Forgot your password?
Document Actions
Stacy Philpott's ideas
Up to Meta-analysis
Posted by
----reposted from email-----
I would suggest that yes, you do need to demonstrate
a thorough search. I would say this is especially important in
meta-analysis because biases may come up when only the most-cited, or
the most easy to find articles are included. So doing and saying
something like the following would work:
I would suggest that yes, you do need to demonstrate
a thorough search. I would say this is especially important in
meta-analysis because biases may come up when only the most-cited, or
the most easy to find articles are included. So doing and saying
something like the following would work:
"We
searched the ISI Web of Science, Biological Abstracts, Agricola, and
Orton Library for all articles for keywords <x, x, x, x, >. We
found a total of 89 articles. From those, we selected only those that
met our specific criteria for the meta-analysis: 1) that they were
conducted in Mesoamerica, 2) that they included natural forest, 3) that
they measured at least 2 taxonomic groups, and 4) that they included a
species list as part of the results [for example]. We were left with a
total of 13 articles that met our search criteria. We extracted data
from these studies and ........"
searched the ISI Web of Science, Biological Abstracts, Agricola, and
Orton Library for all articles for keywords <x, x, x, x, >. We
found a total of 89 articles. From those, we selected only those that
met our specific criteria for the meta-analysis: 1) that they were
conducted in Mesoamerica, 2) that they included natural forest, 3) that
they measured at least 2 taxonomic groups, and 4) that they included a
species list as part of the results [for example]. We were left with a
total of 13 articles that met our search criteria. We extracted data
from these studies and ........"
Besides
being important, it also may be necessary to get enough articles that
meet your specific criteria. It is often surprising to find that out of
a few hundred articles that only 20 or so really have the information
that you need or want, so the more exhaustive the search, the larger
sample size you'll have for the meta-analysis.
being important, it also may be necessary to get enough articles that
meet your specific criteria. It is often surprising to find that out of
a few hundred articles that only 20 or so really have the information
that you need or want, so the more exhaustive the search, the larger
sample size you'll have for the meta-analysis.
in yesterday's chat, we discussed how to go about doing the search, and to what extent we need to standardize it. Also, we decided that we would all like to use papers cited within the papers which result from the search -- at least to go back step into the literature. Some of us have been using this as a main method for getting more papers, and we thought that this would be ok if we could write in the methods something like:
"we searched xxx databases using xxx keywords, and supplemented these with additional papers cited from the resulting papers"
And we would also say that we did the search from February - March 2007, just to make it more definite.
However, we didn't come up with a list of key words yet. Should we put that together here?