TY - JOUR
T1 - A prospective power analysis and review of habitat characteristics used in studies of tree-roosting bats
AU - Lacki, Michael J.
AU - Baker, Michael D.
PY - 2003
Y1 - 2003
N2 - We identified 25 studies published between 1988 and 2001 that measured characteristics of roosting sites of tree-roosting bats, and where measures were compared to characteristics of random or available locations. The most frequently measured habitat characteristics were roost-tree diameter (n = 23), roost-tree height (21), roost-tree canopy cover (16), roost height (14), and slope (10). Habitat characteristics of the roost tree itself were measured more frequently than stand or landscape characteristics; a total of 31 different habitat characteristics was used to describe stand or landscape conditions as opposed to 23 different habitat characteristics used to describe features of the roost tree. The overall mean (± SE) number of habitat characteristics examined per study was 8.0 ± 1.1, with an average of 4.2 ± 0.7 characteristics reported to be significant (P < 0.05). Mean estimated effect size, or the absolute value of the difference between means divided by the population standard deviation, of habitat characteristics ranged from 0.83 to 1.52. A sample size of 11 radio-tagged bats was sufficient to achieve acceptable power, i.e., 0.80, for all habitat characteristics examined when only using the upper limit of the 95% confidence intervals for estimated effect sizes. In contrast, a sample size of 39 radio-tagged bats was sufficient in achieving the same level of power for only 50% of the habitat characteristics evaluated at the lower end of the 95% confidence intervals. We encourage researchers to conduct pilot studies, and estimate effect sizes and variances to assess the level of sampling effort required to evaluate habitat characteristics in studies of tree-roosting bats.
AB - We identified 25 studies published between 1988 and 2001 that measured characteristics of roosting sites of tree-roosting bats, and where measures were compared to characteristics of random or available locations. The most frequently measured habitat characteristics were roost-tree diameter (n = 23), roost-tree height (21), roost-tree canopy cover (16), roost height (14), and slope (10). Habitat characteristics of the roost tree itself were measured more frequently than stand or landscape characteristics; a total of 31 different habitat characteristics was used to describe stand or landscape conditions as opposed to 23 different habitat characteristics used to describe features of the roost tree. The overall mean (± SE) number of habitat characteristics examined per study was 8.0 ± 1.1, with an average of 4.2 ± 0.7 characteristics reported to be significant (P < 0.05). Mean estimated effect size, or the absolute value of the difference between means divided by the population standard deviation, of habitat characteristics ranged from 0.83 to 1.52. A sample size of 11 radio-tagged bats was sufficient to achieve acceptable power, i.e., 0.80, for all habitat characteristics examined when only using the upper limit of the 95% confidence intervals for estimated effect sizes. In contrast, a sample size of 39 radio-tagged bats was sufficient in achieving the same level of power for only 50% of the habitat characteristics evaluated at the lower end of the 95% confidence intervals. We encourage researchers to conduct pilot studies, and estimate effect sizes and variances to assess the level of sampling effort required to evaluate habitat characteristics in studies of tree-roosting bats.
KW - Bats
KW - Effect size
KW - Habitat
KW - Power analysis
KW - Sample size
KW - Tree roosts
KW - Variance
UR - http://www.scopus.com/inward/record.url?scp=1142302216&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=1142302216&partnerID=8YFLogxK
U2 - 10.3161/001.005.0211
DO - 10.3161/001.005.0211
M3 - Article
AN - SCOPUS:1142302216
SN - 1508-1109
VL - 5
SP - 199
EP - 208
JO - Acta Chiropterologica
JF - Acta Chiropterologica
IS - 2
ER -