The process starts with your protocol, in which you:
- Identify data categories to be used for coding and extraction. It's the essential information you need to complete the review; for example, participant demographics, sample size, methods, interventions, and outcomes.
- Describe the methods you will use.
see RESOURCES (box to the right) to help you work through category identification and methods.
- Coding and extraction may be used during the second screening to help select eligible studies.
- Coding and extraction is required for evaluation of the final set of selected eligible studies.
Essential:
- A Recording Form. Data categories are used to create a spreadsheet or special data collection form.
- Consistency. information is recorded in a carefully determined format from each eligible study. (Or, during screening, for every study still under consideration.)
- Coding refers to recording information such as who conducted the study and the design of the study.
- Extraction refers to recording the results of the study.
Recommended Practice:
- Pilot (test) your collection form to make sure it works as planned.
- Record data with synthesis in mind; that is, plan ahead for how you will be using the data.
- Record all information necessary to address the synthesis question, plus any additional information to help with critical appraisal.
- Use a "notes" field to record important information that doesn't fit elsewhere.
- Choose narrow categories over broad categories. Think of age ranges; these can be combined later but will be hard to divide out later.
- But try to balance between under and over-extraction.
- Cochrane methodology advises having two people for most of this work and considers two a mandatory minium for recording results.
- Know that you may have to contact researchers for answers to questions not answered in reports of their research.
Examples of Forms for Coding & Extraction: