[SMARTKeymaster]

SMART recommends 6 days to complete an enumerator training. The 6 days includes the standardization test, field test day, as well as a day for field test feedback and practice and minor questionnaire adjustments based on field test. The Enumerator Training Manual found at the link below on the SMART website provides all information required to conduct an enumerator training as well as a sample agenda on page 8.

https://smartmethodology.org/survey-planning-tools/smart-capacity-building-toolbox/

[SMARTKeymaster]

You should continue to go to all clusters and households after the minimum sample size (697 children) is reached to maintain a representative sample. The following resources found on the website describe in detail how to calculate sample and sampling methods:

Survey Planning Tools/ Sampling for SMART: Sampling Methods & Sample size calculation for the SMART methodology: https://smartmethodology.org/survey-planning-tools/smart-methodology/

Survey Planning Tools/ SMART Capacity Building Toolbox: Survey Manager Module 3 Sampling: https://smartmethodology.org/survey-planning-tools/smart-capacity-building-toolbox/

[SMARTKeymaster]

Information about the Technical Advisory Group and partners can be found on the website in the About SMART- SMART Community section. https://smartmethodology.org/about-smart/community/

[SMARTKeymaster]

There has not been a specific date determined for the launch of ENA for SMART software that includes the mortality plausibility report but it should take place within a few months.

[SMARTKeymaster]

The latest version on ENA (Feb 4, 2015) does not include the mortality plausibility report and ability to generate survey protocols. These additions will be made in subsequent versions of ENA for SMART software. Apologies for the confusion; the necessary changes will be made to the website.

[SMARTKeymaster]

Yes, if only anthropometric data from 6-59 month children is being collected, a shortlist of households with 6-59 month children can be used for random selection. However, it is VERY important that the household list is up to date. If the household list is even 6 months old it can be considered out of date. It is recommended to create an under 5 updated household list and begin data collection as soon as possible after the list has been updated.

[SMARTKeymaster]

A) Villages in two municipalities that are now ODF: You have indicated that these villages have been targeted for your project; therefore, the total population of each of these villages should be included in your sampling frame for one SMART survey. In your report it is important to emphasize that the sampling frame used for your baseline survey was the villages in the two municipalities that were ODF and NOT all the villages (population) of the two municipalities.

B) The third municipality should be a second SMART survey. Since your program will not be operating in this municipality there is no need to include it in your baseline survey of the ODF villages in municipalities 1 and 2. If the third municipality contains some villages that are ODF and others that are not, separate surveys should be performed (or only target ODF/ or non ODF villages depending on objectives) because CLTS intervention has a significant impact on nutrition status. Also, the proposed survey for the third municipality should only take place if an action is likely to take place in a reasonable amount of time (if results indicate actions are required), otherwise, the survey would be unnecessary at this time.

[SMARTKeymaster]

The process of calculating a combined GAM rate is a bit more involved.

When surveys are undertaken in separate strata (zones), the sizes of the populations (and sample size in each stratum) are different, and the distribution of the sample for each strata is not proportional to the sample calculated for the entire population.

Weightings are correction factors to adjust these differences in probability of selection/inclusion (differences in the size of the populations). The weight for each child is calculated taking into account their unequal probability of being selected between strata (zones) and is inversely proportional to their chance of being selected/included in the survey. The standardised weights for each strata are then calculated and used to determine a cumulative GAM. This will correct for the bias of a non-representative sample (differences in the size of the populations in each stratum/zone).

For example, to calculate cumulative GAM for rural (population: 20,000, sample size 800, 15% prevalence) and urban strata (population: 75,000, sample size 600, 5% prevalence):
Calculate the weights for each stratum.
Rural: 20,000/800 = 25.
Urban: 75,000/600=125
Calculate the standardised weighting factors.
Rural: 25 / [(25+125)/2*] = 0.333
Urban: 125/ [(25+125)/2*] = 1.667
*Divide by the number of strata. i.e., if you were to have 4 strata/zones, divide by 4.
In ENA Options tab, select Showing columns for measure, clothes and weighting variables.
Enter data for each of your strata in the Data Entry Anthropometry tab.
Identify strata (zones) by numbers (e.g., 1-4), and
Enter the corresponding standardised weight factor in the WTFACTOR column for each strata.
Select Weighted Analysis of Survey from the Extras menu for your weighted prevalence (it will not appear in the Results tab).
See the ENA Software Manual: https://smartmethodology.org/survey-planning-tools/smart-emergency-nutrition-assessment/

• This reply was modified 11 years, 2 months ago by  SMART.

[SMARTKeymaster]

SMART recommends entering mortality data at the individual level as it allows ENA to stratify rates by age and gender, along with their respective confidence intervals and design effects. Contrary, data entered at the household level is summary data and mortality rates cannot be stratified by age and gender. Also, entering data at the individual level allows ENA to generate demographic data such as age pyramids. When analyzing mortality data from surveys, it is beneficial to see whether rates are higher amongst specific populations (e.g., by age and gender), and this cannot be done as well if data is entered at the household level.

Another reason why it is recommended to enter mortality data at the individual level as opposed to the household level is because during the data collection phase the same SMART mortality survey questionnaire form is used; therefore, it takes the same amount of time to conduct the interview (whether or not the data is entered for household or individual level analysis). In summary, entering data at the household level as opposed to the individual level, is a missed opportunity to analyze important demographic information that the enumerators collected.

[SMARTKeymaster]

In the Results Anthropometry tab if no exclusions is selected, children (data sets) can still be omitted from analysis; therefore, the total number of children analyzed will be lower than the total number of data sets listed in the data entry anthropometry tab. For example, if weight/height and no exclusion are selected but height data is missing from 1 child and weight data is missing from 1 child, these two children will NOT be included in the analysis.

[Author]

Posts