2007
DOI: 10.1016/j.anbehav.2006.12.028
|View full text |Cite
|
Sign up to set email alerts
|

Discriminant function analysis with nonindependent data: consequences and an alternative

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1
1

Citation Types

2
240
0

Year Published

2009
2009
2013
2013

Publication Types

Select...
8

Relationship

0
8

Authors

Journals

citations
Cited by 288 publications
(242 citation statements)
references
References 31 publications
2
240
0
Order By: Relevance
“…Consequently, the actual strength of the individual vocal signature is probably somewhat lower than reported here. The strength of the sex-specific signature, however, is not statistically overestimated since we avoided pseudoreplication by using only mean values per individual (see [46] for details). Interestingly, our results indicated that female S. bilineata which are approximately 15 per cent larger than males [31] tended to call at higher frequencies with shorter call duration compared with males.…”
Section: Discussionmentioning
confidence: 99%
See 1 more Smart Citation
“…Consequently, the actual strength of the individual vocal signature is probably somewhat lower than reported here. The strength of the sex-specific signature, however, is not statistically overestimated since we avoided pseudoreplication by using only mean values per individual (see [46] for details). Interestingly, our results indicated that female S. bilineata which are approximately 15 per cent larger than males [31] tended to call at higher frequencies with shorter call duration compared with males.…”
Section: Discussionmentioning
confidence: 99%
“…A cross-validation procedure was used to correctly assign echolocation call duplets to the respective individuals. We estimated the significance of the classification success by using a two-tailed binomial test (following [46]). In order to test for sex-and location-specific signatures in echolocation calls, we used two different statistical approaches.…”
Section: Methodsmentioning
confidence: 99%
“…In this procedure, each call is classified by the functions derived from all calls other than that one. Since the acoustic data for food-associated calls were two-factorial (caller ID; call type), it has been argued that conventional DFA does not allow for a valid estimation of the overall significance of discriminability (Mundry & Sommer 2007). Therefore, we used a permuted discriminant function analysis (pDFA), using a macro written by R. Mundry and C. Sommer (Mundry & Sommer 2007), to estimate the significance of the number of correctly classified calls (cross-validated).…”
Section: Acoustic Analysesmentioning
confidence: 99%
“…The success rate of classification of call types was highest for barks and peeps, followed by yelps and then peep-yelps (correct classification for barks ¼ 96%, peeps ¼ 94%, yelps ¼ 80%, peepyelps ¼ 74%). We then used a pDFA (Mundry & Sommer 2007) to estimate the significance of the number of correctly classified calls (cross-validated). Results from the pDFA indicated a highly significant level of discrimination (P ¼ 0.001).…”
Section: Acoustic Analysesmentioning
confidence: 99%
“…To test further whether acoustic structure could be assigned to the identity and cycle stage, we performed a discriminant function analysis (DFA). Because cycle stage was confounded by multiple contributions from the same female, we conducted a permutated discriminant function analysis (pDFA) using scripts provided by R. Mundry (Mundry and Sommer 2007). We performed colinearity checks for the 10 parameters and found that all had satisfactory variance inflation factors of <7.0 (Allison 1999;Slocombe and Zuberbühler 2007).…”
Section: Statistical Analysesmentioning
confidence: 99%