The one child in the control group who made two
or more errors relevant to her grammatical target on the
articulation screen did not improve significantly on her
grammatical target (number 4).
Assessment of blinding
As noted above, two of the three testers were blind to
the purpose of the study and were responsible for the
majority of the testing.The third testerwas aware of both
the purpose and group allocation of the subjects. Poststudy
interviews confirmed that the blind testers had
remained blind to the purpose of the testing.Of the eight
children the non-blinded tester tested, all were from the
experimental group. A comparison of the gain scores of
the children tested by the blind testers versus the children
tested by the non-blind tester found no difference. The
mean gain from time 2 to time 3 for children tested by
the blind testers was 7.18 (SD = 5.58) and for nonblind
tester 6.75 (SD = 7.57); t(17) = 0.14 (p > 0.05).
For time 1 to time 3, the mean gain for children tested
by the blind testers was 7.27 (SD = 6.78) and for nonblind
tester 6.75 (SD = 8.07); t(17) = 0.15 (p > 0.05).
It appears that the scores of the non-blind tester were not
affected by her knowledge of the treatment conditions.
Discussion
The purpose of this study was to measure the effectiveness
of an expressive grammar treatment programme
in use in a real-life clinical setting. We tested whether
changes in grammatical outcomes were specific to grammatical
intervention targets or extended to grammatical
morphemes that were not targeted. We found a significant
difference in performance on treated grammatical
targets, pre- and post-treatment for a group of children
who received grammar treatment. In contrast, no
difference in performance was found in the equivalent
period prior to treatment or in a group of children who
received a control treatment. Furthermore, the treatment
effect in the experimental group was specific to
the grammatical targets treated and did not carry over
to untreated grammatical targets. Treatment success was
less pronounced in children with articulation difficulties
which interfered with their ability to produce the
grammatical targets. The significant treatment group
differences corresponded to a clinically significant, large
treatment effect. Importantly, analysis of each child’s
performance in a single-subject design indicated that
the treatment effect was significant for the majority of
children (rather than the results being due to small effects
that were significant when accumulated across the
group but were not for any individual child).
The success of this treatment programme is a clear
indication that expressive grammar treatment in schoolaged
children can be effective in a real-life clinical setting.
This treatment success cannot be attributed either
to the school environment or solely to small group instruction.
All of the children in the study attended a
special school for children with SLI. Classes consist of
12 students, with a teacher and teacher assistant in each
class. Thus, the children were taught by teachers experienced
in working with children with language impairment,
with a ratio of 1 teacher or aide per six students.
Nevertheless, the changes in grammatical performance
cannot be attributed to this enriched educational setting
alone, as we found no change in performance on
the grammar tests in the 8 weeks prior to treatment.
Of course, the teachers were provided with a list of the
grammatical targets to reinforce throughout the day, as
278 Karen M. Smith-Lock et al.
opportunities arose. Therefore, while the environment
of the school did not affect change on the grammatical
tests alone, conscious inclusion of the grammatical targets
in the classroom environment throughout the day
may have contributed to the treatment success. The
change in grammar scores also cannot be attributed to
small group language teaching alone. Both the experimental
and control groups received intensive language
instruction in groups of four or five children.However,
only the group that received grammar instruction improved
their grammar scores.
In fact, the lack of improvement in the experimental
group on untreated grammatical targets emphasizes
the specificity of the treatment effect. Specially designed
treatment focused on a single grammatical target improved
that performance on that target, and that target
only. This suggests that general language enrichment or
indeed intervention targeted to more global language
areas (e.g. grammar, semantics) may not be sufficient
to affect change on specific targets. While the evidence
demonstrates that this specific, focused treatment can
be delivered in group settings, it appears necessary to
identify very specific goals for each individual child
and target these goals in order to maximize each child’s
progress.
While the treatment did result in improvement in
specific grammatical areas, it is important to note that
the content and procedures in the test and treatment
were very different. It is not that case that the children
were simply taught the test. The treatment activities
contained none of the vocabulary items used in
the Grammar Elicitation Tests. This is true, by extension,
of the list of items that the teachers were given for
use in the classroom (which were the items used in the
treatment sessions). Furthermore, the treatment activities
were very different in nature to the Grammar Elicitation
Test. In the one-on-one test situation, children
were asked questions designed to elicit the grammatical
target in response to a picture, whereas the treatment
sessions consisted of group play activities. Therefore, we
can be confident that the children learned the targeted
grammatical construction, not simply vocabulary items
or test procedure.
The treatment delivered consistent results across
most children. All but one of the children who passed the
Articulation Screening Test improved on the Grammar
Elicitation Test, with ten of 14 of these demonstrating a
significant treatment effect on a stringent single-subject
analysis. This is a striking outcome, given the known
heterogeneity of population with SLI.
Of the children who did not show improvement on
the Grammar Elicitation test, five (out of six) failed the
Articulation Screening Test. That is, they were unable
to produce consistently the final consonant clusters required
to use their grammatical target. In fact, only one
of the children who failed the Articulation Screening
Test demonstrated a treatment effect and this did not
reach significance on a single-subject analysis. Again,
this shows remarkable consistency in a heterogeneous
population. It is important to note that the majority
of children in the study had some articulation difficulties.
However, a child failed the Articulation Screening
Test only if their articulation errors interfered with production
of their identified grammatical target. Thus,
articulation difficulties per se did not impede progress,
only articulation difficulties which interfered with the
production of grammatical targets. Thus, in spite of
the intensive modelling, options for the production and
corrective feedback, treatment targeted at grammar goals
did not improve articulation. It appears that articulation
needs to be addressed directly in these children.
After articulation was taken into account, there remained
only one child who showed no treatment gains.
This child was one of the highest functioning children.
His errors consisted primarily of the [əd] allomorph in
his grammatical target of past tense. It is possible that
because treatment included items spread across all three
allomorphs ([t], [d], [əd]), this child did not receive
enough exemplars to improve on [əd].
There are several factors that are likely to have contributed
to the overall treatment success. The key factors
that distinguished the treatment programme from
the general classroom programme were: (1) the identification
of specific targets for each individual, (2) the
repeated modelling of grammatical targets, followed by
(3) multiple opportunities for each child to produce
the targets, followed by (4) feedback to the child and
(5) opportunities for the child to correct him/herself.
When children participated in activities which allowed
application of these treatment techniques to a specific
grammatical goal, test scores on their selected grammatical
targets improved. The on-going reinforcement of
the goals throughout the day, in the manner outlined
above, is also likely to have contributed to the treatment
success.
While the study did not formally assess the acceptability
of the interventions for stakeholders, the school
involved in the research was pleased with the intervention
on a number of levels. From an administrative
point of view, having objectively measured and documented
treatment success provided administrators with
evidence that their school provides an effective service
that deserves continued financial support. Froma teaching
point of view, staff found the preset activities and
materials convenient and practical. In addition, staff
found the professional development enlightening and
useful. In fact, it was this ongoing professional support,
including communication of goals and hands-on teaching
of techniques that school staff identified as the most
valuable outcome of the project. We were told we had
Expressive grammar treatment 279
provided teaching staff with knowledge ‘that would last
a lifetime’. Since the study, the school has continued to
implement the programme and has shared it with other
schools.