Moderate Need, Acute Need: Valid categories for humanitarian needs assessments? Evidence from a recent needs assessment in Syria

from Assessment Capacities Project
Published on 26 Mar 2015 View Original


What this is about

Needs assessments in emergencies seek to establish, among other elements of the situation ,the number of persons in need. “Persons in need” is a broad and fuzzy concept; counts ,proportions and other measures capturing how many are in need, and how intensely are necessarily imprecise. Some sectors achieve greater precision by applying specific standards, such as in malnutrition surveys. However, such methods are not feasible in types .of assessments that depend largely on local key informant estimates.

Greater depth, if not outright precision, may be achieved in other ways. Categories can be refined. Instead of just the binary “persons in need” / “not in need”, needs may be graded. Several recent needs assessments in Syria have done so by distinguishing between persons in moderate need” and “persons in acute need”. This note is about this kind” of refinement in persons-in-needs (PiN) estimates in the most recent assessment. The Syria Multi-Sectoral Needs Assessment (MSNA), released in October 2014, provides PiN estimates for 126 of the 270 sub-districts of the country. Persons in moderate and in acute need have been estimated separately for five sectors - food security, shelter, non-food .items (NFI), health and safe water.

Empirical tests

Whether estimates of persons in acute need add value consistently - in all sectors, and in relation to other measures of need - is an open question. We seek to answer it by submitting the MSNA :estimates to two tests .

Proportions of persons in need, calculated for each assessed area, are correlated among sectors • Areas with relatively more persons in need of food assistance tend of have more persons needing shelter, etc. This is so because extreme events (and even more so persistent crises) trigger .cascading deprivations. In groups with multiple deprivations, needs grow acute more rapidly Thus, in the MSNA, we expect the correlation pattern of persons in acute need to be at least as .strong as that of all persons in need .

Needs assessments produce estimates of the severity of unmet needs in each of several sectors • The enumerators in the MSNA rated the severity on a scale with seven levels. The number of persons in need influenced their judgments. If the enumerators applied the “acute need” category consistently, we expect that the proportions of persons in acute need had a significant influence ,on their severity ratings. We expect to find this effect even after we take into account, statistically .the effect from all persons in need

The “acute needs” concept passes both tests. It passes them for all five sectors. This is a strong indication that the enumerators used the concept consistently, and that their estimates of persons in .acute need are informative

What causes acute needs?

That is not enough. One likes to understand what causes acute needs. We expect to find them closely associated with certain factors in the disaster / emergency / conflict environment as well as with the .configuration of needs in other sectors

However, the MSNA data do not exhibit some of the associations that we expect in the Syrian case. Notably, the proportions of IDPs in sub-districts are not, or are negatively, correlated with the proportions of persons in acute need in the different sectors. This may be so because IDPs flock to areas with better conditions. To test for this, we replaced the IDPs with proportions of the preconflict population that fled their sub-districts. Statistical estimates confirm the expected effect on the current populations (= the remaining populations plus the IDPs they are hosting) for the shelter and health sectors. However, the opposite tendency was found for acute needs of clean water: the greater .the population loss, the lower the proportion of persons currently in acute need

These tangled associations lead us to conclude that the causal pathways from the conflict environment to acute needs are complex and varied. Also, there is evidence that regional particularities matter. We obtained different patterns across the governorates that the MSNA covered. This diversity shifts the .causal validation of acute needs from universal factors to local interpretation

By contrast, our model estimates confirm, for each sector, the association of acute needs with the average” need levels in the other four sectors. This is in line with the idea of cascading deprivations that push needs to acute degrees.

Recommendations for future use

Overall, we find theoretical and empirical support for the inclusion of the acute-needs measures in future rapid needs assessments. Such measures, especially the proportion of persons in acute need of assistance in sector X in social group Y and area Z, provide a .sharper image of the pattern of unmet needs

We recommend the inclusion of such measures with these precautions:

• Adopt “persons in acute need” in the design only if assessment workforce and key .informants are such as to expect reasonably consistent and reliable estimates

• Collect estimates of the number of persons in acute need. In the analysis, make .meaningful comparisons of proportions of such persons in current populations

• Compare persons in acute need to all persons in need, rather than to those in moderate need. Statistics in terms of moderate needs can be misleading; low proportions can .be due to either low overall impact or to high proportions of persons in acute need

• Expect a significant number of areas reporting zero persons in acute needs in some .or all sectors. While debriefing enumerators, ask specifically why none were reported

• Limit the analysis of acute needs in the context of other needs and of likely causal variables to tabulations involving not more than three factors. Purposive sampling and other selection effects are likely to make statistical modeling unwieldy, slow or .unproductive Information about acute needs is helpful for planners and operators. Yet refined categories ,come at a cost. This is true also of needs assessments that work with three, instead of two ,levels: none / moderate / acute. They demand more detailed attention in instrument design training, supervising, debriefing, processing and reporting. The assessment consumers too .need to absorb more complex information .

The nature of this note is conceptual and statistical. There is also an ethical side to it Category changes in social policies create new information that can motivate new interpretations and new priorities. They can redraw the lines of social inclusion and exclusion. Donors may - incorrectly - conclude that “all persons in need” overestimate the real needs, and that “persons in acute need” are more accurate measures and closer to priorities. Assessment teams must avoid language that might favor such a perception .and should focus on the important question of securing humanitarian access to all in need

The MSNA findings reinforce our belief that “acute needs” is a feasible and productive concept. It can improve the measurement of unmet needs under conditions that rarely permit exact classification. As long as everybody remains aware that the distinctions are fuzzy, and estimates need interpretation in context, such measures will strengthen judgment .on humanitarian priorities