Conducting Time Series Research

�Briefly describe a time series design that you would use to investigate the impact of
your intervention, and explain how you would collect and record the additional data. Be
�Select a method for graphically representing the provided data. Use the method to
create a graph.
�Analyze the data and provide a narrative description of your analysis.

Conducting Time Series Research

Before an intervention is administered, it is important to track the symptom status of a
psychotherapy patient. This helps in determining whether the intervention works and how
effective such a treatment is. This is the essence of time series research and as Nash et al
assert, it qualifies as a true experiment that can stand alongside other group designs such as
randomized control trial (Borckardt et al, 2008). In this study, three observational periods,
off-task behaviors, physical aggression, and verbal aggression are considered. On Monday
during a math class, the observation is made between 11:20 and 11:30 am using a 10-second
partial interval to assess the frequency of the behaviors. The same procedure is repeated on
Wednesday during recess and Friday during a group activity and the results are as shown in
the tables.
Since there is need for reliable information and the time of three days is limited, a
single-subject time series design is applied. The research relies on a third party to collect the
data as Kaya continues with her activities. To help in the analysis of the effectiveness of the
intervention, graphs are created and analyzed. Finally, a decision is reached on whether the
analysis is effective enough. In summary, this paper poses the research question: “What
impact does a designed intervention have in reducing off-task behavior, physical aggression,
and verbal aggression for the identified child?”

Conducting Time Series Research 2

The methodology shown above is a single-subject type that is rigorous and applicable
for establishment of evidence-based practices (Horner, 2005). This research is mostly
experimental and involves both dependent and independent variables. Since the dependent
variable in single-subject research is the observable behavior, Kaya’s case has several
variables. The dependent variables are off-task behavior, physical aggression, and verbal
aggression. These variables qualify in this case because they present the qualities as outlined
in the Council for Exceptional Children. According to this text, dependent variables have
social significance, are accessed for consistency in the experiment, and are operationally
defined (Horner, 2005).
Since the targets of the intervention have been identified as off-task behaviors,
physical aggression, and verbal aggression, the research now seeks to operationalize of the
problem. To measure behavior, this research uses frequency as the sole characterization.
Monitoring Kaya
Because the study aims at checking whether the client is making any meaningful
progress, monitoring the behavior is necessary. Here, the research uses a specific design
referred to as a B-design. This design has only an intervention phase. It is during this
intervention that the class teacher records repeated measurements of behavior in a log.
Data Collection
As insinuated previously, Kaya’s class teacher will help in data collection. This is
necessary considering that there are consequences if a stranger takes up this role instead. The
teacher will be given a log in which to fill data as time counts. Acting as a social work
researcher, the teacher will observe Kaya’s behavior and fill data in the log.

Conducting Time Series Research 3
One problem with collecting data is reactivity, a situation where the subject
systematically reduces behavior. Luckily, this is not present in this study because Kaya is not
aware the teacher is observing her behavior.
Another problem associated with this form of data collection is feasibility of the
process. This issue would arise if the study takes too much of the teacher’s time. To limit this
problem, each data collection is limited to a maximum of 30 minutes.
The sensitivity of the method chosen to collect data is another consideration. This
research is confident that the three problems under study occur frequently so the 30 minutes
time limit collects credible data.
Analysis of the Data
In the engagement with Kaya, the goal or concern is whether the intervention is
providing any improvements. If there is no meaningful change in the girl’s behavior, either an
increased intensity of the intervention or its complete abandonment will be considered. One
way of helping in making this decision is a simple visual examination of the graphed data.
There are also other options such as standard deviation-band and time series, which is used in
this case. In using time series analysis, this research seeks to ascertain the clinical
significance of the results of the intervention.
From visual inspection of the graphical results, the intervention seems to work for a
single behavior only. Graph 1 shows that the frequency of off-task increased on Wednesday
and Friday, with the greatest intensity on Friday. The same case applies for verbal aggression,
which increased both on Wednesday and on Friday. It is only in physical aggression that
Kaya showed some progress on Friday, where the frequency of the behavior reduced
drastically from 20% on Wednesday to 7% on Friday.

Conducting Time Series Research 4

Implication and Conclusion
The apparent reduction in physical aggression from Wednesday to Friday indicates
that the intervention on Kaya can solve this antisocial behavior within a matter of days.
Alternatively, the increase in verbal and off-task misbehaviors from Monday to Friday is a
pointer that this intervention is not appropriate for such behaviors for short-term results.
Therefore, this paper suggests further research regarding the two behaviors that failed to
respond to the intervention with the aim of determining factors that aggravate them, and their
relation to the intervention.
The time-series model used here is restricted to the class setting and is done for three
days only. A better method is to take daily totals of the aggressive behaviors’ data and then
analyze those using methods such as multilevel modeling and growth curve analysis. As
Nugent & Ely attest, this would introduce the important aspect of systematic differences
(2010). Therefore, this paper recommends the use of such models in further research on
aggression behavior intervention.

Conducting Time Series Research 5


Borckardt, J. J. et al (2008). Clinical practice as natural laboratory for psychotherapy
research-a guide to case-based time-series analysis. American Psychologist, 63(2), 77-

  1. (Borckardt et al, 2008)
    Horner, R. H. (2005). The use of single-subject research to identify evidence-based practice
    in special education. Council for Exceptional Children, 71(2), 165-179.
    Nugent, W. R., & Ely, G. (2010). The effects of aggression replacement training on
    periodicities in antisocial behavior in a short-term shelter for adolescents.Journal of
    the Society for Social Work and Research, 1(3), 140-158.