[Jonathan Seiden]

Hi Sarah,

I’d recommend using KoBo over Tangerine for administering IDELA. The XLSForm that is available here is easily adaptable is already optimized for use in KoBo. We may be updating those templates to a more user-friendly version, but they work perfectly fine as-is. XLS forms make adding additional languages very easy as you just have to translate the “labels” for questions and choice answers.

It’s quite common to have the instructions in a colonial language and then the child-facing language in a local language. Just ensure you translate the words in **bold** into local language.

If you make a French version in XLSForm, we’d love for you to share the final version so that others can use it in the future.

[Jonathan Seiden]

We have detailed instructions for how to clean and analyze IDELA data here.

In essence, each domain score (Emergent Numeracy, Emergent Literacy, Motor, and Social-Emotional) is calculated by the average percent correct for the tasks in each domain. So, in the Motor domain, if a child hopped 7/10 steps, drew 6/8 body parts, drew a good triangle with 3 sides and folded 1/4 folds, they would score 70% on Hopping, 75% on Drawing a Person, 100% on Drawing a Shape and 25% on Folding. The Motor score would be (70%+75%+100%+25%)/4= 67.5%. The Total IDELA score is simply the average for the four core domain scores.

The Excel Template also automatically will calculate the scores based on the default IDELA assessment. You can access it here.

[Jonathan Seiden]

We don’t have specific recommendations for technology for IDELA. You definitely do not need “top-shelf” devices like iPads to work. At the same time, it’s important to buy devices that will not break too quickly. If you have a complex form, newer tablets will typically run faster, so more RAM is generally helpful.

KoBo Toolbox (our recommended electronic data collection platform) has general guidelines on devices here: http://help.kobotoolbox.org/collecting-data/recommended-devices-for-data-collection

[Jonathan Seiden]

Hi Sarah,

Great questions! Let me address them one by one:

For the household measure, is there a composite measure/index of responses that Save has found corresponds to a household environment that is conducive for child development?

We don’t have a standardized way of measuring the overall household environment, but we often create individual indices for reading materials, toys, and home learning activities. We usually create indices for the “total number of types of X” reported by the caregiver. We can then generate variables for the total number of types of toys and reading materials (and sometimes combine this index to create the overall “Home Learning Environment”). With the home learning activities, we usually create a composite for “any” caregiver, and then also analyze it by the mother and father reported activities.

Same question for socioeconomic status. Has Save the Children used IDELA-HE to create an index of SES?

We take a similar approach to this issue as well. We can create a measure of “total number of types of household possessions” as a proxy for SES.

Obviously, these approaches are a bit reductionist because we are assuming that all types of possessions, toys, reading materials, etc. contribute equally. You may want to use more sophisticated measures such as factor analysis to address these limitations. The trade off is then how to communicate your results. It’s relatively easy to understand that, on average, one additional reading material was associated with an X p.p. change in IDELA score. It’s more complicated when you condense variables with other methods. One option would be to use those other methods, but then rely on quantiles to communicate your results.

• This reply was modified 7 years, 4 months ago by Jonathan Seiden.

[Jonathan Seiden]

This is a really important question that (while I was still at SC), we spent a lot of time thinking, discussing and, at least for me, worrying about.

As Frannie and Lauren mentioned, it’s possible to set benchmarks at the 25% and 75% level to create three broad categories that generally show that children are able to do almost none, some, or almost all of the IDELA activities correctly. However, this comes with a couple of limitations. First of all, we should be *very* careful comparing domains against each other–25% on Motor does not indicate the same level of development as 25% on Emergent Literacy. Second of all, we need to be very cautious about interpreting the results because they are not age-adjusted–of course younger children are going to be considered “struggling” more often than older kids, even when they may be developmentally very much on track.

Also as Lauren alludes to, the ideal benchmarking situation will consist of (unfortunately) a much more involved process, but also one that will yield much more informative and culturally and policy relevant information. That process, simply stated, involves convening a panel of experts in a given context, reviewing the individual activities that children are asked to do on the assessment, and coming to a consensus of what children would be expected to be able to do at various ages. After deciding on the individual activities, these can be

The biggest advantage of this approach is that it generates conceptually comparable benchmarks for different domains and ages of children–you might find that a score of 50% on Motor would generate worry for a 6-year-old, but that on the Emergent Literacy Domain (which is substantially “harder”) would be considered on-track.

This non-statistical benchmarking approach is generally known as “Policy linking” and there are a variety of approaches that can be used. When done in a systematic way across contexts, it can yield a much more nuanced and informative benchmark.

[Author]

Replies