Investigating the effects of dimensionality and referent variability on word learning in autism and typical development.

Investigating the effects of dimensionality and referent variability on word learning in autism and typical development.

Fiona Smith

2015

The ability to learn words from pictures could give children another forum to develop
their lexical understanding and vocabulary. This is particularly important for children
with developmental disorders such as Autism. This research investigated how word
learning processes (referent selection, retention and generalisation) in autism and
typical development are influenced by learning from pictures and objects, including
single and multiple exemplars of symbols. The participants in this study were 16
typically developing children, M age=3.68, the TD group was composed of 7 males
(43.75%) and 8 females (56.25%). And 16 children diagnosed with ASD, M
age=9.37, 8 males (50%) and 8 females (50%). Participants looked at pictorial and
object referents. This was to differentiate whether there was a preference in word
acquisition and retention, depending on the structure of the stimuli. It was expected
that word referent selection, retention and generalisation would be more accurate in
the object condition compared to the picture condition, as participants would not be
relying of picture-word-associations. Participants also examined words paired with
either single or multiple exemplars of referents, to determine whether multiple
exemplars of shaped matched referents would promote shape-based generalisation
in the ASD group, which has been shown to be impaired (Hartley and Allen, 2014).
It was expected that retention would be superior when learning directly from objects
in both the ASD and TD groups, which was found in this research. We also
anticipated that labelling from multiple exemplars, rather than single exemplars,
may scaffold more consistent shape-based generalisation. We found that referent
selection was more accurate in both groups in the multiple exemplar condition
compared to the single exemplar condition. The implications of this research are
that we can further understanding of how symbols or objects benefit word learning,
retention and generalisation in ASD or TD children. And whether there are any
cognitive differences in the ASD and TD groups when it comes to word learning
processes.

Dimensionality, referent variability, word learning.

Participants
The participants in this study were 16 minimally verbal children with ASD (M age =
10.42 years, SD = 3.29) and 16 typically developing children (M age = 3.64, SD =
1.64).
Children with ASD were recruited from the specialist schools Dee Banks School in
Chester, and Hinderton School in Ellesmere Port. Typically developing children were
recruited via opportunity sampling, via the social media platform Facebook through
advertisement.
All the children with ASD received their diagnosis from a qualified clinical or
educational psychologist. This was obtained using standardised instruments (i.e.
Autism Diagnostic Observation Scale and Autism Diagnostic Interview—Revised;
(Lord, Rutter & Le Couteur, 1994, Lord, Rutter, DiLavore & Risi, 2002) and expert
judgment. Clinical diagnosis was confirmed for children with autism using the
Childhood Autism Rating Scale (CARS; Schopler, Van Bourgondien, Wellman &
Love, 2010), which was completed by a class teacher (Raw Score M score = 37.26,
Raw Score range = 27 – 53.5). The ASD were tested for non-verbal vocabulary using
the British Picture Vocabulary Scale (BPVS; Dunn, Dunn, Whetton, & Burley, 1997),
which was conducted by the experimenter. Mean receptive vocabulary of children
with autism was years 2.84 (range = 6 years – 2 years 4 months).
Some of the children diagnosed with ASD who participated in this study were current
PECS-users with impaired expressive language skills. Most of the children with ASD
who participated in this study were functionally non-verbal (no spoken words),
although, some produced speech of 1–2 words in length (however, much of this was
echolalia) and one child could speak some short phrases over three words in length.
Therefore, the sample was linguistically representative of children with ASD who
receive and may benefit from picture-based communication interventions. Participants
had 1–6 years’ experience of using PECS.
When recruiting the children diagnosed with ASD, the experimenter emailed
specialist schools, explaining the study and whether the school would be interested in
participating. When recruiting the TD children, advertisements were put on social
media platforms such as Facebook (see Appendix A). The information poster
instructed the parents to contact the experimenter via email if they were interested in
their child participating.
The study was approved by the Lancaster University Ethics Committee and informed
consent was obtained from parents before children were included in the study.
See Appendix B for completed and approved Lancaster University Ethics Committee
form
Materials
For the warm up test trials in all tests the participants were shown three familiar
objects (for example; dog, bus, chair), these were small laminated pictorial symbols.
In the picture, single and multiple exemplar conditions the participants were shown 12
laminated pictorial symbols, 6 familiar and 4 novel. The participants saw each novel
symbol once and the novel symbol twice. Participants saw the same named novel
symbols in the retention test trial, in this trial the named novel objects were shown to
each participant twice. In the generalisation test trial, the participants saw shape
matches (same object or picture, for example both would be paperclips) to the named
novel objects from the referent selection test trial and retention test trial, however they
were different colour variations (for example a red and blue paperclip). In the object
condition participants followed the same test layout and number of referents as the
other conditions, the difference being that the stimuli were actual objects compared to
pictorial symbols. The words for the familiar stimuli were gathered using the CDI
database (Fenson, Dale, Reznick, Bates, Thal, Pethick, Stiles, 1994) and appropriately
aged matched to the non-verbal age range of the ASD children and the chronological
age of the TD children (See Appendix C). The words for the novel stimuli were
picked from the NOUN database (Horst & Hout, 2016), these were picked to all be
two syllables long and words to have different phonological sounds per set. In the
picture condition were, Gloop, Virdex, Akar and Teebu. For the novel words for the
object condition were, Fiffin, Tranzer, Brisp and Pentants. For the single exemplar
condition the novel word were, Tulver, Kaki, Jefa and Blicket. For the multiple
exemplar condition the novel word were, Zepper, Toma, Modi and Chatten (see
Appendix D)
Objects were obtained through the equipment assistant at Lancaster University and
purchased through amazon. Appendix E is an example warm up selection trial which
a participant saw, and the response form completed by the experimenter. Appendix F
is an example of a referents selection trial, which participant saw, and the response
form completed by the experimenter. Appendix G is an example of a retention test
trial a participant will have seen, and the response form completed by the
experimenter. Appendix H is an example of the generalisation test trial which a
participant saw, and the response form completed by the experimenter. All test trials
were pseudorandomised per participant per condition and trial. Therefore, while all
the participants will have seen the same number of familiar and novel objects or
pictures. And each picture or object will have had the same name per shape matched
object they will have been in a different order. Therefore, a different response form
was required per participant, for the change in referent location and set order.
Procedure
Prior to the children participating the parents received, the information sheet (see
Appendix I), and the consent form (see Appendix J). On the last day of experiments
the experimenter brought the debrief forms (see Appendix K).
Participants were test individually, in their schools for the children with ASD or in
their own homes for the TD children, and were always accompanied by a familiar
adult, teaching assistant or parent. The participants were seated at a table opposite the
experimenter; the materials were placed within reaching distance of the participants.
Children were reinforced throughout the session; correct performance was only
reinforced during the warm up trial. The first test examined the picture condition vs
the object condition, the second test examined single vs multiple exemplars. The tasks
were between participants, as they were examining the results of the TD group
compared to the ASD group, however for the analysis some within participants
analysis was carried out to determine accuracy between test conditions (e.g. picture vs
object). Each task always consisted of a warm up stage, referent selection trial,
distracter familiarisation trial, retention test trial and generalisation test trial. The test
trials were based on that done by Horst and Samuelson in 2008, with the extension of
the generalisation trial which was not included in the Horst and Samuelson (2008)
study.
Picture Condition vs Object Condition Tests
Warm Up Stage
Participants were shown three sets of three familiar objects, in the object condition, in
the picture condition participants were shown three familiar pictures. Participants
were asked to identify each in turn, the warm up objects or pictures were
pseudorandomised per participant, changing the order and location per participant per
condition. The pictures or objects were removed and reordered after each set, and the
participants response recorded.
Referent Selection Trial
Participants were shown four sets of stimuli (pictures for the picture condition and
objects for the object condition) the sets of stimuli were different per condition, each
consisting of two familiar items and one novel item, each set was shown four times,
the novel referent was shown twice and the two familiar referents once. The order and
location of the sets was pseudorandomised for each participant, the location of the
novel object was never in the same location twice consecutively, and a novel or
familiar object or picture was never requested more than twice consecutively. Sets
were not presented twice in a row.
Distractor Familiarisation
To control for novelty or familiarity preferences in the subsequent test trials, children
were shown all the novel objects that used in generalisation test trials. The new novel
objects were a different colour variation of a previously seen novel object, which was
named in the referent selection trial. Novel objects or pictures were shown against a
previously named novel objects or pictures, which was not a shape or colour match to
the new novel object. Objects or pictures were shown so one previous named novel
object was shown against a new novel object or picture. The objects were not shape or
colour matched, the objects or pictures were placed in front of the participant, they
were not asked to identify them just to “look”.
Retention Test Trial
Retention trials will assess children’s memory of the newly-learned word-referent
pairings. Participants were shown four sets; each set was shown twice with the target
object requested twice. The sets were made up of three named novel objects, names
were picked from the NOUN database (Horst & Houst, 2016), each made up of two
syllables, objects or pictures were picked on the basis that participants items that
would be novel to them, for instance gym or plumbing equipment. Objects and
pictures which were not shape or colour matches to each other and were shown in the
referent selection test trial. The order and location of each object or picture per set
was pseudorandomised per participant per trial. The location of the novel object was
never in the same location twice consecutively, and a novel or familiar object or
picture was never requested more than twice consecutively. Sets were not presented
twice in a row.
Generalisation Test Trial
Generalisation trials will assess children’s extension of labels to new items.
Participants were shown four sets; each consisting of three objects or pictures, each
set was shown twice with the target object being requested twice. The objects or
pictures in the sets were shape matches to the objects or pictures shown in the referent
selection, and retention trials, but different colour variations. All the shape matched
objects or pictures were also colour matched to a non-shape matched object from the
previous conditions. The order and location of each object or picture per set was
pseudorandomised per participant per trial. The location of the novel object was never
in the same location twice consecutively, and a novel or familiar object or picture was
never requested more than twice consecutively. Sets were not presented twice in a
row.
Single vs Multiple Exemplars Tests
Warm Up Trial
Participants were shown three sets of three familiar pictures in both the single and
multiple exemplar conditions. Participants were asked to identify each in turn, the
pictures were pseudorandomised per participant, changing the order and location per
participant per condition. The pictures were removed and reordered after each set, and
the participants response recorded.
Referent selection Trial
Participants were shown four sets of stimuli, the sets of stimuli were different per
condition, each consisting of two familiar items and one novel item, each set was
shown four times, the novel referent was shown twice and the two familiar referents
once. In the multiple exemplar trial, two differently-coloured versions of each
unfamiliar object were named (one per novel trial for each set). The order of the sets
was pseudorandomised for each participant. The order and location of each object or
picture per set was pseudorandomised per participant per trial. The location of the
novel object was never in the same location twice consecutively, and a novel or
familiar object or picture was never requested more than twice consecutively. Sets
were not presented twice in a row. The order and location of the sets was
pseudorandomised for each participant, the location of the novel object was never in
the same location twice consecutively, and a novel or familiar object or picture was
never requested more than twice consecutively. Sets were not presented twice in a
row.
Distractor Familiarisation
To control for novelty or familiarity preferences in the subsequent test trials, children
were shown all the novel pictures that used in generalisation test trials. The new novel
pictures were a different colour variation of a previously seen novel picture referent,
which was named in the referent selection trial. Novel pictures were shown against a
previously named novel pictures, which was not a shape or colour match to the new
novel picture. Pictures were shown so one previous named novel referent was shown
against a new novel picture. The referents were not shape or colour matched, the
pictures were placed in front of the participant, they were not asked to identify them
just to “look”.
Retention Test Trial
Retention trials will assess children’s memory of the newly-learned word-referent
pairings. Participants were shown four sets; each set was shown twice with the target
referent requested twice. The sets were made up of three named novel objects, names
were picked from the NOUN database (Horst & Houst, 2016), each made up of two
syllables, pictures were picked on the basis that participants items that would be novel
to them, for instance gym or plumbing equipment. Pictures which were not shape or
colour matches to each other and were shown in the referent selection test trial. The
order and location of each picture per set was pseudorandomised per participant per
trial. The location of the novel object was never in the same location twice
consecutively, and a novel or familiar object or picture was never requested more than
twice consecutively. Sets were not presented twice in a row.
Generalisation Test Trial
Generalisation trials will assess children’s extension of labels to new items.
Participants were shown four sets; each consisting of three pictures, each set was
shown twice with the target object being requested twice. The pictures in the set were
shape matches to the picture shown in the referent selection, and retention trials, but
different colour variations. All the shape matched pictures were also colour matched
to a non-shape matched object from the previous conditions. The order and location
of each picture per set was pseudorandomised per participant per trial. The location of
the novel object was never in the same location twice consecutively, and a novel or
familiar picture was never requested more than twice consecutively. Sets were not
presented twice in a row. In the multiple exemplar condition the generalisation test
trial introduced the shape matched referent in a third colour that was coloured
matched to a referent of a different shape matched seen in the referent selection or
retention test trial.

Lancaster University

Data/SPSS.sav

Smith2015

Rebecca James

Open

None

English

Data

LA1 4YF

Calum Hartley

MSC

Cognitive, Developmental Psychology

16 minimally verbal children with ASD and 16 typically developing children

ANOVA, Correlation, quantitative, t-test