ArticlePDF Available

Edinburgh Handedness Inventory - Short Form: A revised version based on confirmatory factor analysis

Authors:

Abstract and Figures

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 10-item inventory had poor fit for the data. This study also detected model misspecification in the previously-proposed 7-item modification. A 4-item Edinburgh Handedness Inventory - Short Form had good model fit with items modelled as both continuous and ordinal. Despite its brevity, it showed very good reliability, factor score determinacy, and correlation with scores on the 10-item inventory. By eliminating items that were modelled with considerable measurement error, the short form alleviates the concern of the 10-item inventory over-categorising 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 4-item 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.
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),
164-177, available online: http://www.tandfonline.com/10.1080/1357650X.2013.783045
EDINBURGH HANDEDNESS INVENTORYSHORT 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 10-item inventory had poor fit for the data. This study also detected model
misspecification in the previously-proposed 7-item modification. A 4-item Edinburgh
Handedness Inventoryshort 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 10-item inventory. By eliminating items that
were modeled with considerable measurement error, the short form alleviates the concern of
the 10-item inventory over-categorizing 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 4-item inventory.
The proposed Edinburgh Handedness Inventoryshort 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 right-handers 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 test-retest 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 box-lid were modeled with a high proportion of
Edinburgh Handedness InventoryShort 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 7-item 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 5-point Likert response options, an estimation technique for
ordered-categorical 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 internet-based 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 gender-variant 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
5-point 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
Yuan-Bentler robust maximum likelihood estimation and mean- and variance-adjusted
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: Satorra-Bentler 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 InventoryShort 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 over-sensitive 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
df
p
CFI
TLI
RMSEA
SRMR
ρ
α
FSD
10 items
35
< .001
.84
.79
.145
.05
.95
.95
.98
7 items
14
< .001
.99
.98
.047
.02
.95
.95
.98
4 items
2
.819
1.00
1.00
.000
.00
.93
.93
.97
4 items, parallel
indicators
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 right-handed use); and
- the match and knife items are the only remaining items pertaining to tasks which require
two handsremoving these will simplify the scale.
Edinburgh Handedness InventoryShort Form
7
Given these considerations, the knife, striking match, and scissors items were
removed. Tables 2 and 3 show the resultant 4-item 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 4-item model are given in Table 4.
Table 4. Model estimates for the 4-item 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 between-countries 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 between-gender 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 4-item 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 10-item inventory and the
4-item 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 10-item and 4-item inventories was kappa = .73.
Edinburgh Handedness InventoryShort Form
9
Table 6. Invariance testing fit statistics for the 4-item 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 Inventoryshort 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 (10-items)
Left n (%)
Mixed n (%)
Right n (%)
Total n
Short form (4-items)
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 10-item 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 7-item inventory, this study provided evidence that a 7-item
inventory does not adequately fit the data. Although Dragovic (2004) did not find evidence
for model misspecification for the 7-item 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 InventoryShort 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 4-item 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 4-item Edinburgh Handedness Inventoryshort 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 10-item 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 10-item
inventory were categorized as either left or right handers by the 4-item inventory. The
percentage classified as mixed hander in the 10-item inventory was 31% and this decreased to
19% in the 4-item short form. With recent evidence that the Edinburgh Handedness
Inventory overestimates the proportion of mixed-handers (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 between-group 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 4-item Edinburgh Handedness Inventoryshort
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 pen-and-paper 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 InventoryShort 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), 617624.
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, 610628. doi:10.1080/13576500903081806
Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodness-of-fit indexes for testing
measurement invariance. Structural Equation Modeling, 9, 233255.
Cicchetti, D. V. (1994). Guidelines, criteria, and rules of thumb for evaluating normed and
standardized assessment instruments in psychology. Psychological Assessment, 6,
284290. doi:10.1037/1040-3590.6.4.284
Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika,
16, 297334. doi:10.1007/BF02310555
Dragovic, M. (2004). Towards an improved measure of the Edinburgh Handedness
Inventory: A one-factor congeneric measurement model using confirmatory factor
analysis. Laterality, 9, 411419. 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, 445459.
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), 7077.
doi:10.1080/1357650X.2010.532801
Gregorich, S. E. (2006). Do self-report instruments allow meaningful comparisons across
diverse population groups? Testing measurement invariance using the confirmatory
factor analysis framework. Medical Care, 44(11 Suppl 3), S78S94.
doi:10.1097/01.mlr.0000245454.12228.8f
Jaimie Veale
14
Grice, J. W. (2001). Computing and evaluating factor scores. Psychological Methods, 6(4),
430450. doi:10.1037/1082-989X.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, 159174.
McFarland, K., & Anderson, J. (1980). Factor stability of the Edinburgh Handedness
Inventory as a function of testretest performance, age and sex. British Journal of
Psychology, 71, 135142. doi:10.1111/j.2044-8295.1980.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, 97113. doi:10.1016/0028-3932(71)90067-4
Raykov, T. (1997). Estimation of composite reliability for congeneric measures. Applied
Psychological Measurement, 21, 173184. doi:10.1177/01466216970212006
Veale, J. F., Clarke, D. E., & Lomax, T. C. (2010). Biological and psychosocial correlates of
adult gender-variant identities: New findings. Personality and Individual Differences,
49, 252257. 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, 293306.
doi:10.1007/BF02293640
Williams, S. M. (1986). Factor analysis of the Edinburgh Handedness Inventory. Cortex, 22,
325326.
Edinburgh Handedness InventoryShort 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
... A general screening questionnaire was used to record participants' demographics (age, gender, education). For the abovementioned reasons, the BDI was used to screen for symptoms of depression 47 , the MMSE o screen for cognitive impairment 36 , and the Veale Edinburgh Handedness Inventory to ascertain right-handed dominance 46 . Questionnaires were administered by one of the authors (JE) with the help of the experimenters reported in the Acknowledgment section. ...
... On the Affective Go/No-Go test, for AGN 'positive' outcome measures involving mean correct latencies, both groups did slightly better during the 'post' session but at the statistical level these variations in the performance were not substantial: indeed there were no significant main effect for Session, F (1,46) = 1.20, p = 0.28, η 2 p = 0.025 nor was there any main effect for Group, F (1,46) = 2.14, p = 0.15, η 2 p = 0.044 or for the interaction effect between Session * Group, F (1,46) = 0.04, p = 0.84, η 2 p = 0.0008. Conversely, for omission scores, while there were no significant main effects of the two factors (Session, F (1,46) = 1.11, p = 0.30, η 2 p = 0.024 or Group, F (1,46) = 0.04, p = 0.83, η 2 p = 0.0009) the two groups behaved in the opposite way, with the EG doing better during the 'post' session while the CG did worse (significant Session * Group interaction effect, F (1,46) = 5.10, p = 0.029, η 2 p = 0.10). ...
... On the Affective Go/No-Go test, for AGN 'positive' outcome measures involving mean correct latencies, both groups did slightly better during the 'post' session but at the statistical level these variations in the performance were not substantial: indeed there were no significant main effect for Session, F (1,46) = 1.20, p = 0.28, η 2 p = 0.025 nor was there any main effect for Group, F (1,46) = 2.14, p = 0.15, η 2 p = 0.044 or for the interaction effect between Session * Group, F (1,46) = 0.04, p = 0.84, η 2 p = 0.0008. Conversely, for omission scores, while there were no significant main effects of the two factors (Session, F (1,46) = 1.11, p = 0.30, η 2 p = 0.024 or Group, F (1,46) = 0.04, p = 0.83, η 2 p = 0.0009) the two groups behaved in the opposite way, with the EG doing better during the 'post' session while the CG did worse (significant Session * Group interaction effect, F (1,46) = 5.10, p = 0.029, η 2 p = 0.10). However, subsequent, post-hoc analysis, using Bonferroni post-hoc test, revealed no specific significant findings in pairwise testing. ...
Article
Full-text available
Regular physical activity is associated with healthier brains and improved cognition among older adults. Yet, the impact of a short bout of exercise on older adults’ cognition still is not fully clarified. The present study explored the effects of 20 min of moderate-intensity aerobic exercise (on a cycle ergometer) on cognition. Forty-eight healthy older adults were randomly assigned to an exercise or a control group and completed four cognitive tests: Affective Go/No-Go (AGN), Simple Reaction Time (SRT), Spatial Working Memory (SWM) and a Backward Counting task. Tests were administered prior to and immediately after 20 min of cycling (exercise group) or rest (control group). Mixed-design 2 × 2 ANOVAs indicated a significant interaction of Group x Session, for commission errors on the positive valence of the Affective Go/No-go task, indicating that the exercise group performed better on one aspect of this inhibition test after cycling (p = 0.004), while the control group’s performance declined after rest. A similar pattern was found for the SWM, with the exercise group showing a significantly better performance after the exercise both for total error (p = 0.027) and the strategy (p = 0.002), while no improvement was observed after rest (controls). The study suggests that inhibitory control functions and working memory may be improved by a single relatively short bout of moderate exercise. However, the null effects of exercise on the other cognitive measures indicate that the neurocognitive benefits of acute exercise for older adults may be selectively sensitive to exercise parameters and to specific aspects of cognition.
... Indeed, the extant literature provides good albeit very limited evidence for frequency detuning among individual motor representations within both M1 cortical and subcortical motor structures. In their EEG study Neuper and Pfurtscheller (13) reported have reported distinct peak resonance frequencies within the beta-band motor rhythm (13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30) for upper versus lower limb sensorimotor representations, with significantly higher peak beta frequencies (> 20 Hz) at recording locations over the lower limb representation compared to locations over the upper representation (< 20 Hz). A comparable result has been obtained for beta band upper versus lower limb representations in the subthalamic nucleus (14). ...
... and 19 typically developing (TD) children (8 females, mean age =11.0 years; SD = 2.5, range 7.5 -16.7) and previously described in (16). All participants were righthanded as assessed by a short version of the Edinburgh handedness questionnaire (17). All procedures were approved by the Macquarie University Human Subjects Research Ethics Committee. ...
Preprint
Full-text available
Many decades after Penfield’s (1937) classic depiction of the motor homunculus, it remains unclear how spatially contiguous and interconnected representations within human sensorimotor cortex might separate their activities to achieve the directed and precise control of distinct body regions evident in activities as different as typing and speaking. One long-standing but relatively neglected explanation draws from models of simple physical systems (like swinging pendulums) to posit that small differences in the oscillatory properties of neuronal populations (termed “frequency detuning”) can result in highly effective segregation of their activities and outputs. We tested this hypothesis by comparing the peak frequencies of beta-band (13-30 Hz) motor rhythms measured in a magnetoencephalographic neuroimaging study of finger and speech movements in a group of healthy adults and a group of typically developing children. Our results confirm a peak frequency task difference of about 1.5 Hz in the beta motor rhythms of both left and right hemispheres in adults. A comparable task difference was obtained in children for the left but not for the right hemisphere. These results provide novel support for the role of frequency detuning in the functional organisation of the brain and suggest that this mechanism should play a more prominent role in current models of bodily representations and their development within the sensorimotor cortex. NEW & NOTEWORTHY This work is the first evidence for frequency detuning of an intrinsic motor rhythm in spatially contiguous regions of primary motor cortex associated with hand and speech movements, supporting a mechanism of functional segregation that has been largely overlooked by current models of motor cortex organization. Our results from children further provide the first evidence for frequency detuning in the developing brain, indicating an early-developing mechanism with later refinements of hemispheric control.
... For the control group, participants remained untrained throughout the study period. Arm dominance was determined by asking the participant which arm is their dominant arm, and this was confirmed by the Edinburgh Handedness Inventory [23]. ...
Article
Full-text available
Training one limb with a high‐load has been shown to augment strength changes in the opposite limb training with a low‐load (via cross‐education of strength), indicating that within‐subject models can be problematic when investigating strength changes. This study examined if the cross‐education of strength from unilateral high‐load training could augment the strength changes in the opposite arm undergoing the same unilateral high‐load training. 160 participants were randomized to one of four groups: (1) training on the dominant arm followed by the non‐dominant arm (D + ND), (2) training on the dominant arm only (D‐Only), (3) training on the non‐dominant arm only (ND‐Only), and (4) a non‐exercise control. All exercise groups performed 18 sessions of unilateral high‐load elbow flexion exercise over 6 weeks. Participants were compared for changes in 1RM strength and muscle thickness. Changes in strength of the non‐dominant arm were greater in D + ND (2.7 kg) and ND‐Only (2.6 kg) compared to D‐Only (1.5 kg) and control (−0.2 kg), while the changes were greater in D‐Only compared to control. The same finding was observed in the dominant arm. Only the arms being directly trained observed increases in muscle thickness. Unilateral high‐load resistance training increased strength in the opposite untrained arm, without changes in muscle thickness. This cross‐education of strength did not augment the strength changes in the contralateral arm undergoing the same unilateral high‐load training. However, it does not necessarily indicate that within‐subject models are methodologically sound to investigate strength change if both limbs are trained with a high‐load.
... Handedness was assessed using the 10-item Edinburgh Handedness Inventory (Oldfield, 1971), with scores ≥ + 60 being assigned as right-handed, scores between + 59 to -59 being assigned as mixed-handed, and scores ≤ -60 being assigned as left-handed. These cut-off points were based on research by Milenkovic and Dragovic (2012) and Veale (2013). The mixed-handed individuals were recruited as they all stated a strong right-hand preference for writing, with writing being a key predictor of handedness (Bryden, 1977;Rigal, 1992). ...
Article
Full-text available
Right-handed individuals (RHIs) demonstrate perceptual biases towards their right hand, estimating it to be larger and longer than their left. In addition, RHIs estimate that they can grasp larger objects with their right hand than their left. This study investigated whether visual information specifying handedness enhances biases in RHIs’ perceptions of their action capabilities. Twenty-two participants were placed in an immersive virtual environment in which self-animated, virtual hands were either presented congruently to their physical hand or mirrored. Following a calibration task, participants estimated their maximum grasp size by adjusting the size of a virtual block until it reached the largest size they thought they could grasp. The results showed that, consistent with research outside of virtual reality, RHIs gave larger estimates of maximum grasp when using their right physical hand than their left. However, this difference remained regardless of how the hand was virtually presented. This finding suggests that proprioceptive feedback may be more important than visual feedback when estimating maximum grasp. In addition, visual feedback on handedness does not appear to enhance biases in perceptions of maximum grasp with the right hand. Considerations for further research into the embodiment of mirrored virtual limbs are discussed.
... Based on citation analysis of several studies that use Confirmatory Factor Analysis (CFA), it can be concluded that CFA has a very important role in scientific literature. Two examples of research mentioned, namely " (Marsh et al., 2014)" regarding exploratory structural equation modeling and research by Veale (2014) regarding the revision of the Edinburgh Handedness Inventory, show that CFA is used to measure, confirm, or revise conceptual models in various research context. The high number of citations from these two studies indicates that the contribution of CFA in validating and understanding the factor structure of a concept or measurement instrument is highly appreciated and recognized by the scientific community. ...
Article
Full-text available
This research aims at bibliometric analysis on the topic of confirmatory factor analysis in the field of education. Researchers used the RStudio and Vosviewer applications to map bibliometrics results. From the results of the analysis that has been carried out, it is concluded that overview of the bibliometric analysis of 203 research documents, published between 1987 and 2023, from various sources in the Scopus database. Publication growth shows an increase of 7.61% every year. In contrast, it saw a peak in growth in 2018, with 24 publications, reflecting significant research interest and focus. The University of Tasmania from Australia dominates with 13 articles. Nurse Education Today was the highest contributing journal, contributing 7 articles, highlighting the important role of nursing in this research. This analysis illustrates the diversity of contributions of researchers from various institutions in Thailand in the exploration of CFA in education. The highest document citation is "(Marsh et al., 2014)" regarding exploratory structural equation models. There are five clusters that discuss CFA in the education sector. There are keywords multivariate analysis, exploratory factor analysis, surveys and questionnaires, reliability, and construct validity as the latest keywords that can be used as recommendations for further research in the field of CFA in education.
... All participants in the NT group had no clinical, psychological or developmental disorders. The Edinburgh Handedness Inventory (Veale, 2014) was used to determine each participant's handedness and confirmed that all participants were right-handed; they also had normal vision (20/20) because we used eye-tracking glasses, ensuring that all children had normal vision without wearing any glasses for visual acuity correction was essential. ...
Article
There has been a lot of controversy regarding mirror neuron function in autism spectrum disorder (ASD), in particular during the observation of biological motions (BM). Here, we directly explored the link between visual attention and brain activity in terms of mu suppression, by simultaneously recording eye‐tracking and EEGs during BM tasks. Nineteen autistic children (15 boys, mean age = 11.57 ± 4.28 years) and 19 age‐matched neurotypical (NT) children (15 boys, mean age = 11.68 ± 5.22 years) participated in the study. Each participant's eye movement and EEG were simultaneously recorded while watching four BM stimuli (walking, cartwheeling, free‐throwing and underarm throwing) and a scrambled condition. Mu (8–13 Hz) suppression index (SI) for central regions was calculated. Fixation counts and percent of fixation time were calculated as indices of eye movements. EEG results revealed significant mu suppressions in the central region in both groups for all BM actions. Eye‐tracking results showed that NT children had greater fixation counts and a higher percentage of fixation time than autistic children, indicating greater overall visual attention to BM. Notably, correlational analyses for both groups further revealed that individuals' fixation time and fixation counts were negatively correlated with the mu suppression index for all actions, indicating a strong association between visual attention and mu SI in the central region. Our findings suggest a critical role of visual attention in interpreting mu suppression during action perception in autism.
... All participants reported German as their first language, and reported 251 that they had no proficiency in speaking or reading a language that uses a 252 non-alphabetic script. Participants' handedness was assessed via the revised short form 253 of the Edinburgh Handedness Inventory(Veale, 2014), with participants only permitted to 254 take part if they scored a laterality quotient of at least +40 indicating right handedness.All participants completed education to at least level 4 in the European Qualifications Framework (i.e., in Germany, Abitur). No participants reported diagnosis of any disorder details). ...
Preprint
Full-text available
Letter processing plays a key role in visual word recognition. However, word recognition models typically overlook or greatly simplify early perceptual processes of letter recognition. We suggest that optimal transport theory may provide a computational framework for describing letter shape processing. We use representational similarity analysis to show that optimal transport cost (Wasserstein distance) between pairs of letters aligns with neural activity elicited by visually presented letters <225 ms after stimulus onset, outperforming an existing approach based on shape overlap. We additionally show that optimal transport can capture the emergence of geometric invariances (e.g., to position or size) observed in letter perception. Finally, we demonstrate that Wasserstein distance predicts neural activity similarly well to features from artificial networks trained to classify images and letters. However, whereas representations in artificial neural networks emerge in a computationally unconstrained manner, our proposal provides a computationally explicit route to modeling the earliest orthographic processes.
Article
Full-text available
Researchers commonly compare means and other statistics across groups with little concern for whether the measure possesses strong factorial invariance (i.e., equal factor loadings and intercepts/thresholds). When this assumption is violated, inaccurate inferences associated with statistical and practical significance can occur. This manuscript emphasizes the importance of testing for measurement invariance (MI) and provides guidance when conducting these tests. Topics discussed are potential causes of noninvariant items, the difference between measurement bias and invariance, remedies for noninvariant measures, and considerations associated with model estimation. Using a sample of 491 teachers, a demonstration is also provided that evaluates whether a newly constructed behavior and instructional management scale is invariant across elementary and middle school teachers. Analyses revealed that the results differ slightly based on the estimation method utilized although these differences did not greatly influence the latent factor mean difference conclusions. Additional implications and considerations related to invariance testing are discussed.
Article
Let Ω be a domain in ℝ n whose boundary is C 1 if n≥3 or C 1,β if n=2. We consider a magnetic Schrödinger operator L W,q in Ω and show how to recover the boundary values of the tangential component of the vector potential W from the Dirichlet to Neumann map for L W,q . We also consider a steady state heat equation with convection term Δ+2W·∇ and recover the boundary values of the convection term W from the Dirichlet to Neumann map. Our method is constructive and gives a stability result at the boundary.
Article
A structural equation model is described that permits estimation of the reliability index and coefficient of a composite test for congeneric measures. The method is also helpful in exploring the factorial structure of an item set, and its use in scale reliability estimation and development is illustrated. The modeling. estimator of composite reliability it yields does not possess the general underestimation property of Cronbach's coefficient a.