Content uploaded by Jaimie F. Veale
Author content
All content in this area was uploaded by Jaimie F. Veale on Jun 27, 2014
Content may be subject to copyright.
_____________________________
Correspondence: Department of Psychology, University of British Columbia, 2136 West
Mall, Vancouver, BC, V6T 1Z4, Canada; email: jaimie@jaimieveale.com
This is an Accepted Manuscript of an article published in Laterality: Asymmetries of Body, Brain and Cognition 2014 19(2),
164177, available online: http://www.tandfonline.com/10.1080/1357650X.2013.783045
EDINBURGH HANDEDNESS INVENTORY—SHORT FORM: A
REVISED VERSION BASED ON CONFIRMATORY FACTOR
ANALYSIS
Jaimie F. Veale
Department of Psychology, University of British Columbia.
ABSTRACT
While the Edinburgh Handedness Inventory has been widely used, there have been few
studies assessing its factorial validity. There is evidence that the original instructions and
response options are difficult to understand. Using simplified instructions and response
options, the Edinburgh Handedness Inventory was administered on a sample of 1514
participants using an online questionnaire. In accordance with previous research, a model of
the 10item inventory had poor fit for the data. This study also detected model
misspecification in the previouslyproposed 7item modification. A 4item Edinburgh
Handedness Inventory—short form had good model fit with items modeled as both
continuous and ordinal. Despite its brevity, it showed very good reliability, factor score
determinacy, and correlation with scores on the 10item inventory. By eliminating items that
were modeled with considerable measurement error, the short form alleviates the concern of
the 10item inventory overcategorizing mixed handers. Evidence was found for factorial
invariance across level of education, age groups, and regions (USA and Australia/New
Zealand). There generally appeared to be invariance across genders for the 4item inventory.
The proposed Edinburgh Handedness Inventory—short form measures a single handedness
factor with an inventory that has brief and simple instructions and a small number of items.
KEY WORDS: handedness, Edinburgh Handedness Inventory, confirmatory factor analysis,
invariance testing, psychometric properties
Jaimie Veale
2
Handedness, or the preference to use one hand more than the other, is usually
measured by questionnaire in the social sciences. The Edinburgh Handedness Inventory
(Oldfield, 1971) is the most widely used of these questionnaires (Fazio, Coenen, & Denney,
2012). Oldfield’s 1971 article currently has almost 11000 citations in Scopus. To complete
the Edinburgh Handedness Inventory, respondents endorse hand preference for ten everyday
tasks. In the original inventory, participants are given a list of the tasks with adjacent “left”
or “right” columns. They are asked write “+” in the appropriate corresponding column. If the
preference is so strong that they “would never try to use the other hand unless absolutely
forced to” they are to write “++” instead. If they are “really indifferent” they are instructed to
put an “+” in both columns (Oldfield, 1971, p. 112).
Despite its widespread usage, there have been few studies assessing the factorial
validity of the Edinburgh Handedness Inventory. Such assessment is important because it
tests that the inventory measures the unidimensional handedness factor it purports to. Studies
that have been conducted have generally noted that some of the questionnaire items are
problematic (Bryden, 1977; Dragovic, 2004; McFarland & Anderson, 1980; S. M. Williams,
1986). From exploratory factor analysis, Bryden found three items (opening box, using
broom, and striking a match) loaded onto a single factor “in which there was considerable
disagreement among righthanders as to which hand was generally used, and items in which
the subject had to think very carefully before giving an answer” (p. 621). Also from
exploratory factor analysis, McFarland and Anderson noted “poor” loading of the knife,
broom, and box lid items on a single handedness factor. They reported evidence for factor
loading stability across age and gender, and testretest for most of the items, but noted
instability on the box, broom, match, and scissors item. However, McFarland and Anderson
did not test whether allowing these factor loadings to differ between groups significantly
improved the model fit and they did not test the equality of item intercepts or error variances
which are requirements for testing whether findings of group mean differences can be
accounted for by measurement biases (Gregorich, 2006).
Confirmatory factor analysis is a powerful tool for assessing the factorial validity of a
questionnaire. Advantages of confirmatory factor analysis are that it allows modeling of
error variance, testing for item uniqueness, and testing for acceptable fit of the factor
structure (Brown, 2006). Two studies have tested the Edinburgh Handedness Inventory using
confirmatory factor analysis. From a community sample of 203, Dragovic (2004) found it
was redundant to include both the writing and drawing items, as these appeared to be almost
perfectly collinear, and that the broom and boxlid were modeled with a high proportion of
Edinburgh Handedness Inventory—Short Form
3
error variance. He concluded that this high error variance was due to these items measuring
another factor or ambiguity of item interpretation. Dragovic proposed a 7item revised
inventory with these three items removed. Milenkovic and Dragovic (2012) replicated these
findings on a sample of 1224 high school students.
There were limitations of these previous two confirmatory factor analysis studies.
These studies used the maximum likelihood estimation technique that assumes indicators are
continuous. However, with 5point Likert response options, an estimation technique for
orderedcategorical data is also appropriate (Kline, 2011). These two previous studies
administered the Edinburgh Handedness Inventory using the original instructions. These
instructions have been criticized as “somewhat lengthy and confusing” (Fazio et al., 2012, p.
71) and there is evidence that the majority of male inmates who were administered the
inventory did not understand or follow them correctly (Fazio et al., 2012). Fazio et al.
suggested a Likert scale adaptation of the Edinburgh Handedness Inventory could markedly
improve instruction adherence.
METHOD
Participants
Participants were recruited for an internetbased study examining the development of
gender identity and sexuality (Veale, Clarke, & Lomax, 2010) through lesbian, gay, bisexual,
and transgender related online forums and mailing lists, Google online advertising, and a
press release. The Edinburgh Handedness Inventory was completed by 1514 participants.
Participants’ demographics are outlined in Table 1. Age ranged from 16 to 81. A
significant proportion of participants had gender identities not consistent with their biological
sex (e.g. transsexual, transgender). For invariance testing between genders, 160 biological
males and 266 biological females who did not report having gendervariant identities were
used.
Measure
The Edinburgh Handedness Inventory (Oldfield, 1971) was administered with revised
instructions and response options. The instructions given were “Please indicate your
preferences in the use of hands in the following activities. Some of the activities require both
hands. In these cases the part of the task, or object, for which hand preference is wanted is
Jaimie Veale
4
indicated in brackets”. Response options given were “always right”, “usually right”, “both
equally”, “usually left”, and “always left”.
Table 1. Participants’ demographics
Ethnicity
%
Residence
%
Level of education
%
White/Caucasian
92
USA
60
3 years high school
7
East Asian
3
New Zealand
18
4 years high school
10
Hispanic/Latino
3
Great Britain
8
5 years high school
11
Indigenous American
3
Canada
5
Diploma/trade qualification
20
Black/African
2
Australia
2
Bachelor’s degree
31
South/other Asian
2
Other
7
Master’s degree
14
Other
2
Doctoral degree
5
Data Analysis
Confirmatory factor analysis of the Edinburgh Handedness Inventory was conducted
using Mplus software version 5.1 (Muthén & Muthén, 2008). Some authors have suggested
5point response scales be treated as ordinal (Kline, 2011), while others have suggested
treatments as continuous (Blunch, 2008). Analyses for this article were conducted using both
YuanBentler robust maximum likelihood estimation and mean and varianceadjusted
weighted least squares (WLSMV) estimation. The former estimates and fit indices are
adjusted to allow for missing data and variations of multivariate normality. The latter allows
for ordinal categories, missing data, and uses a polychoric correlation matrix (Kline, 2011).
Model fit was assessed using absolute fit indices: SatorraBentler χ² likelihood ratio
(SBχ²), RMSEA, and SRMR and relative fit indices: CFI and TLI. Model misspecification
was detected by a statistically significant p value less than .05 on the χ² test. CFI and TLI
values lower than .9, RMSEA values greater than .05 and SRMR values greater than .08 were
also used as indicators that the proposed model did not fit the data (Kline, 2011).
Reliability was estimated using Cronbach’s (1951) α and Raykov’s (1997) factor ρ
which is a composite reliability coefficient that is calculated as the ratio of variance explained
by the factor to the total variance. Factor score determinacy was also estimated by
calculating the squared multiple correlation of the proposed indicators for predicting the
handedness factor. This gives information about the extent to which the true factor score can
be determined in the model (Grice, 2001). This is useful because while confirmatory factor
Edinburgh Handedness Inventory—Short Form
5
analysis models with a small number of indicators may be more likely to fit the data, they are
also more likely to have factor indeterminacy (Brown, 2006).
In accordance with Milenkovic and Dragovic’s (2012) study, laterality quotients were
calculated. These ranged from 100 (left handed) to 100 (right handed). Participants with a
laterality quotient score between 60 and 60 (inclusive) were classified as mixed handers.
Testing for parallel indicators was also conducted to test whether each item measures the
handedness construct equally. Such testing is desirable to understand whether summing
scores on each item equally (as per the laterality quotient) will accurately estimate a person’s
handedness.
The inventory’s stability across a number of groups was also assessed using factorial
invariance testing. The majority of participants sampled in this thesis lived in two regions:
the USA (60%) or Australia and New Zealand (20%). Invariance testing was conducted
between these groups to test for differences in item meaning for participants living in these
regions. The median age of participants was 37. Invariance testing was conducted between
those above and below the median age. Finally, invariance testing was conducted between
males and females (participants with gender identities atypical to their biological sex are
removed from this analysis).
Metric invariance and scalar invariance were tested by comparing these to a model
where these invariance constraints were not imposed (the unconstrained model). A
statistically significant change in scaled difference χ² likelihood ratio test may indicate scale
noninvariance (Kline, 2011). However, because this test can be oversensitive when sample
size is large, Cheung and Rensvold’s (2002) criterion of a decrease in CFI of greater than .01
is evidence for measurement noninvariance was also used.
Data were missing for 1% of responses. Analyses were conducted using Mplus’s full
information maximum likelihood method (Asparouhov & Muthén, 2010).
RESULTS
Model Testing and Selection
As outlined in Tables 2 and 3, confirmatory factor analysis using all ten items of the
Edinburgh Handedness Inventory elicited similar concerns to previous studies. The model
had poor performance on all fit indices with the exception of the CFI and TLI using WLSMV
estimation. There was collinearity between the drawing and writing items, r = .97, and high
proportion of residual error for the broom (.64 using robust maximum likelihood, .44 using
Jaimie Veale
6
WLSMV estimation), knife (.54, .35), and lid opening (.54, .34) items. The covariance
matrix for the 10 items is given in supplementary electronic data. Dragovic’s proposed 7
item inventory, in which the drawing, broom, and box lid opening items are removed, had
signs of model misspecification on the χ² likelihood ratio using both estimation techniques
and also the RMSEA when using WLSMV estimation, so a revised model was also tested.
Table 2. Fit statistics, and reliability and factor score determinacy estimates for versions of
the Edinburgh Handedness Inventory with robust maximum likelihood estimation
Model
SBχ²
df
p
CFI
TLI
RMSEA
SRMR
ρ
α
FSD
10 items
1144.19
35
< .001
.84
.79
.145
.05
.95
.95
.98
7 items
59.95
14
< .001
.99
.98
.047
.02
.95
.95
.98
4 items
0.40
2
.819
1.00
1.00
.000
.00
.93
.93
.97
4 items, parallel
indicators
47.04
5
< .001
.98
.97
.075
.08
.93
.97
Note. N = 1514; ρ = Raykov’s composite reliability; FSD = factor score determinacy
Table 3. Fit statistics, and reliability and factor score determinacy estimates for versions of
the Edinburgh Handedness Inventory with WLSMV estimation
Model
χ²
df
p
CFI
TLI
RMSEA
10 items
271.87
21
< .001
.99
1.00
.089
7 items
103.67
14
< .001
1.00
1.00
.065
4 items
1.97
2
.373
1.00
1.00
.000
4 items, parallel indicators
158.53
4
< .001
.99
1.00
.160
Note. N = 1514
In selecting a revised model, the following were considered:
 As outlined above, previous research has noted problems with the knife, striking match,
and scissors items;
 although this past evidence is more limited for the knife and scissors items, the knife item
was modeled with significant error variance in the previous analysis and scissor use
preference may be effected by practice with the tools available (i.e. many modern scissors
are designed for righthanded use); and
 the match and knife items are the only remaining items pertaining to tasks which require
two hands—removing these will simplify the scale.
Edinburgh Handedness Inventory—Short Form
7
Given these considerations, the knife, striking match, and scissors items were
removed. Tables 2 and 3 show the resultant 4item model had adequate fit for the on all fit
tests using both estimation techniques. Constraining factor loadings to be equal amongst
indicators (parallel indicators) resulted in a model with poor fit using both estimation
techniques. Reliability and factor score determinacy estimates are also given in Table 2.
Parameter estimates for the 4item model are given in Table 4.
Table 4. Model estimates for the 4item Edinburgh Handedness Inventory
Parameter
Unstandardized
SE
Standardized
Unstandardized
SE
Standardized
MLR factor loadings
WLSMV factor loadings
Writing
1.00a
__
.89
1.00a
__
.97
Throwing
0.83
.03
.82
0.89
.01
.87
Toothbrush
0.94
.02
.88
0.95
.01
.92
Spoon
1.00
.02
.91
0.97
.01
.94
MLR factor variance
WLSMV factor variance
Handedness
32.29
2.23
1.00
0.95
.01
1.00
Note. SE = standard error. MLR = robust maximum likelihood. a This parameter is fixed
because this item is used as a marker variable, therefore it is not tested for statistical
significance. All other estimates were statistically significant p < .001.
Invariance Testing
Tables 5 and 6 outline the results of invariance testing results. These analyses showed
no significant change towards poorer model fit in the χ² likelihood ratio test. While the
change in this test for betweencountries metric invariance achieved statistical significance
(χ²(3) = 8.13, p = .04) when using WLSMV estimation, this represented an improvement in
model fit. Although not detected by the χ² likelihood ratio test, change in CFI for between
genders invariance testing was notable. This can be explained by constraining the toothbrush
item between genders resulting in a significant scaled difference χ²(1) = 4.08, p = .04,
suggesting toothbrush hand preference has a slightly higher factor loading in males.
Unfortunately, WLSMV betweengender invariance testing could not be conducted because
of a not positive definite correlation matrix, likely due to a low number of “both equally” and
“usually left” response options to the writing variable.
Jaimie Veale
8
Table 5. Invariance testing fit statistics for the 4item Edinburgh Handedness Inventory using
robust maximum likelihood estimation
Model
SBχ²
df
p
χ²SD
∆df
p
CFI
∆CFI
Country: USA/Australia or New Zealand
n = 1214
Unconstrained
2.14
4
.710



1.000

Metric invariance
2.82
7
.901
0.30
3
.960
1.000
.000
Scalar invariance
10.94
10
.362
4.88
6
.560
.999
.001
Age median split
N = 1514
Unconstrained
4.14
4
.388



1.000

Metric invariance
7.17
7
.412
1.24
3
.744
1.000
.000
Scalar invariance
15.54
10
.114
6.31
6
.389
.997
.003
Gender
n = 426
Unconstrained
6.48
4
.166



.994

Metric invariance
16.27
7
.023
4.61
3
.203
.977
.017
Scalar invariance
22.54
10
.013
8.45
6
.207
.968
.026
Level of education
n = 1446
Unconstrained
3.96
6
.683



1.000

Metric invariance
11.88
12
.455
3.21
6
.782
1.000
.000
Scalar invariance
19.91
18
.338
6.46
12
.891
.999
.001
Note. SD = scaled difference
Laterality Quotient
Figure 1 plots participants’ laterality quotient scores on the 10item inventory and the
4item short form. Participants with any missing data were excluded from this analysis. The
relationship between these two scales was Spearman’s ρ = .90 and r2 = .94.
Results of assignment to handedness categories are given in Table 7. The level
agreement of this categorization between the 10item and 4item inventories was kappa = .73.
Edinburgh Handedness Inventory—Short Form
9
Table 6. Invariance testing fit statistics for the 4item Edinburgh Handedness Inventory using
WLSMV estimation
Model
χ²
df
p
χ²Difference
∆df
p
CFI
∆CFI
Country: USA/Australia or New Zealand
n = 1214
Unconstrained
48.82
9
< .001



.998

Metric invariance
45.33
10
< .001
8.13
3
.043
.998
.000
Scalar invariance
40.20
11
< .001
6.60
4
.159
.998
.000
Age median split
N = 1514
Unconstrained
52.69
10
< .001



.998

Metric invariance
46.26
11
< .001
6.53
3
.089
.998
.000
Scalar invariance
45.28
12
< .001
8.37
4
.079
.998
.000
Level of education
n = 1446
Unconstrained
35.23
17
.005



.999

Metric invariance
31.61
19
.035
4.91
5
.427
.999
.000
Scalar invariance
35.39
21
.026
11.68
9
.232
.999
.000
DISCUSSION
This is the first study to use confirmatory factor analysis to examine the validity the
Edinburgh Handedness Inventory when using simplified instructions and a response scale.
The proposed Edinburgh Handedness Inventory—short form, including instructions and
response scale, is reproduced in the Appendix. Because the remaining items relate to tasks
that only require one hand, only simplified instructions need to be retained. The simpler
instructions and response options, and lower number of items mean the short form is notably
less burdensome to participants.
Table 7. Assignment of categorical handedness groups in the original and short form
Edinburgh Handedness Inventories
Original version (10items)
Left n (%)
Mixed n (%)
Right n (%)
Total n
Short form (4items)
Left
87 (95)
28 (6)
0 (0)
115
Mixed
5 (5)
259 (60)
4 (0)
268
Right
0 (0)
146 (34)
867 (100)
1013
Total
92
433
871
1396
Jaimie Veale
10
Figure 1. Scatter plot of laterality quotients for the Edinburgh Handedness Inventory and its
short form
Note. n = 1396. Jitter function used to illustrate overlapping points.
In accordance with previous research, this study found that the 10item Edinburgh
Handedness Inventory did not adequately measure a single underlying handedness factor.
Unlike previous studies using confirmatory factor analysis by Dragovic (2004; Milenkovic &
Dragovic, 2012) that suggested a 7item inventory, this study provided evidence that a 7item
inventory does not adequately fit the data. Although Dragovic (2004) did not find evidence
for model misspecification for the 7item inventory, this is likely to be because this study’s
sample size (N = 203) lacked the power detect this. This can be surmised because the two
Edinburgh Handedness Inventory—Short Form
11
later studies with larger samples (Milenkovic & Dragovic, 2012 and the present study) found
evidence for this misspecification on the χ² likelihood ratio test.
The 4item inventory was proposed based on consideration of items’ face validity,
performance of items in this study’s initial analysis, and the findings previous factor analyses
(Bryden, 1977; Dragovic, 2004; McFarland & Anderson, 1980; Milenkovic & Dragovic,
2012). This inventory passed the requirements of all of the fit tests and indices when
indicators were modeled as both continuous and ordinal.
The 4item Edinburgh Handedness Inventory—short form performed well on further
validation analyses. Despite it being a short scale, it showed very high reliability on
measures of factorial composite reliability and Cronbach’s alpha (Cicchetti, 1994). The
estimate of the quality of factor scores, factor determinacy, was also high. This is noteworthy
because this is an area which can be of concern when using a low number of indicators for
factor estimation (J. S. Williams, 1978). Also, because the findings of this and another recent
study (Milenkovic & Dragovic, 2012) indicated that the inventory does not have parallel
indicators, researchers may prefer to calculate factor scores to more accurately assess
handedness.
When laterality quotients were calculated, the short form also predicted a large
proportion (94%) of the variance of the 10item inventory. This is around the same
proportion that Milenkovic and Dragovic (2012) found from the same analysis with the 7
item inventory. The level of agreement of categorization of participants into left, mixed, and
right handers between original and revised versions in this study (kappa = .73) was lower
than in Milenkovic and Dragovic’s study (kappa = .80), but this finding still suggests
“substantial agreement” (Landis & Koch, 1977). Table 7 illustrates that the main source of
disagreement in this study was participants categorized as mixed handers by the 10item
inventory were categorized as either left or right handers by the 4item inventory. The
percentage classified as mixed hander in the 10item inventory was 31% and this decreased to
19% in the 4item short form. With recent evidence that the Edinburgh Handedness
Inventory overestimates the proportion of mixedhanders (Büsch, Hagemann, & Bender,
2010; Dragovic, Milenkovic, & Hammond, 2008), it seems that by eliminating items with
notable measurement error, the short from inventory alleviates this concern.
This was the first study to test betweengroup factorial invariance for the Edinburgh
Handedness Inventory using confirmatory factor analysis. Both metric and scalar invariance
was found between those older than and younger than age 37, those with different levels of
education, and those living in the USA versus Australia and New Zealand. This can be
Jaimie Veale
12
interpreted as evidence for the same factor structure across these groups, further supporting
the validity of the short form. These findings are especially important given recent work
suggesting that a notable proportion of inmates, who are likely to have lower levels of
education on average, did not understand or follow the original Edinburgh Handedness
Inventory correctly. There was some indication that the toothbrush item loaded higher on the
factor in males than females. Because this difference was of marginal significance, only
detected on one of the two fit indices, and there is no obvious theoretical explanation for this
finding, replication of this finding would be prudent before attempting to interpret it.
Nevertheless, researchers interested in the study of gender differences in handedness should
be aware of the possibility of factorial noninvariance for this item.
This was the first confirmatory factor analysis study to validate the Edinburgh
Handedness Inventory using simplified instructions and response options. This was also the
first study to model the response options as ordinal and test factorial invariance statistically.
Results of the analyses suggested the revised 4item Edinburgh Handedness Inventory—short
form had very good reliability, factor score determinacy, and model fit. It also appeared to
have enhanced performance in handedness categorization. This short form will be of interest
to researchers wanting to measure a single handedness factor with an inventory that has brief
and simple instructions and a small number of items. This research was conducted using an
online questionnaire; future research could assess the Edinburgh Handedness Inventory—
short form using penandpaper format, and use biomarkers of handedness to further assess
its validity.
ACKNOWLEDGEMENTS
Contributors: Matt Williams critically reviewed the manuscript; Dave Clarke, Terri
Lomax, and Steve Humphries supervised the larger project; Ellen Stephenson gave helpful
comments regarding face validity of the Edinburgh Handedness Inventory.
This research was supported by a Massey University Doctoral Scholarship and a Post
Doctoral Research Fellowship with the assistance of the Government of Canada/ avec l’appui
du gouvernement du Canada.
Edinburgh Handedness Inventory—Short Form
13
REFERENCES
Asparouhov, T., & Muthén, B. O. (2010, September 29). Multiple imputation with Mplus.
Retrieved from http://statmodel.com/download/Imputations7.pdf
Blunch, N. J. (2008). Introduction to structural equation modelling using SPSS and AMOS.
London: Sage.
Brown, T. A. (2006). Confirmatory factor analysis for applied research. New York: Guilford
Press.
Bryden, M. P. (1977). Measuring handedness with questionnaires. Neuropsychologia, 15(4
5), 617–624.
Büsch, D., Hagemann, N., & Bender, N. (2010). The dimensionality of the Edinburgh
Handedness Inventory: An analysis with models of the item response theory.
Laterality, 15, 610–628. doi:10.1080/13576500903081806
Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodnessoffit indexes for testing
measurement invariance. Structural Equation Modeling, 9, 233–255.
Cicchetti, D. V. (1994). Guidelines, criteria, and rules of thumb for evaluating normed and
standardized assessment instruments in psychology. Psychological Assessment, 6,
284–290. doi:10.1037/10403590.6.4.284
Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika,
16, 297–334. doi:10.1007/BF02310555
Dragovic, M. (2004). Towards an improved measure of the Edinburgh Handedness
Inventory: A onefactor congeneric measurement model using confirmatory factor
analysis. Laterality, 9, 411–419. doi:10.1080/13576500342000248
Dragovic, M., Milenkovic, S., & Hammond, G. (2008). The distribution of hand preference is
discrete: A taxometric examination. British Journal of Psychology, 99, 445–459.
doi:10.1348/000712608X304450
Fazio, R., Coenen, C., & Denney, R. L. (2012). The original instructions for the Edinburgh
Handedness Inventory are misunderstood by a majority of participants. Laterality:
Asymmetries of Body, Brain and Cognition, 17(1), 70–77.
doi:10.1080/1357650X.2010.532801
Gregorich, S. E. (2006). Do selfreport instruments allow meaningful comparisons across
diverse population groups? Testing measurement invariance using the confirmatory
factor analysis framework. Medical Care, 44(11 Suppl 3), S78–S94.
doi:10.1097/01.mlr.0000245454.12228.8f
Jaimie Veale
14
Grice, J. W. (2001). Computing and evaluating factor scores. Psychological Methods, 6(4),
430–450. doi:10.1037/1082989X.6.4.430
Kline, R. B. (2011). Principles and practice of structural equation modeling. New York:
Guilford Press.
Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical
data. Biometrics, 33, 159–174.
McFarland, K., & Anderson, J. (1980). Factor stability of the Edinburgh Handedness
Inventory as a function of test–retest performance, age and sex. British Journal of
Psychology, 71, 135–142. doi:10.1111/j.20448295.1980.tb02739.x
Milenkovic, S., & Dragovic, M. (2012). Modification of the Edinburgh Handedness
Inventory: A replication study. Laterality: Asymmetries of Body, Brain and Cognition,
Advance online publication. doi:10.1080/1357650X.2012.683196
Muthén, L. K., & Muthén, B. O. (2008). Mplus: User’s guide 5.1. Los Angeles: Muthén &
Muthén.
Oldfield, R. C. (1971). The assessment and analysis of handedness: The Edinburgh inventory.
Neuropsychologia, 9, 97–113. doi:10.1016/00283932(71)900674
Raykov, T. (1997). Estimation of composite reliability for congeneric measures. Applied
Psychological Measurement, 21, 173–184. doi:10.1177/01466216970212006
Veale, J. F., Clarke, D. E., & Lomax, T. C. (2010). Biological and psychosocial correlates of
adult gendervariant identities: New findings. Personality and Individual Differences,
49, 252–257. doi:10.1016/j.paid.2010.03.045
Williams, J. S. (1978). A definition for the common factor model and the elimination of
problems of factor score indeterminacy. Psychometrika, 43, 293–306.
doi:10.1007/BF02293640
Williams, S. M. (1986). Factor analysis of the Edinburgh Handedness Inventory. Cortex, 22,
325–326.
Edinburgh Handedness Inventory—Short Form
15
APPENDIX
Edinburgh Handedness Inventory  Short Form
Please indicate your preferences in the use of hands in the following activities or
objects:
Always
right
Usually
right
Both
equally
Usually
left
Always
left
Writing
Throwing
Toothbrush
Spoon
Scoring:
For each item: Always right = 100; Usually right = 50; Both equally = 0; Usually left = 50;
Always
left = 100
To calculate the Laterality Quotient add the scores for the four items in the scale and divide
this by four:
Writing score
Throwing score
Toothbrush score
Spoon score
Total
Total ÷ 4 (Laterality Quotient)
Classification:
Laterality Quotient score:
Left handers
100 to 61
Mixed handers
60 to 60
Right handers
61 to 100