Tutorhelpdesk.

co
m

NONPARAMETRIC
METHODS for Analyzing
LONGITUDINAL DATA
BIOSTATISTICS ASSIGNMENT GUIDE
Introduction Tutorhelpdesk.com

WHAT IS LONGITUDINAL DATA?

DEFINITION
Longitudinal data is data that has been collected on the same
subjects continuously and at multiple intervals. This type of data
is used to track the changes and patterns which is essential for
the field of biostatistics notably in medical and public health
research.

2
Introduction Tutorhelpdesk.com
Introduction

EXAMPLES
• Week-long recording of a patient’s blood pressure.
• Measure a child’s height and weight and compare them later
to ensure they follow the right growth pattern.
• Measuring how well a new drug is performing in comparison
to older drugs after two years of usage.

KEY FEATURES
• Observations made multiple times over a period on the same
individual.
• Correlations between measurements on the same subject.
3
Why nonparametric method is used Tutorhelpdesk.com
Introduction
for Longitudinal data?

• No Assumptions About Data Distribution: Non-

parametric techniques are techniques in which data does
not need to be normally distributed.
• Robustness: Particularly suitable for skewed data, or data
having outliers, or small sample sizes.
• Flexibility: Applicable on categorical, ordinal, and other
qualitative data.
SCENARIOS WHERE NONPARAMETRIC METHODS ARE
• When the population isUSEFUL:
small.
• When the distribution of the population from which the
sample has been drawn cannot be ascertained or, it is non-
normal.
4 • Whenever you are working with the rankings or scores or
ordinal outcome.
Common Nonparametric Techniques in Tutorhelpdesk.com
Introduction
Analysis of Longitudinal Data

Sign Test and Wilcoxon Kruskal-Wallis

Signed Rank Test Test
Friedman Test Kaplan-Meier
Estimator

5
1. Sign Test and Wilcoxon Signed Tutorhelpdesk.com
Introduction
Rank Test
• Purpose: Compare two sets of PROCEDURE
results on the same subjects – • Order the differences between
paired data analysis (Before and paired measurements.
After treatment). • Don’t consider the signs of
• Example Scenario: Conducting differences (either positive or
hypothesis whether a special diet negative).
type does or does not influence • Rather than the raw values, it
the cholesterol levels in the is the ranks that have to be
patients.
EXAMPLE IN R analyzed.
# Hypothetical cholesterol levels before and after intervention
before <- c(200, 220, 180, 210, 195)
after <- c(190, 210, 185, 200, 185)
# Wilcoxon Signed-Rank Test
6 wilcox.test(before, after, paired = TRUE, exact = FALSE)
2. Kruskal-Wallis Test Tutorhelpdesk.com
Introduction
• Purpose: Compare three or EXAMPLE IN PYTHON (USIN
more groups for ordinal or
from scipy.stats import kruskal
continuous data.
# Recovery rates in three
• Example Scenario:
treatment groups
Statistically assessing whether
group1 = [3.2, 2.8, 3.5, 3.0]
patient samples from three
group2 = [3.8, 3.7, 4.0, 4.1]
treatment groups have
group3 = [2.9, 3.1, 3.3, 3.2]
PROCEDURE
different recovery rates over
# Kruskal-Wallis Test
time.
• Analyse all the data and rank stat, p = kruskal(group1, group2,
them in order. group3)
• compare mean ranks between print(f"Kruskal-Wallis statistic:
groups. {stat}, p-value: {p}")
7
3. Friedman Test Tutorhelpdesk.com
Introduction
• Purpose: The many-measures EXAMPLE IN R
data are analyzed over multiple
# Pain scores across time points
conditions.
pain_scores <- matrix(c(4, 3, 2,
• Example Scenario: Pain scores
5, 4, 3,
are measured at 3 different
6, 5, 4),
periods post surgery on patients
nrow = 3, byrow =
(e.g., Day 1, Day 3, and Day 7).
# Perform Friedman test
PROCEDURE friedman.test(pain_scores)
• Rank data for each subject
across time points.
• Compute the test statistics
using rank sums.
8
4. Kaplan-Meier Estimator Tutorhelpdesk.com
Introduction Survival Analysis)
(Nonparametric
• Purpose: The estimate of EXAMPLE IN PYTHON
time-to-event data under no
habit of distribution. from lifelines import
• Example Scenario: A drug KaplanMeierFitter
is administered to the import matplotlib.pyplot as plt
patients, and the survival # Time-to-event data
rates are studied. time = [5, 6, 6, 2, 4, 8, 10]
event = [1, 1, 0, 1, 0, 1, 1] # 1:
PROCEDURE Event occurred, 0: Censored
• Estimate survival probability # Kaplan-Meier Estimator
at each time point. kmf = KaplanMeierFitter()
• Plot survival curves. kmf.fit(time, event)
kmf.plot_survival_function()
9 plt.title("Kaplan-Meier Survival
Curve")
Difficulties in the analysis of Tutorhelpdesk.com
Introduction
longitudinal data

• Complexity of Dependencies: Analzying the correlations between the same subject

across time is not easy.
• Reduced Power: Nonparametric tests can however be lower in terms of power as
compared to parametric tests.
• Computational Intensity: Certain techniques are computationally intensive
particularly when handling big data.

10
Strengths of Nonparametric Tutorhelpdesk.com
Introduction
Techniques

• Versatile: Can manage many types of data (ordinal and categorical).

• Fewer Assumptions: One does not need to normality or equal variance.
• Robustness: Overall it is less affected by outliers and skewed data.

TIPS FOR STUDENTS

• Understand the structure of data before
proposing the method.
• View the data using charts, like a box plot or
time series plot for any trends or patterns.
• Always check that the parametric
assumptions are met before they prefer a
nonparametric alternative.
11 • Become conversant with statistical packages
such as R, Python, SPSS, or SAS.
Looking for Tutorhelpdesk.com
Introduction
Biostatistics Assignment Help?
Our biostatistics assignment help is more than just a provider of solutions-it is your
gateway toward mastery of the subject. By entrusting us with this task, students receive
precise, well-structured solutions, which act as self-explanatory guide to understand how
to solve biostatistical problems correctly. The guidance provided will assist in completing
assignments and skills for the real-world application and knowledge of biostatistics.
Our support covers popular statistical software, including R, Python, SPSS, SAS and
Stata, ensuring you are well prepared for sound data analyses. Our team of
biostatistics experts start from the basics of what hypothesis testing is, regressions,
survivalWe
analysis, and other important topics.
cater to all study levels-from novices who have problems with the most-
basic concepts to advanced learners holding complex datasets. Empower
yourself with theoretical and practical knowledge for excelling in biostatistics,
which will not only help you through your course but also boost your
12 employment options. Choose us for dependable, educational support and
pursue your academic issues with confidence.
Conclusion Tutorhelpdesk.com

Nonparametric methods are powerful and indispensable yet highly

potent tools for longitudinal data analysis when data assumptions are
violated. They show robustness, flexibility, and applicability to small or
non-normal datasets. By knowing these methods, students will have an
easier understanding of analyzing complex biostatistical data and
improve their research and analysis skills.

13
Questions for Self-Review Tutorhelpdesk.com

1.why nonparametric methods are more appropriate for

skewed data in longitudinal studies?
2. Compare the differences between the Wilcoxon
Signed-Rank Test and the Kruskal-Wallis Test?
3.In what way does the Kaplan-Meier estimator deal
with censored data in survival analysis?

14
Recommended Textbooks Tutorhelpdesk.com

“Nonparametric Statistical Methods" by Myles Hollander

and Douglas A. Wolfe
"Applied Longitudinal Analysis" by Garrett M.
Fitzmaurice, Nan M. Laird, and James H. Ware

"Practical Nonparametric Statistics" by W.J. Conover

"Introduction to the Practice of Statistics" by Moore,

McCabe, and Craig

15
 Tutorhelpdesk.co
m
Thank You
www.tutorhelpdesk.co
m
+1-617-807-0926

[email protected]
m