Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
12 views

Data Processing and Analysis

Uploaded by

melesse jenbolla
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
12 views

Data Processing and Analysis

Uploaded by

melesse jenbolla
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 19

Methods of Data

Processing & Analysis

By: Ephrem Mannekulih


(MSc. in Biostatistics & Health/Inf., Asst. Prof.)
Prerequisites for analysis
1. More focus to the objectives of study
2. Knowledge of type of variables and measurement scale
(dependent Vs. independent, qualitative Vs.
quantitative…)
3. Assessing the nature and distribution of data
4. Knowledge of statistical analysis to be done
Focus on study objectives
A research is made principally to answer study questions
We should be aware of;
 Results should answer the objectives (study questions)
 Discussion should interpret what it mean by the results answering
the objectives
 Conclusion should be based on the answer to the objectives
 Recommendation also should be based on finding but not on wish
Cont’d….
Results should answer the objectives (study questions)
Eg.
 To determine prevalence of TB in a community
 Assess factors associated with HIV/ AIDS
 Measure effect of multiple partner on HIV/AIDS prevalence
Knowledge about variables
Before performing any analyses, you must first get to
know;
 Types of variables and
 Measurement scales of your data.
Because data analysis, summarization and presentation
techniques are depends on:
 Both types of variable and data measurement scales
Variable can also be classified into dependent and
independent
Assessing Statistical Assumptions
Based on probability distribution of a random variable
statistical methods are classified as;
 Parametric method
 Used when the measurements of the groups are normally distributed,
 Non-parametric method
 Used when the measurements of the groups are not normally distributed
7 How to Check the Assumptions
Normality of data can be assessed using:
1. Using of graphs
 Histogram, normal Q-Q plot, steam & leaf plot, box & whiskers plot,
frequency polygon
2. Comparison of statistic like mean, mode and median or using
skewedness measure
3. Using statistical tests for normality
 Shapiro-Wilk, Kolmogorov-Smironov

AHMC 9/26/2024
8 Assumptions…Cont’d
Normality assumption
 For example;
 The distribution of response variable within each treatment group should be
normally distributed.

AHMC 9/26/2024
9 Assumptions…Cont’d
Equality of variances

 With larger variances, you can expect a greater number of


observations at the extremes of the distributions
 This can have real implications on inferences we make from
comparisons between groups

AHMC 9/26/2024
10 Assumptions…Cont’d
Testing equality of variances: homoscedasticity
 Homogeneity of variances is determined by the following
approaches:
 Comparison of graphs
 Box plots or frequency polygon
 Comparison of variance, standard deviation and IQR
 Using statistical tests for homogeneity of variances
 Levene’s test
Other statistical Assumption and tests??????????????

AHMC 9/26/2024
Stages of Data Processing & Analysis

Perform data processing tasks


 Editing
 Coding
 Entry to a computer
 Transforming
 Classification or categorizing
 Cleaning
 Computing
 Counting
 Ranking etc
Stages of Data analysis
Descriptive data analysis
 Frequency distribution or tabulation
 Absolute frequency
 Relative frequency
 Cumulative frequency
 Numerical summary measures
 Measure of central tendency,
 Measure dispersion,
 Measure of relative standing and shape
Stages of Data analysis
Inferential statistical analysis
 Estimation of parameter value
 Point estimation
 Interval Estimation
 Hypothesis testing
 Parametric
 Non-parametric
Data Analysis…Cont’d
Hypothesis testing
 It is the most common application of statistical analysis in research
setting
 The purposes are;
1. Measuring the differences between group means/proportions
2. To test whether or not the difference exists between two or more groups
means/proportions is statistically significant
Data Analysis…Cont’d
Statistical methods to test the difference of group means
I. Parametric Methods
One sample t-test: used to test or compare a single sample
mean with a known/hypothesized population mean,
Two sample t-test: used to measure and test the differences
between two groups mean.
Analysis of variance (ANOVA): used to measure and test
differences among group means.
 Where the groups are three and more
Cont’d…
Statistical methods to test the difference of group means
Non-Parametric Methods
Sign Test for single sample
Sign Test for paired samples For Ordinal Scale Data
Wilcoxon signed-rank test, for paired samples.
Wilcoxon rank sum test (Mann-Whitney test) for
independent samples
Kruskal-Wallis Test: to compare differences among group
means where the groups are three and more
Data Analysis…Cont’d
Causal analysis
 Correlation and cross tabulation
 Pearson, Spearman correlation
 Chi-Squere, MackNeimar, Fiesher
 Regression
 Linear
 Logistic
 Cox…..
Data Presentation Techniques
Data presentation techniques for qualitative variables
 Tabulation
 Figures/charts
 Pie chart
 Bar chart
Data Presentation Techniques
Data presentation techniques for quantitative variables
 Figure/chart for quantitative data
 Histogram
 Stem-and-leaf plot
 Box plot
 Scatter plot
 Line graph
 Others

You might also like