How to start a Secondary Data Review?

(Chataigner 2018).

Undertaking an SDR is heavy on resources, as doing an SDR is time consuming. This means that SDRs are best done at places where it does not drain resources from the field, for example at HQ. Prior to starting collecting secondary data, the analytical plan needs to be in place, which will clearly define what data will be collected, for whom, and for what purpose.

1. Locate your data sources

These can be websites, archives, newspapers, I/NGO and UN reports). This is a time consuming step as depending on the topic there may be a lot of information available, and this may come in different formats (Excel, Word, PDF, etc.). The extraction of the pieces of data that is actually useable from these sources, and putting that into a workable format in order to do analysis, is lengthy.

2. Read through your sources thoroughly

Usually references within your sources can direct you to further information on the topic.

3. Store and tag your data

4. Assess the collected data

This step starts going into the exploratory stage (or first stage) of analysis. Read more on this here

Questions on this include

1. How useful is my data?
2. How reliable is my source?
3. What is the validity of the data collection methods used?
4. How useable is this data compared to my analysis framework?
5. Look at what you have available after you have assessed and filtered through your data. Is your data consistent? Triangulate any important data you have found.

6. Identify information gaps

What information gaps do you have left? What are they? Can they be filled?

An information gap is generally data that is missing against the analysis framework, and critical to complete analysis. Information gaps may also occur when information acquired is either too qualitative or too quantitative, and the other is missing. Sometimes, data is too aggregated and not informative enough to do analysis.

When such gaps occur, an analyst can:

1. Find more information by spending more resources
2. Use proxies to replace missing data
3. Contact experts for more insight/expert judgment
4. Look for lessons learnt from similar contexts
5. Do primary data collection.

7. Follow the steps in the analysis spectrum

When conducting an SDR, keep in mind:

• The research process is time consuming and can continue endlessly. Know when to stop SDR in light of time- and resource constraints.
• Scrutinize information and identify the underlying details of important facts, patterns, trends, significant differences or anomalies that are not always readily visible. Consider the details.
• Separate the matter into key parts and/or essential elements; break things down; identify causes/key factors or features/possible results.
• Ensure there is enough time to turn data into information. Often a great deal of time is spent collecting information, but too little time given to preparing for data collection, or analysing it.
• Challenge your own assumptions and conclusions. Discuss your findings with your colleagues and reach consensus on conclusions.
• Consider bias and reliability/credibility. Don’t rely on one source only.
• Be sceptical when dealing with comparisons. Always remember that correlation DOES NOT imply causation.
• Be careful of the actual meaning of terms used. Terms such as ‘affected’, ‘household’, or ‘community’ can mean different things in different areas. Definitions may change over time and where this is not recognised, erroneous conclusions may be drawn. Provide a definition for potentially confusing or sensitive terms.
• If you use technical terms, make sure you define them correctly. E.g., specify which type of malnutrition you are referring to (stunting, wasting, etc.)
• Ensure the SDR is properly referenced. A well-documented SDR and analysis allows for easier use of the material by other interested parties and allows for greater credibility of the product.
• Clearly define when information is based on assumptions instead of on facts or sufficiently verified information.
• Think about whether or not your findings make sense (Does it fit in with the history and context? Does it make sense to the people living there? Etc.) (ACAPS 06/2011)