p
    Study finds climate determines shapes of river basins

    There are more than 1 million river basins carved into the topography of the United States, each collecting rainwater to feed the rivers that cut through them. Some basins are as small as individual streams, while others span nearly half the continent, encompassing, for instance, the whole of the Mississippi river network.

    River basins also vary in shape, which, as MIT scientists now report, is heavily influenced by the climate in which they form. The team found that in dry regions of the country, river basins take on a long and thin contour, regardless of their size. In more humid environments, river basins vary: Larger basins, on the scale of hundreds of kilometers, are long and thin, while smaller basins, spanning a few kilometers, are noticeably short and squat.

    The difference, they found, boils down to the local availability of groundwater. In general, river basins are shaped by rainfall, which erodes the land as it drains down into a river or stream. In humid environments, a large fraction of rainfall seeps into the Earth, creating a water table, or a local reservoir of groundwater. When that groundwater seeps back out, it can also cut into a basin, further eroding and shifting its shape.

    The researchers found that smaller basins that are formed in humid climates are heavily shaped by the local groundwater, which acts to carve out shorter, wider basins. For much larger basins that cover a more expansive geographic area, the availability of groundwater may be less consistent, and therefore plays less of a role in a basin’s shape.

    The results, published today in the Proceedings of the Royal Society A, may help researchers identify ancient climates in which basins originally formed, both on Earth and beyond.

    “This is the first time in which the shape of river networks has been related to climate,” says Daniel Rothman, professor of geophysics in MIT’s Department of Earth, Atmospheric, and Planetary Sciences, and co-director of MIT’s Lorentz Center. “Work like this may help scientists infer the kind of climate that was present when river networks were initially incised.”

    Rothman’s co-authors are first author and former graduate student Robert Yi, former visiting graduate student Álvaro Arredondo, graduate student Eric Stansifer, and former postdoc Hansjörg Seybold of ETH Zurich.

    A climate connection

    In previous work published in 2012, Rothman and his colleagues identified a surprisingly universal connection between groundwater and the way in which rivers split, or branch. The team formulated a mathematical model to discover that, in regions where erosion is caused mainly by the seepage of groundwater, rivers branch at a common angle of 72 degrees. In follow-up work, they found that this common branching angle held up in humid environments, but in dryer regions, rivers tended to split at narrower angles of around 45 degrees.

    “River networks form these beautiful branched structures, and previous work has helped explain the angles at which rivers join together to form these structures,” Yi says. “But each river is also intimately connected to a basin, which is the area of land that it drains rainwater from. So we suspected that the shapes of bains could contain some similar geometric curiosities.”

    The team set out to find a similar universal pattern in the shape of river basins. To do this, they accessed datasets containing detailed maps of all the rivers and basins in the contiguous United States — more than 1 million in total — along with datasets containing two climatic parameters for every region in the country: precipitation rate and potential evapotranspiration, or the rate at which surface water would evaporate if it were present.

    The datasets contained estimates of each river basin’s area, which the researchers combined with the length of each basin’s river to calculate a basin’s width. They then noted for each basin, an aspect ratio — the ratio of a basin’s length to width, which gives an idea of a basin’s overall shape. They also calculated each basin’s aridity index — the ratio between the regional precipitation rate and potential evapotranspiration — which indicates whether the basin resides in a humid or dry environment.

    When they plotted each basin’s aspect ratio against the local aridity index, they found an interesting trend: Basins in dry climates, regardless of size, took on long, thin shapes, as did large basins in humid environments. However, smaller basins in similarly humid regions looked significantly wider and shorter. 

    “We found that arid basins roughly kept their shape with size, but humid basins got narrower as they grew larger,” Yi says. “That confused us for a long time.”

    Answers in the ground

    The researchers suspected that the dichotomy between dry- and humid-type shapes stemmed from their previous observations of branching rivers: In humid climates, groundwater plays an additional role to rainfall in creating wider branches of a rivers, compared with in drier climates. They reasoned that groundwater may play a similar role in widening a river’s basin.

    To check their hypothesis, they looked at characteristics of each basin’s geology, such as the types of rock and soil underlying the basin, and the depth to which groundwater might penetrate. In general, they found that in drier climates, any rainwater that seeped into the ground would dribble deep below the surface, like liquid running through a Brillo pad. Any resulting reservoir, or water table, would be too deep for groundwater to come back up to the surface.

    In contrast, in more humid environments, water is more likely to saturate the soil, like tap water soaking a damp sponge. In these climates, water would seep into the ground, creating large water tables close to the surface.

    The team then computed the extent to which stream locations corresponded to locations where groundwater emerged. They found a greater correspondence where there was more groundwater seeping out around river basins in humid climates, versus in drier climates. This suggests that groundwater plays a bigger role in carving out humid basins, creating wider, more squat shapes, in contrast to the longer, thinner shapes of dry-climate river basins.

    This groundwater effect may be especially pronounced at smaller, more local scales over several kilometers. At much larger scales, spanning nearly half the continent, the group found river basins, even in humid environments, took on long, thin contours, which may be attributed to the fact that, over such a vast area, the interaction between groundwater and the large-scale structure of river networks is relatively weak.

    “Our paper establishes a new, large-scale connection between hydrogeology and geomorphology,” Rothman says. “It also represents an unusual application of the physics of pattern formation. … All this turns out to be connected with fractal geometry. Thus in some sense we are finding a surprising connection between climate and the fractal geometry of river networks.”

    This research was supported, in part, by the U.S. Department of Energy Office of Science, Office of Basic Energy Sciences, Chemical Sciences, Geosciences and Biosciences Division.

    << Previous Day 2018/07/17
    [Calendar]
    Next Day >>

MIT Research News   About LJ.Rossia.org