front |1 |2 |3 |4 |5 |6 |7 |8 |9 |10 |11 |12 |13 |14 |15 |16 |17 |18 |19 |20 |21 |22 |23 |24 |25 |26 |27 |28 |review

For eligible studies, outcomes of interest as well as other study characteristics have to be coded. Issues to consider include (1) the number of items that will be coded, (2) how one will assess inter-coder bias if more than one person is coding the studies, and (3) what items should be coded. Ideally, one should conduct a pilot study by coding several eligible studies prior to the actual investigation taking place. In this way, modifications can be made to the codebook in order to ensure a thorough and accurate recording of relevant information. If resources are available, it’s recommended that at least two people code the same studies and then review every data point for accuracy and consistency. Discrepancies can be resolved by consensus as well as with the help of outside experts in the field.