Instructions for Case Assignments
(50% of course grade)
General Instructions
● If you are part of a research group, work on the case assignments within your group.
If there are more than 4 members in the group, break up into 4 or less people.
● If you are working on a case report as your research option, you can choose to work
alone or form a group (of 4 or less people) with others who are also working on case
reports.
● Stay in your group (or work alone) for all the cases.
Instructions for Each Case (points deducted if not followed)
● For each case, refer to your notes and handouts from Research 1, recommended
article(s) and the internet to guide your answers.
● In each case, you are given the name of a statistical test.
● Design an experiment/research study which requires the use of the statistical test for
analyses.
● The experiment must be related to your research project topic -- the age group of your
subjects, the disease or impairment, and the evaluation (balance control, exercise,
craniovertebral stability, etc) etc.
● Cite your sources/references (handouts in Research 1, textbook, recommended
article(s) and internet). There is no limit to the number of resources that you should use.
● Use the APA style for in-text citations and Reference list.
● Use the future tense in your answers.
Submitting your work
● Save each work as a Word document using the last and first names of the group
members and the case number, e.g., "ChongRaymondBlaggMargaretCase1".
● Email the document to me on or before the stated deadline.
● Copy everyone in the group.
Good to know
● You may ask for my written feedback by submitting a complete draft of your practice
and each of the first 3 cases. For the last two cases, you can still consult with me but I
will not vet your work
● Do not ask your research advisor for help. They are not responsible for this part of the
course
How to do well in the case assignments
● Follow all instructions, including proper formatting of in-text citation and Reference list.
● Use the recommended articles to design your experiments.
● For Q1 and Q3 of each case, you do not have to explain the meaning of the statistical
test or research concepts. Rather, explain how they apply to your study. Still with me?
PRACTICE CASE
Name of statistical test: independent t-test
1a. What is the purpose of using this statistical test for your hypothetical experiment,
i.e., what will it help you to determine?
1b. State your research question, i.e., the purpose of your experiment. For example,
"The purpose of this experiment/study is to …. "
Hint: The statement should be stated such that it provides a clue to the reader as to the
statistical test which should be used to answer the research question.
1c. What are the appropriate PICO search terms for your research question?
1d. State the null and experimental hypotheses (based on your research question in
1b).
2. Describe your experimental design. Include the following concepts where applicable:
sampling method (purposive, random, etc); independent variable(s); age-group(s) of
subjects, dependent variable(s) (include the unit of measurement).
3. Explain what result you hope to obtain for each of the following: a) standard
deviation; b) effect size; c) p value; d) type 1 error.
4. Show a graph to illustrate how the result of your experiment would like if your
experimental hypothesis (stated in 1d) was confirmed.
Use the following article as a resource to design your hypothetical study:
Chong, R. K., Barbas, J., Garrison, K., Herolz, A., Teheng, R., & Sethi, K. (2001). Does
balance control deficit account for walking difficulty in Parkinson's disease? International
Journal of Clinical Practice, 55(6), 411-412. Available for download at the Research
website:
http://www.gru.edu/alliedhealth/pt/researchcourse/articles.php
CASE 1 (10% of course grade)
Name of statistical test: receiver operating curve (ROC) analysis
1a. What is the purpose of using this statistical test for your hypothetical experiment,
i.e., what will it help you to determine?
1b. State your research question, i.e., the purpose of your experiment. For example,
"The purpose of this experiment/study is to …. "
1c. What are the appropriate PICO search terms for your research question?
1d. State the null and experimental hypotheses (based on your research question in
1b).
2. Describe your experimental design. Include the following concepts where applicable:
sampling method (purposive, random, etc), grouping (within-subject, independent group
or mixed design); independent variable(s); age-group(s) of subjects, dependent
variable(s) (include the unit of measurement).
3. Explain what result you hope to obtain for each of the following: a) sensitivity and
specificity, b) cut-off score (include an explanation on how you will make your choice),
c) positive/negative predictive value, d) likelihood ratio (include an explanation on how
you will make your choice).
4. Show a graph to illustrate how the result of your experiment would like if your
experimental hypothesis (stated in 1d) was confirmed.
Use the following article as a resource to design your hypothetical study:
Chong, R. K., Morgan, J., Mehta, S. H., Pawlikowska, I., Hall, P., Ellis, A. V., et al.
(2011). Rapid assessment of postural instability in Parkinson's disease (RAPID): A pilot
study. European Journal of Neurology, 18(2), 260-265. doi: 10.1111/j.14681331.2010.03115.x
Available for download at
http://www.ncbi.nlm.nih.gov/pubmed/20586791 or the Research website:
http://www.gru.edu/alliedhealth/pt/researchcourse/articles.php
CASE 2 (10% of course grade)
Name of statistical test: 2 x 2 ANOVA (analysis of variance). Choose an independent,
repeated, or mixed ANOVA design, i.e., 2 x 2 repeated ANOVA, 2 x 2 mixed ANOVA or
2 x 2 independent ANOVA design.
1a. What is the purpose of using this statistical test for your hypothetical experiment,
i.e., what will it help you to determine?
Hint: Think about why we don't we do two 1-way ANOVAs instead, i.e., what can we
learn from a 2-way ANOVA that we cannot from separate 1-way ANOVAs? Also, keep
in mind that the repeated design has more statistical power and therefore you should
choose this option as the first choice (we discussed this in Research 1). If it is not
appropriate for your hypothetical study, then consider the other options. However, do
not simply pick the repeated design option just because it has more statistical power.
There are justifiable instances where the mixed or independent designs are more
appropriate instead. When you explain your choice, cite the textbook and the handouts
from Research 1.
1b. State your research question, i.e., the purpose of your experiment. For example,
"The purpose of this experiment/study is to …. "
Hint: The statement should be stated such that it provides a clue to the reader as to the
statistical test which should be used to answer the research question.
1c. What are the appropriate PICO search terms for your research question?
1d. State the null and experimental hypotheses (based on your research question in
1b).
2. Describe your experimental design. Include the following concepts where applicable:
sampling method (purposive, random, etc), grouping (within-subject, independent group
or mixed design); independent variable; age-group(s) of subjects, dependent variable(s)
(include the unit of measurement).
2b. Draw the ANOVA table showing the independent and dependent variable.
3. Explain what result you hope to obtain for each of the following: a) main effect, b)
interaction effect, c) p value, d) post-hoc test(s). Explain your answers.
4. Show a graph to illustrate how the result of your experiment would like if your
experimental hypothesis (stated in 1d) was confirmed. Refer to the ANOVA graphs in
Research 1 to help you.
Use the following articles as resources to design your hypothetical study:
Chong, R. K., Mills, B., Dailey, L., Lane, E., Smith, S., & Lee, K. H. (2010).
Specific interference between a cognitive task and sensory organization for stance
balance control in healthy young adults: visuospatial effects. Neuropsychologia, 48(9),
2709-2718. Available for download at
http://www.ncbi.nlm.nih.gov/pubmed/20478320?dopt=Citation or the Research
website: http://www.augusta.edu/alliedhealth/pt/researchcourse/articles.php
Chong, R. K., Chastan, N., Welter, M. L., & Do, M. C. (2009). Age-related
changes in the center of mass velocity control during walking. Neuroscience Letters,
458(1), 23-27. Available for download at
http://www.ncbi.nlm.nih.gov/pubmed/19442871?dopt=Citation or the Research
website: http://www.gru.edu/alliedhealth/pt/researchcourse/articles.php
CASE 3 (10% of course grade)
Name of statistical test: multiple regression analysis
1a. What is the purpose of using this statistical test for your hypothetical experiment,
i.e., what will it help you to determine?
Hint: Think about why we don't do separate simple (single) regression tests instead, i.e,
what can we learn from doing a multiple regression test that we cannot with multiple
simple regression tests?
1b. State your research question, i.e., the purpose of your experiment. For example,
"The purpose of this experiment/study is to …. "
Hint: The statement should be stated such that it provides a clue to the reader as to the
statistical test which should be used to answer the research question.
1c. What are the appropriate PICO search terms for your research question?
1d. State the null and experimental hypotheses (based on your research question in
1b).
2. Describe your experimental design. Include the following concepts where applicable:
sampling method (purposive, random, etc), grouping (within-subject, independent group
or mixed design); independent variables; age-group(s) of subjects, dependent
variable(s) (include the unit of measurement).
Hint: Come up with predictor (independent) variables that you think are not correlated
with each other.
3. Explain what result you hope to obtain for each of the following: a) r, b) r2, c) forward
stepwise analysis, d) criterion variable, e) predictor variable.
4. Show a graph to illustrate how the result of your experiment would like if your
experimental hypothesis (stated in 1d) was confirmed. Refer to the Jacobs article to
help you.
Use the following article as a resource to design your hypothetical study:
Jacobs, J. V., Horak, F. B., Tran, V. K., & Nutt, J. G. (2006). Multiple balance tests
improve the assessment of postural stability in subjects with Parkinson's disease.
Journal of Neurology, Neurosurgery and Psychiatry, 77(3), 322-326. Available for
download at
http://www.ncbi.nlm.nih.gov/pubmed/16484639?dopt=Citation or the Research
website:
http://www.gru.edu/alliedhealth/pt/researchcourse/articles.php
CASE 4 (10% of course grade)
Name of statistical test: factor analysis
1a. What is the purpose of using this statistical test for your hypothetical experiment,
i.e., what will it help you to determine?
Hint: Think about the rationale of doing the factor analyses statistical test not in terms of
the statistical purpose but the clinical application.
1b. State your research question, i.e., the purpose of your experiment. For example,
"The purpose of this experiment/study is to …. "
Hint: The statement should be stated such that it provides a clue to the reader as to the
statistical test which should be used to answer the research question.
1c. What are the appropriate PICO search terms for your research question?
1d. State the null and experimental hypotheses (based on your research question in
1b).
2. Describe your experimental design. Include the following concepts where applicable:
sampling method (purposive, random, etc), grouping (within-subject, independent group
or mixed design); independent variable(s); age-group(s) of subjects, dependent
variables (include the unit of measurement).
Hint: Come up with about 9-12 dependent variables divided into groups of 3 or 4. Within
each group (i.e., factor), the variables should be highly correlated with each other but
not with the variables in the other groups (factors). Go over the Chong resource article
cited below and see what variables were used. Don't copy the variables blindly but use
them to help you think about how to come up with highly correlated variables in your
hypothetical study. After that, you need to label each factor that represents the theme of
the variables in each factor. See how the Chong resource article label their factors.
Illustrate your factors using a table. You should also refer to the Research 1 handout on
Factor Analysis and follow the procedures.
3. Explain what result you hope to obtain for each of the following: a) factor loading, b)
explained variance, c) Cronbach's alpha.
4. Show an example table of the above results (refer to the Factor Analysis table from
Research 1 to help you).
Use the following article as a resource to design your hypothetical study
Chong, R. K. (2008). Factor analysis of the functional limitations test in healthy
individuals. Gait & Posture, 28(1), 144-149. Available for download at
http://www.ncbi.nlm.nih.gov/pubmed/18164620?dopt=Citation or the Research
website: http://www.gru.edu/alliedhealth/pt/researchcourse/articles.php
CASE 5 (10% of course grade)
Name of statistical test: reliability. Choose intra-rater, inter-rater, subject test-retest or
instrument test-retest reliability.
1a. What is the purpose of using this statistical test for your hypothetical experiment,
i.e., what will it help you to determine?
1b. State your research question, i.e., the purpose of your experiment. For example,
"The purpose of this experiment/study is to …. "
1c. What are the appropriate PICO search terms for your research question?
1d. State the null and experimental hypotheses (based on your research question in
1b).
2. Describe your experimental design. Include the following concepts where applicable:
sampling method (purposive, random, etc), grouping (within-subject, independent group
or mixed design); independent variable(s); age-group(s) of subjects, dependent
variable(s) (include the unit of measurement), gold standard of comparison.
Include an explanation of how you will control for the other factors of reliability. For
example, if you choose to work on intra-rater reliability, explain how you will minimize or
control for variability between raters, subjects, equipment, etc which may confound the
study.
Hints:
If you choose …
reliability of the rater/raters
reliability of the subjects
reliability of a test/equipment
Good to know …
● You can't test/retest the raters if the equipment is automated/computerized
unless the rater is involved in some way which causes the equipment to take
bad readings (e.g. incorrect placement of a sensor on the subject's body).
● Explain how subjects will produce a consistent performance across multiple
tests. You can however, carry out a reliability of the rater/s if you are
comparing for example, experienced versus inexperienced raters. Also,
explain how the variance of the instrument/equipment and the subjects will
not confound the study.
● Explain how the rater/s and the equipment will both be reliable so that the
majority of the variance is primarily attributed to the subjects.
● Explain how the subjects' performances will be consistent and how the
rater/s across multiple measurements will be reliable.
► Make sure you're not mixing two or three reliability tests in the same study, otherwise
you can no longer tell where the source(s) of variance came from.
3. Explain what result you hope to obtain in regards to the ICC value.
Use the following articles as resources to design your hypothetical study:
Hsu, A. L., Tang, P. F., & Jan, M. H. (2002). Test-retest reliability of isokinetic
muscle strength of the lower extremities in patients with stroke. Archives of Physical
Medicine and Rehabilitation, 83(8), 1130-1137. Available for download at
http://www.ncbi.nlm.nih.gov/pubmed/12161836 or the Research website:
http://www.gru.edu/alliedhealth/pt/researchcourse/articles.php
Ryan, C. G., Grant, P. M., Tigbe, W. W., & Granat, M. H. (2006). The validity and
reliability of a novel activity monitor as a measure of walking. British Journal of Sports
Medicine, 40(9), 779-784. doi: 10.1136/bjsm.2006.027276. Available for download at
http://www.ncbi.nlm.nih.gov/pubmed/16825270 or the Research website:
http://www.gru.edu/alliedhealth/pt/researchcourse/articles.php