You are on page 1of 18

DATA ANALYSIS AND EVALUATION

NAFDAC TRAINING SERIES Keffi, 2011

Objectives
At the end of this session participants should be able to:

Identify and define the basic concepts and procedures required for data analysis and interpretation. Analyze and interpret the data collected. Draw conclusions related to the objectives of your study.

Aim of analysis
Analysis is the systematic breakdown into components parts and goes further to describe relationships.

It is a form of exploration that is ALWAYS --SYSTEMATIC; --RIGOUROUS AND --CONSISTENT

Cont:

Analysis thus systematically extract Key ideas Theories Concepts Methodological assumptions

Next step
Next step is SYNTHESIS in order to answer the research question. It entails making connections between the broken down parts at the stage of analysis Rearrange back Search for new order Produce a new pattern Bring out a new question Provide new insights

Methods of Analysis
1. Descriptive: This includes the frequency distribution, measurement of central tendency and dispersion 2. Inferential: This concentrates on the nature of the relationship between two variables and on the construction of measures of relationship. E.g. correlation, regression, significant tests

Types of Descriptive Analysis


Frequency Distribution --It reports through a straightforward calculation how the respondents are distributed on all the items of the investigation; a simple example is 20% of participants reacted to ACT within 24hrs of oral drug intake, whereas 80% did not. --This can be illustrated in a table, bar chart, histogram, graph or pie chart.

Cont: Measures of Central tendency --This is applied when you have a group of data and you wish to find the most typical value for the group or the score which all other scores are evenly clustered around. These are known as mean, mode and median.

Cont:
Measurements of Dispersion This is a type of analysis that shows the degree by which numerical data tend to spread about an average value. It is called variation, dispersion or standard deviation.

Inferential Analysis
Data Classification

--Parametric: whose parameters are normally distributed --Non-Parametric: used for ranked data and does not rely on normal distribution

The t-test It is a parametric test and appropriate to compare the difference between the mean scores of two groups of data. The data has to be of interval type and normally distributed. It can be used to find out if there is a significant difference in the unit cost between two types of contracts or two types of drug intervention.

Cont:
The chi-test It is a test of association between two sets of data which is nominal or ordinal. It predicts how many subject in each group fall into certain categories e.g. how many people react to AA immediately and how many goes into sub-acute or chronic states

The spearmen ranking correlation (rho) This is a non parametric test for measuring the difference between two groups of respondents scored on a number of issues, attitudes or factors. The significant thing here is whether there is a significant difference in their ranking

Cont:
The Pearson Product moment correlation This determines whether there is a strong relationship between two sets of data. It can be positive or negative and the strength is measured on a scale varying from plus one through zero to minus one, e.g. if speed of construction is related to the size of the project.

Some Commonly Used Statistical Tests Normal theory based test Corresponding nonparametric test Purpose of test t test for independent samples:MannWhitney U test; Wilcoxon rank-sum test Compares two independent samples

Paired t test: Wilcoxon matched pairs and signed-rank test. Examines a set of differences Pearson correlation coefficient & Spearman rank correlation coefficient: Assesses the linear association between two variables.

One way analysis of variance (F test): KruskalWallis analysis of variance by ranks Meant to Compares three or more groups Two way analysis of variance: Friedman Two way analysis of variance. Compares groups classified by two different factors

Hypothesis testing
The Null Hypothesis --Single tailed --Two tailed Errors of Hypothesis Type 1 and type 11

You might also like