The Veterinary Journal 196 (2013) 247–252

Contents lists available at SciVerse ScienceDirect

The Veterinary Journal

journal homepage: www.elsevier.com/locate/tvjl

Evaluation of a novel feather scoring system for monitoring feather damaging

behaviour in parrots
Yvonne R.A. van Zeeland ⇑, Madeleine J. Bergers, Lisette van der Valk, Nico J. Schoemaker,
Johannes T. Lumeij
Division of Zoological Medicine, Department of Clinical Sciences of Companion Animals, Faculty of Veterinary Medicine, Utrecht University, Yalelaan 108, 3584 CM Utrecht,
The Netherlands

a r t i c l e i n f o a b s t r a c t

Article history: Feather damaging behaviour is common in captive psittacine birds and there is a need for reliable meth-
Accepted 28 August 2012 ods to evaluate the efficacy of therapeutic and preventive interventions. This study compared the inter-
and intra-observer reliabilities of a novel feather scoring system with an existing system to assess the
plumage of grey parrots (Psittacus erithacus). Regions of the body were photographed separately at
Keywords: 1 week intervals and shown at random to 35 examiners (avian veterinarians and veterinary students),
Psittacine who used the two scoring systems to assess plumage. Since the quality of the photographs was insuffi-
Plumage
cient to allow accurate assessment of the individual flight and tail feathers, the novel scoring system was
Feather picking
Feather plucking
only evaluated for its reliability regarding covert and down feathers. Inter- and intra-observer reliabilities
Scoring system were determined using the intra-class correlation coefficient. Bland–Altman analysis was performed to
determine absolute reliabilities for both systems.
Correlation coefficients were 0.90 and 0.95 for intra-observer reliability and 0.83 and 0.89 for inter-
observer reliability for the existing and novel feather scoring systems, respectively. When using the novel
system, a change in plumage condition of P10% was needed to ensure that the change reflected a real
difference in 95% of cases, while a change of P15% was needed for the existing system. Since it may take
from 4 weeks (covert or down feathers) to over 1 year (flight or tail feathers) for feathers to regrow, suf-
ficient time should be allowed to elapse between two scoring sessions to reliably evaluate the efficacy of
preventive or therapeutic interventions for feather damaging behaviour.
Ó 2012 Elsevier Ltd. All rights reserved.

Introduction
Feather damaging behaviour (also referred to as feather picking
or plucking) is one of the most common behavioural problems in
captive parrots, with an estimated prevalence of 10% (Grindlinger
and Ramsay, 1991; Gaskins and Bergman, 2011). Feather damaging
behaviour includes plucking, chewing, fraying and/or biting,
resulting in loss of or damage to the feathers. Contour and down
feathers (which are often plucked or pulled) are usually the pri-
mary targets, but tail and flight feathers (which are often chewed)
may also be affected (Nett and Tully, 2003). Feather damage usu-
ally occurs in the readily accessible regions of the neck, chest, flank,
inner thigh and wing web (Harrison, 1986; Van Hoek and King,
1997; Nett and Tully, 2003), while the feathers on the parrot’s head
remain intact (Harrison, 1986).
A variety of causes of feather damaging behaviour have been
suggested, including socio-environmental, neurobiological, genetic
⇑ Corresponding author. Tel.: +31 30 2534542.
E-mail address: Y.R.A.vanZeeland@uu.nl (Y.R.A. van Zeeland).
and medical causes (van Zeeland et al., 2009). Prior to addressing
behaviour and environment, ‘underlying’ medical problems should
be ruled out (van Zeeland et al., 2009). Management of (non-med-
ical) feather damaging behaviour is challenging. Many treatment
modalities have been proposed, including behavioural modifica-
tion (Davis, 1995), environmental enrichment with toys and forag-
ing activities (Meehan et al., 2003; Lumeij and Hommers, 2008),
pharmacological intervention (Ramsay and Grindlinger, 1994;
Seibert et al., 2004; Seibert, 2007) and local application of foul
tasting substances and/or collars (Rosskopf and Woerpel, 1996;
Kaleta, 2003).
It is necessary to assess changes in feather damaging behaviour
over time to evaluate the efficacy of preventive or therapeutic
interventions. Ramsay and Grindlinger (1994) quantified the time
spent on feather damaging by direct behavioural observation, but
this method does not appear to be reliable, since it is difficult to
distinguish between preening and feather damaging behaviour
(Van Hoek and King, 1997). Furthermore, bouts of feather damag-
ing behaviour may not be observed clinically, since they often take
place at night (Meehan et al., 2003). Feather scoring systems,
which measure feather damaging behaviour indirectly by assessing

1090-0233/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.tvjl.2012.08.020
248 Y.R.A. van Zeeland et al. / The Veterinary Journal 196 (2013) 247–252

Table 1
Ten-point feather scoring assessment (Meehan et al., 2003).

Score Description
(a) Scoring system used for chest/flank, back and legs
0 All or most feathers removed, down removed and skin exposed, evidence of skin or tissue injury
0.25 All or most feathers removed, down removed and skin exposed, no evidence of skin or tissue injury
0.5 All or most feathers removed, some down removed, patches of skin exposed
0.75 All or most feathers removed, down exposed and intact or feathers removed from more than half of the area, some down removed, patches of skin exposed
1.0 Feathers removed from less than half of the area, some down removed and skin exposed
1.25 Feathers removed from more than half of the area, down exposed and intact
1.5 Feathers removed from less than half of the area, down exposed and intact
1.75 Feathers intact with fraying or breakage
2.0 Feathers intact with little or no fraying or breakage
(b) Scoring system used for wings
0 All or most primary feathers, secondary feathers and coverts removed, down removed, skin exposed, evidence of skin or tissue injury
0.5 All or most primary feathers, secondary feathers and coverts removed, down removed, skin exposed, no evidence of injury
1.0 More than half of coverts removed, down exposed and intact or more than half of primary and secondary feathers removed, down exposed and intact
1.5 Fewer than half of coverts removed, down exposed and intact or fewer than half of primary and secondary feathers removed, down exposed and intact or
primary and secondary feathers intact with significant breakage and fraying
2.0 Feathers intact with little or no fraying or breakage
(c) Scoring system used for tail
0 All or most tail feathers removed or broken
1 Some tail feathers removed or broken or significant fraying of tail feathers
2 Feathers intact with little or no fraying or breakage

Total score (0–10) = (chest/flank) + (back) + (legs) + (wings) + (tail).

plumage condition, may be used as an alternative to direct obser-
vations. Similar systems have been used in feather pecking laying
hens (Bilcik and Keeling, 1999; Kjaer et al., 2011).
For any test that involves scoring by humans, there is a possibil-
ity that bias or error related to human judgement is introduced.
This error may be present within a single person performing
multiple ratings of a subject and/or between people scoring the
same subject (Bruton et al., 2000). Prior to using a scoring method,
it should be evaluated for inter- and intra-observer reliabilities
(Bland and Altman, 1986).
Meehan et al. (2003) developed a 10-point feather scoring
system for evaluation of feather damaging behaviour in parrots (Ta-
ble 1); using the Pearson correlation coefficient, the inter-observer
reliability for two observers was >0.76. However, this coefficient
only considers relative position (i.e. linear correlation between rat-
ings) rather than agreement in specific segments or absolute agree-
ment of ratings (Keating and Matyas, 1998). Several drawbacks
have also been encountered when using the scoring system of Mee-
han et al. (2003): (1) the use of descriptive terminology increases
the degree of subjective interpretation by the observer; (2) the lim-
ited, incomplete, range of descriptions may result in choosing a
suboptimal alternative; (3) the commonly involved ventral surfaces
of the wings are not assessed; (4) the relative size of each body part
is omitted, and (5) no distinction is made between the types of
damage to the feathers, although a different time frame is needed
for replacement of different feather groups. In the present study,
a novel feather scoring system was developed and evaluated by
comparison with the scoring system of Meehan et al. (2003).
Materials and methods
Observers
A total of 35 observers with a veterinary background were recruited from Eur-
ope and North America. The observers were either avian veterinarians examining at
least one psittacine bird per week for a period >1 year (experienced in evaluating
plumage; n = 16) or fourth-year veterinary students (inexperienced in evaluating
plumage; n = 19). None of the observers had prior experience in using either feather
scoring system.
Feather scoring systems
The novel feather scoring system consists of two parts: (1) Part A assesses the
covert and down feathers, and (2) Part B assesses the flight feathers; for both parts,
a maximum score of 100 can be obtained (Table 2).
Birds for assessment
Twenty-five privately-owned grey parrots (Psittacus erithacus erithacus) with
various degrees of feather damaging behaviour, ranging from almost perfectly
feathered to almost completely bald, were selected at random from the patient pop-
ulation of the Utrecht University Veterinary Teaching Hospital. All birds were pre-
sented to the clinic for consultation regarding feather damaging behaviour, which
included a behavioural assessment and a routine physical examination to assess
the bird’s plumage condition and to exclude underlying medical problems.
During the physical examination, standardised photographs were taken of each
body part to be assessed: (1) chest and neck (front); (2) back; (3) ventral surfaces of
the left and right wings in fully extended position; (4) dorsal surfaces of the left and
right wings in fully extended position; (5) front view of the left and right legs in ex-
tended position, and (6) tail with feathers spread.
To prevent bias from recognition, pictures were trimmed to show only the body
part of interest (without background detail) and subsequently mixed and ordered at
random (using a random sequence generator programme1) to create a set of pic-
tures representing a complete parrot.
Study design
At least 7 days prior to the start of the study, the examiners were asked to as-
sess photographs of three parrots (not included in the trial) to familiarise them-
selves with the scoring procedure. For the assessment of intra-observer
reliability, nine students and nine veterinarians were randomly selected from the
groups using a random sequence generator programme1. These examiners were
asked to score the plumage condition of 10 parrots twice using both scoring systems.
Photographs of the various body parts were distributed in batches of five parrots at a
time. A minimum of 7 days elapsed between the first and second scoring dates to en-
sure that enough time had elapsed to minimise the influence of memory (Streiner and
Norman, 2003). Of the observers participating in the intra-observer trial, three veter-
inarians and four students scored only 5/10 parrots twice.
To determine inter-observer reliability, the other 17 examiners used both scor-
ing systems for a single assessment of the plumage condition of 15 parrots. Photo-
graphs of the parrots were distributed in batches of five parrots at a time, with an
interval of 7 days between the first and second scoring dates.
Scoring sessions were offered digitally to the examiners in an on-line survey
(Lime-survey, Carsten Schmitz2). Depending on the body part that was involved,
1–4 multiple choice questions were used to assess the plumage condition with both
scoring systems, resulting in a total of 20 questions for each set of photographs (see
Appendix A: Supplementary material 1). Based on the answers of the examiners, the
corresponding scores were calculated for each body part and added up to a total score
for each individual parrot.

The feather scoring system of Meehan et al. (2003) involves a 10-point scoring
1
system for assessment of the feather condition of five separate body areas, i.e. See: http://www.random.org.
2
chest/flank, back, legs, tail and wings (Table 1). See: http://docs.limesurvey.org/.
 Y.R.A. van Zeeland et al. / The Veterinary Journal 196 (2013) 247–252 249

Table 2
Novel feather scoring system.
(1) Score determination table for coverts and down feathers; used for chest/neck/flank, back, legs, dorsal and ventral surface of the wings.

Coverts Down feathers

No down removed <50% of down removed >50% of down removed All down removed
All coverts intact 100 85 70 60
Fraying or breakage of feathers 95 80 65 55
<25% of coverts removed 90 75 60 50
25–50% of coverts removed 80 65 50 40
50–75% of coverts removed 70 55 40 30
75–90% of coverts removed 60 45 30 20
>90% of coverts removed 50 35 20 10
The percentage of damage to the covert and down feathers is assessed for each body part separately.
Deduct 10 points from the score if skin damage is present.
Total body plumage score (0–100) = 0.25  chest/flank + 0.17  back + 0.10  legs + 0.28  dorsal wings + 0.20  (ventral wings).
To determine the total body plumage score, the scores for each body part are corrected for their relative body surface percentage, similar to scoring systems used in
human burn victims (Lund and Browder, 1944). These percentages (expressed as percentage of the total body surface area excluding the surface area of the head and
unfeathered parts of the legs) were determined in six grey parrots according to Mitchell (1929). Mean (± standard deviation) values for the various body parts were
25 ± 1.2% (chest/neck/flank), 17 ± 1.5% (back), 10 ± 1.2% (legs), 28 ± 2.2% (dorsal surface of the wings, up to the level of the tertiary feathers) and 20 ± 1.9% (ventral
surface of the wings, up to the level of the tertiary feathers).
(2) Score determination for flight feathers; used for tail feathers and primary and secondary feathers (wing feathers).

Score Description
0 Flight feather with signs of fraying and/or breakage over >50% of the original length
1 Flight feather with signs of fraying and/or breakage over <50% of the original length
2 Flight feather with little or no damage present
Damage to individual flight feathers is assessed.
Total flight feather score (0–100) = (primary + secondary feathers left wing) + (primary + secondary feathers right wing) + (tail feathers).
The maximum score is dependent on total number of flight feathers of the bird. In general, each wing has 10 primary feathers and 10 secondary feathers (remiges),
whereas the tail has 10–12 flight feathers (rectrices). As each individual flight feather is awarded a score from 0–2, the score will range from 0–40 for each wing and
from 0–20 (or 0–24 in the case of 12 tail feathers) for the tail, respectively.

No questions were included on the status of the flight and tail feathers (needed
to determine the score for part B of the novel scoring system), since it was impos-
sible to spread the wings and tail sufficiently to allow accurate assessment of the
individual flight and tail feathers from the photographs.
Statistical analysis
Statistical analyses were performed using SPSS version 16.0. The relative inter-
and intra-observer reliabilities of each scoring system were assessed using the in-
tra-class correlation coefficient (ICC). A two-way random effects model for single
measurements (ICC[2,1]) was used to calculate the ICC and corresponding 95% con-
fidence intervals (CIs) for the total plumage scores (Fleiss and Cohen, 1973; Shrout
and Fleiss, 1979). For ICC, values range from 0 to 1, with a value of ‘0’ representing
no agreement and a value of ‘1’ representing perfect agreement. In general, values
>0.75 and >0.90 indicate good and excellent agreement, respectively (Fleiss, 1981;
Portney and Watkins, 2000).
Bland–Altman analysis was performed to determine absolute reliability (Bland
and Altman, 1996; Weir, 2005). The inter-observer standard error of measurement
(SEM) was calculated using the scores of all examiners simultaneously. The intra-
observer SEM was determined as the mean (± standard deviation) of the individual
SEMs (which were calculated using the repeated scores of 10 parrots for each exam-
iner separately). Subsequently, the SEM was used to determine the minimum real
p
difference, with a reliability of 95% (SEM  1.96  2), for qualitative interpretation
of the results (Weir, 2005). Any change in a parrot’s plumage score above or below
the previous plumage score that was greater than this minimum difference was
considered to reflect a real difference in 95% of the cases.
An independent samples t test was conducted to compare the relative inter- and
intra-observer reliabilities obtained with each scoring system. To compare absolute
inter- and intra-observer reliabilities (SEM, minimum difference) an F test was per-
formed. Normality of distribution for each parameter was evaluated using the Shap-
iro–Wilk test. Standard deviations, necessary for the t test, were calculated
according to the method described by Kenney and Keeping (1962). P values <0.05
were considered to be statistically significant. For ‘post hoc’ comparisons of the
scoring methods within the two subgroups, a Bonferroni correction was performed,
leading to an adjusted significance level of 0.025.
Results
Relative reliability and inter- and intra-observer reliabilities
ICC[2,1] was 0.90 for Meehan’s scoring system and 0.95 for the
novel scoring system. These values indicate excellent intra-observer
reliability (>0.90). The novel scoring system had significantly higher
reliability, especially when comparing the consistency in scoring for
the group of veterinarians who were experienced in examining
avian plumage (P < 0.01; Table 3). Both systems had overall good in-
ter-observer reliability (Meehan’s system 0.83; novel scoring sys-
tem 0.89) and there were no significant differences between the
two scoring systems (Table 4). No significant differences in intra-
or inter-observer reliabilities were found between the two observer
groups for either scoring system (Tables 3 and 4).
Absolute reliability, measurement error and minimum change
The SEM for individual observers ranged from 0.19 to 0.89
(mean 0.53 ± 0.19) on a scale of 0–10 for Meehan’s scoring system
and from 1.5 to 8.4 (mean 3.6 ± 1.7) on a scale of 0–100 for the no-
vel scoring system (Table 3). The minimum real difference (i.e. the
minimum difference between two measurements that reflects a
real difference at a confidence level of 95%) was 1.5 ± 0.5 points
(14.7 ± 5.3%) on the 10-point scale for the scoring system of
Meehan. For the novel scoring system, the absolute intra-observer
reliability was significantly higher (9.9 ± 4.7 points on a scale of
0–100; P < 0.01; Table 3; Fig. 1).
The SEM for inter-observer reliability was 0.72 for the 10-point
scale of Meehan’s scoring system and 5.8 for the 100-point scale of
the novel scoring system. There was no significant difference
between the absolute inter-observer reliability of the novel scoring
system (13.5%) and Meehan’s scoring system (17%; Table 4).
Discussion
The inter- and intra-observer reliabilities of both scoring
systems in our study were >0.75 and >0.90, respectively, which is
considered to reflect good and excellent agreement (Fleiss, 1981;
Portney and Watkins, 2000). Similar to other studies (Leggin
et al., 1996; Sokoloff and Blumberg, 2002), intra-observer reliabil-
250 Y.R.A. van Zeeland et al. / The Veterinary Journal 196 (2013) 247–252

Table 3
Relative and absolute intra-observer reliabilities (mean ± standard deviation) for the total feather score using Meehan’s feather scoring system and the novel feather scoring
system.

Observer Feather scoring system of Meehan et al. (2003) Novel feather scoring system
ICC[2, 1] (95% CI) SEM (0–10) Minimum difference (%) ICC[2, 1] (95% CI) SEM (0–100) Minimum difference (%)
Inexperienced observers (n = 9) 0.92 (0.88–0.95) 0.51 ± 0.24 14.0 ± 6.7 0.94 (0.90–0.98) 4.0 ± 2.1 11.0 ± 5.8
a a a
Experienced observers (n = 9) 0.89 (0.84–0.93) 0.56 ± 0.16 15.5 ± 4.3 0.96 (0.94–0.98) 3.2 ± 1.2 8.9 ± 3.3
a a a
All observers (n = 18) 0.90 (0.87–0.93) 0.53 ± 0.19 14.7 ± 5.3 0.95 (0.93–0.97) 3.6 ± 1.7 9.9 ± 4.7

ICC[2,1], Intra-class correlation coefficient; SEM, Standard error of measurement; CI, Confidence interval.
a
Significant difference (P <0.01) from the feather scoring system of Meehan et al. (2003).

Table 4
Relative and absolute inter-observer reliabilities for the total feather score using Meehan’s scoring system (Meehan et al., 2003) and the novel feather scoring system.

Observers Feather scoring system of Meehan et al. (2003) Novel feather scoring system
ICC[2, 1] (95% CI) SEM (0–10) Minimum difference (%) ICC[2, 1] (95% CI) SEM (0–100) Minimum difference (%)
Inexperienced observers (n = 10) 0.82 (0.67–0.93) 0.70 16 0.88 (0.77–0.95) 5.86 13.6
Experienced observers (n = 7) 0.82 (0.65–0.93) 0.76 18 0.91 (0.81–0.96) 5.71 13.3
All observers (n = 17) 0.83 (0.70–0.93) 0.72 17 0.89 (0.80–0.95) 5.82 13.5

ICC[2,1], Intra-class correlation coefficient; SEM, Standard error of measurement; CI, Confidence interval.

Fig. 1. Intra-observer data for (A) Meehan’s feather scoring system (Meehan et al., 2003) and (B) the novel scoring system. The first and second score for each individual
parrot (n = 10) by the different observers (n = 18) are plotted against each other to depict intra-observer reliability for experienced observers (dots) and inexperienced
observers (asterisks). The graph furthermore depicts the ideal correlation between the two measurements (x = y; continuous line) and the minimum difference considered to
be real in 95% of cases (interrupted lines).

ities were higher than inter-observer reliabilities, indicating that
the highest level of reliability can be found when consecutive scor-
ings of a bird are performed by the same examiner. The signifi-
cantly higher intra-observer reliability obtained with the novel
scoring system may be linked to the use of objectively definable
criteria to quantify the amount of feather damage compared to
the more subjective terminology used in Meehan’s scoring system.
However, a high ICC does not always imply an acceptable reli-
ability (Costa-Santos et al., 2011). When large ranges in scores
are present, a high ICC may be obtained, while actual scores display
large within-subject differences. In addition, factors such as the de-
sign of the study, variance within the sampled population and the
number of observers and/or subjects may influence the height of
the ICC, limiting the ability to extrapolate the findings more widely
(Müller and Büttner, 1994; Walter et al., 1998; Saito et al., 2006).
The complementary use of an absolute measure of reliability, e.g.
the coefficient of variation (Portney and Watkins, 1993) or limits
of reliability (Bland and Altman, 1986), is therefore recommended.
In our study, a Bland–Altman analysis was performed to deter-
mine the SEM and the minimum real difference in 95% of cases. The
smallest minimum real differences (i.e. 9.9 ± 4.7 points or 10 ± 4.7%
on a scale of 0–100) were achieved on consecutive scoring by the
same observer using the novel scoring system. On the basis of
our experience, this minimum difference of 10% is considered to
be clinically important, rendering the scoring system suitable for
use in practice. Using Meehan’s scoring system, a significantly
greater difference between scores (at least 15 ± 5.3% instead of
10 ± 4.7%) is necessary to ascertain that this difference is real in
95% of cases. Therefore, the novel scoring system may be able to
detect real changes at an earlier stage than Meehan’s system. The
mean values obtained for each system may be used as a guide
for interpretation of scores in clinical practice. However, since
 Y.R.A. van Zeeland et al. / The Veterinary Journal 196 (2013) 247–252
the minimum real difference varies greatly between observers, it is
advisable that the minimum real difference is assessed for each
individual observer.
The minimum real differences for both scoring systems indicate
that slight changes in feather damaging behaviour (<10% for the
novel scoring system and <15% for Meehan’s scoring system) can-
not be detected reliably. Since it may take from 4 weeks (covert or
down feathers) (Wang, 1943) to >1 year (flight or tail feathers)
(Juniper and Parr, 1998) for feathers to regrow, sufficient time
should elapse between two scoring sessions to evaluate the
efficacy of preventive or therapeutic interventions for feather
damaging behaviour. Further studies are needed to determine
whether smaller differences can be detected by direct evaluation
of the parrot’s plumage condition instead of using photographs.
In addition to its higher relative and absolute intra-observer reli-
abilities, the novel scoring system may also result in a more accu-
rate representation of the birds plumage condition due to (1) the
assessment of feather damage in percentages; (2) the incorporation
of relative body surface area of the affected body parts (a method-
ology that is routinely used in human burn victims); and (3) the use
of a separate system to address damage to the flight and tail feath-
ers, which usually involves a different type of damage (chewing or
fraying) and a different time frame (until the next moult) to restore
damage (Nett and Tully, 2003). Thus, when high accuracy and/or
high consistency of the observations are needed, the novel scoring
system may be preferable to the system of Meehan.
However, the choice for a specific scoring system may also de-
pend on other factors, such as simplicity of the system or the time
needed to complete the scoring, which are advantages of Meehan’s
scoring system (although scoring time may be decreased consider-
ably for the novel scoring system when using an automated pro-
gramme to calculate the scores; see Appendix A: Supplementary
file 2). The choice may also depend on the conditions under which
the scoring can be performed or the purpose for which the scoring
is used. The scoring system of Meehan is ideal for evaluation of the
plumage condition of a bird from a distance (e.g. in the wild or in
an aviary), which hinders the accurate estimation of percentage of
feathers damaged, as well as assessment of the ventral surface of
the wings. Conversely, the novel scoring system is designed to en-
able the feather condition to be evaluated either by using photo-
graphs or by direct visualisation while the bird is manually
restrained during clinical examination.
Due to the inability to accurately assess the individual flight
feathers from the photographs, evaluation of damage to the pri-
mary, secondary and tail feathers with the novel scoring system
was excluded in this study, thereby omitting the reliability of this
part of the scoring system. Direct visualisation of a manually re-
strained bird would be necessary to allow proper evaluation of
the individual flight and tail feathers, but this was not feasible in
the present study due to the use of assessors at different institutes.
Although the novel scoring system has not been fully evaluated,
the assessment of flight and tail feathers is relatively straightfor-
ward (i.e. no damage, <50% of the feather damaged or P50% of
the feather damaged). Therefore no marked influence on the out-
come of the present study would be expected by inclusion of these
feathers in the analysis. Furthermore, the tail and flight feathers do
not need to be routinely assessed when evaluating feather damag-
ing behaviour, since the damage inflicted to these feathers (chew-
ing, fraying) does not resolve until the next moult and therefore is
not useful for evaluating short-term changes in behaviour.
Conclusions
In a comparison of a novel feather scoring system with the
existing (Meehan) system, there was good to excellent agreement
251
between total feather scores within and between observers for
both scoring systems. The relative and absolute intra-observer reli-
abilities were highest for the novel system, demonstrating its value
for monitoring feather damaging behaviour over time. Since both
systems had an overall good consistency, the preference for a cer-
tain scoring system will be based on additional factors, such as the
clinical setting in which a parrot’s plumage condition is assessed.
Since intra-observer reliability was higher than inter-observer reli-
ability for both scoring systems, it is recommended that individual
birds should be assessed by the same observer.
Conflict of interest statement
None of the authors has any financial or personal relationships
that could inappropriately influence or bias the content of the
paper.
Acknowledgements
The authors would like to thank all the participants for their
time and contribution to the study, J. Fama and the veterinary tech-
nicians of the Division of Zoological Medicine of Utrecht University
for assistance with taking the photographs of the parrots and Dr
J.C.M. Vernooij, Dr G.P.A. Bergers and Mrs M. de Kam for help with
statistical analysis. This study was made possible with the help of a
grant received from the Dutch Ministry of Economic affairs, Agri-
culture and Innovation (Grant No. 1400001780).
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.tvjl.2012.08.020.
References
Bilcik, B., Keeling, I.J., 1999. Changes in feather condition in relation to feather
pecking and aggressive behaviour in laying hens. British Poultry Science 40,
444–451.
Bland, J.M., Altman, D.G., 1986. Statistical methods for assessing agreement
between two methods of clinical measurement. Lancet 327, 307–310.
Bland, J.M., Altman, D.G., 1996. Measurement error. British Medical Journal 313,
744.
Bruton, A., Conway, J.H., Holgate, S.T., 2000. Reliability: What is it, and how is it
measured? Physiotherapy 86, 94–99.
Costa-Santos, C., Bernardes, J., Ayres de Campos, D., Costa, A., Costa, C., 2011. The
limits of agreement and the intraclass correlation coefficient may be
inconsistent in the interpretation of agreement. Journal of Clinical
Epidemiology 64, 264–269.
Davis, C., 1995. Behavior modification counselling – An alliance between
veterinarian and behaviour consultant. Seminars in Avian and Exotic Pet
Medicine 4, 39–42.
Fleiss, J.L., Cohen, J., 1973. The equivalence of weighted kappa and the intraclass
correlation coefficient as measures of reliability. Educational and Psychological
Measurement 33, 613–619.
Fleiss, J.L., 1981. Statistical Methods for Rates and Proportions, Second Ed. John
Wiley, New York, USA, 352 pp.
Gaskins, L.A., Bergman, L., 2011. Surveys of avian practitioners and pet owners
regarding common behavior problems in psittacine birds. Journal of Avian
Medicine and Surgery 25, 111–118.
Grindlinger, H.M., Ramsay, E., 1991. Compulsive feather picking in birds. Archives of
General Psychiatry 48, 857.
Harrison, G.J., 1986. Disorders of the integument. In: Ritchie, B.W., Harrison, G.J.,
Harrison, L.R. (Eds.), Avian Medicine and Surgery: Principles and Application.
W.B. Saunders, Philadelphia, Pennsylvania, USA, pp. 509–524.
Juniper, T., Parr, M., 1998. Guide to Parrots of the World. Pica Press, Robertsbridge,
East Sussex, UK, 584 pp.
Kaleta, E.F., 2003. Verhalten und verhaltensstorungen. In: Kaleta, E.F., Krautwald-
Junghanns, M.E. (Eds.), Kompendium der Ziervogelkrankheiten. Schlutersche,
Hannover, Germany, pp. 41–45.
Keating, J., Matyas, T., 1998. Unreliable inferences from reliable measurements.
Australian Journal of Physiotherapy 44, 5–10.
Kenney, J.F., Keeping, E.S., 1962. Chapter 6.5: The standard deviation; Chapter 6.6:
Calculation of the standard deviation. In: Kenny, J.F. (Ed.), Mathematics of
Statistics, Pt. 1, Third Ed. Van Nostrand, Princeton, New Jersey, USA, pp. 77–80.
252 Y.R.A. van Zeeland et al. / The Veterinary Journal 196 (2013) 247–252
Kjaer, J.B., Glawatz, H., Scholz, B., Rettenbacher, S., Tauson, R., 2011. Reducing stress
during welfare inspection: Validation of a non-intrusive version of the LayWel
plumage scoring system for laying hens. British Poultry Science 52, 149–154.
Leggin, B.G., Neumann, R.M., Iannotti, J.P., Williams, G.R., Thompson, E.C., 1996.
Intrarater and interrater reliability of three isometric dynamometers in
assessing shoulder strength. Journal of Shoulder and Elbow Surgery 5, 18–24.
Lumeij, J.T., Hommers, C.J., 2008. Foraging ‘enrichment’ as treatment for
pterotillomania. Applied Animal Behaviour Science 111, 85–94.
Lund, C.C., Browder, N.C., 1944. The estimation of areas of burns. Surgery,
Gynecology and Obstetrics 79, 352–358.
Meehan, C.L., Millam, J.R., Mench, J.A., 2003. Foraging opportunity and increased
physical complexity both prevent and reduce psychogenic feather picking by
young Amazon parrots. Applied Animal Behaviour Science 80, 71–85.
Mitchell, H.H., 1929. The surface area of single comb white leghorn chickens.
Journal of Nutrition 5, 443–449.
Müller, R., Büttner, P., 1994. A critical discussion of intraclass correlation
coefficients. Statistics in Medicine 13, 2465–2476.
Nett, C.S., Tully, T.N., 2003. Anatomy, clinical presentation and diagnostic approach
to feather-picking pet birds. Compendium on Continuing Education for the
Practicing Veterinarian 25, 206–219.
Portney, L.G., Watkins, M.P., 1993. Foundations of Clinical Research: Applications to
Practice. Appleton and Lange, Norwalk, Connecticut, USA, 722 pp.
Portney, L.G., Watkins, M.P., 2000. Foundations of Clinical Research: Applications to
Practice, Second Ed. Prentice Hall Health, Upper Saddle River, New Jersey, USA,
742 pp.
Ramsay, E.C., Grindlinger, H., 1994. Use of clomipramine in the treatment of
obsessive behavior in psittacine birds. Journal of the Association of Avian
Veterinarians 8, 9–15.
Rosskopf, W.J., Woerpel, R.W., 1996. Feather picking and therapy of skin and feather
disorders. In: Rosskopf, W.J., Woerpel, R.W. (Eds.), Diseases of Cage and Aviary
birds, Third Ed. Williams and Wilkins, Baltimore, Maryland, USA, pp. 397–405.
Saito, Y., Sozu, T., Hamada, C., Yoshimura, I., 2006. Effective number of subjects and
number of raters for inter-rater reliability studies. Statistics in Medicine 25,
1547–1560.
Seibert, L.M., 2007. Pharmacotherapy for behavioural disorders in pet birds. Journal
of Exotic Pet Medicine 16, 30–37.
Seibert, L.M., Crowell-Davis, S.L., Wilson, G.H., Ritchie, B.W., 2004. Placebo-
controlled clomipramine trial for the treatment of feather-picking disorder in
cockatoos. Journal of the American Animal Hospital Association 40, 261–269.
Shrout, P.E., Fleiss, J.L., 1979. Intraclass correlations: Uses in assessing rater
reliability. Psychological Bulletin 86, 420–428.
Sokoloff, G., Blumberg, M.S., 2002. Contributions of endothermy to huddling
behavior in infant Norway rats (Rattus norvegicus) and Syrian gold hamsters
(Mesocricetus auratus). Journal of Comparative Psychology 116, 240–246.
Streiner, D.L., Norman, G.R., 2003. Health Measurement Scales: A Practical Guide to
their Development and Use. Oxford University Press, New York, USA, 296 pp.
Van Hoek, C.S., King, C.E., 1997. Causation and influence of environmental
enrichment on feather picking of the crimson-bellied conure (Pyrrhura perlata
perlata). Zoo Biology 16, 161–172.
Van Zeeland, Y.R.A., Spruijt, B.M., Rodenburg, T.B., Riedstra, B., van Hierden, Y.M.,
Buitenhuis, B., Korte, S.M., Lumeij, J.T., 2009. Feather damaging behaviour in
parrots: A review with consideration of comparative aspects. Applied Animal
Behaviour Science 121, 75–95.
Walter, S.D., Eliasziw, M., Donner, A., 1998. Sample size and optimal designs for
reliability studies. Statistics in Medicine 17, 101–110.
Wang, H., 1943. The morphogenetic functions of the epidermal and dermal
components of the papilla in feather regeneration. Physiological Zoology 16,
325–349.
Weir, J.P., 2005. Quantifying test-retest reliability using the intraclass correlation
coefficient and the SEM. Journal of Strength and Conditioning Research 19, 231–
240.