DATA ANALYSIS & STATISTICAL PACKAGES. Daniel Inusa Yakmut ICT Directorate Federal University Lafia

Transcription

1 DATA ANALYSIS & STATISTICAL PACKAGES Daniel Inusa Yakmut ICT Directorate Federal University Lafia

2 Expectations Understand the basic principle of Data Analysis Relationship of Research Design & Type of Data Not to be an expert statistician Not to become an expert in any statistical Package Understand the kind of statistics to run Understand the statistical package to use

3 Introduction The lecture will give an overview on the following: What is Data Analysis Types of Research Design Types of Data Analysis of Data Types of Statistical Tests Statistical Packages (Software)

4 What is Data Analysis? Data Analysis the process of systematically examining data with the purpose of producing useful information Complete Analysis of Data will enable a researcher to: Determine the impact of his/her work Assess the quality of his/her programming Communicate results to stakeholders

5 Research Designs A research design is chosen to align with the purpose of the research and the type of data to be collected. Research design is broadly situated as follows: True experimental Quasi-experimental Non-experimental

6 Types of Data There are two types of : Qualitative consists of word and narratives Quantitative numerical in nature

7 Qualitative Data Analysis of Qualitative data is in form of: highlighting keywords extracting themes elaborating on concepts Strengths and Limitations

8 Quantitative Data Analysis of quantitative data requires statistical techniques. Strengths and Limitations

9 Types of Variables Questions asked in a survey are referred to as Variables Type of variables leads to kind of data and guides type of analysis. There majorly two types of variables Categorical variables Continuous Variables

10 Categorical Variables Categorical variable is made of categories typically a set of categories a respondent can select from, with each category distinct from the other e.g. marital status. It can be in the form: Binary this variable lists two distinct mutually exclusive choices, e.g gender (male or female). Nominal usually lists more than two categories to select from, e.g. what is your religion (christianity, islam, african religion, atheist), short, dwarf)? Ordinal a variable with categories that can be put in a logical order, but it does not tell the differences between the categories, e.g. describe your income (high, medium, low)

11 Continuous Variables Continuous variable takes any score or value within a measurement scale. The difference between each value has real meaning e.g temperature, height and weight. It has two types: Interval variable it can be ordered, distance or level between each category is equal and static, e.g. what is the average daytime temperature during the rains in lafia (28deg, 30deg, 32deg, 34deg) Ratio variable similar to interval variable, with one difference the ratio of the scores makes sense, and have clear 0 point, e.g on a scale 0 10 what is you stress level?

12 Analysis of Data Quantitative Data use numeric responses to variables. The numeric data is analysed. the type of analysis depends on research design, type of variables and distribution of data. There are two types of quantitative analysis: Descriptive Inferential

13 Descriptive Analysis Descriptive Analysis - the analysis reveals the basic qualities of the data. Descriptive analysis includes descriptive statistics such as: range minimum maximum frequency measure of central tendency mean, media, mode standard deviation

14 Data example Descriptive Analysis

15 Descriptive Analysis Measure of central tendency this gives an idea of how participants are responding in general Measures of Central Tendency Example Mean: The average or the sum of the values divided by the number of values. In the scores above the Mean = 11.1 ( )/9 Median: The middle score of data after they are arrange in numerical order. in the scores above the Median = 10 Mode: The most frequently occurring score in a data set. In the scores above the Mode = 5

16 Descriptive Analysis In descriptive statistics it may be required to also find how far apart or close together responses are. To do, that Standard Deviation will be require. Standard Deviation shows how well the mean represents all of the data, which represents the average amount that a given score deviates from the mean score.

17 Inferential Analysis In inferential analysis, inferences are made based on the data. Statistical tests are use to check whether an observed pattern is due to chance or the program or intervention effects. Inferential analysis is used to determine if there is a relationship/strength of the relationship between an intervention and an outcome

18 Inferential Analysis The first step of inferential analysis is to show how the data distribution looks. Always do this, before doing an inferential analysis! The type of test to be conducted is guided by the distribution of the data. Data distribution falls into two categories: Normal Non - normal

19 Inferential Analysis Normal Distribution it looks like a Bell Curve A curve drawn over the distribution and it fits indicate normal distribution. If data is normal, it shows that the data is clustered one number or value If data is normal, we use parametric tests.

20 Inferential Analysis Non-Normal Distributions an usual set of responses are causes for data not to be normally distributed. If data is non-normal, we use non parametric test. The two types of non-normal distribution: Skewed Kurtosis

21 Inferential Analysis (Types of Non Normal Distribution) Skewed It does not take bell shape, but distribution is skewed negatively or positively Negatively Skewed most scores are at the higher end of possible scores Positively Skewed most scores are at the lower end of possible scores

22 Inferential Analysis (Types of Non Normal Distribution) Kurtosis this is when a distribution is either too peaked (pointy) or too flat. Pointy Flat

23 Types of Statistical Tests There are a wide range of statistical tests, which test to use depends: research design distribution of data type of variable If data is normally distributed, choose test from parametric tests. If data is non-normal, choose test from non-parametric tests.

24 Types of Statistical Tests (non-parametric tests)

25 Types of Statistical Tests (non-parametric tests)

26 Types of Statistical Tests (non-parametric tests)

27 Types of Statistical Tests (non-parametric tests)

28 Choosing a Statistical Test To choose any of test described earlier, there are some basic question a researcher needs to needs to ask:

29 Statistical Test Example The table below shows an example of how to select a statistical test:

30 Statistical Packages These are computer software that can assist in the analysis of quantitative and qualitative. Examples of Statistical packages: Proprietary MS Excel SPSS Minitab SAS Free Software LibreOffce Calc EpiInfo PSSP R

31 Mostly used in Academia Use of data analysis software in academic publications as measured by hits on Google Scholar.

32 Microsoft Excel

33 Microsoft Excel COST Individual License for Microsoft Office Professional $350 Microsoft Office University Student License: $99 Volume Discounts available for large organizations and universities Free Starter Version available on some new PCs PRO Nearly ubiquitous and is often pre-installed on new computers User friendly Very good for basic descriptive statistics, charts and plots CON Costs money Not sufficient for anything beyond the most basic statistical analysis

34 Statistical Packages for Social Sciences (SPSS)

35 SPSS COST From $1000 to $12000 per license depending on license type. CON Very expensive Not adequate for modeling and cutting edge statistical analysis PRO Easy to learn and use More powerful then Minitab One of the most widely used statistical packages in academia and industry Has a command line interface in addition to menu driven user intefrace One of the most powerful statistical package that is also easy to use.

36 Minitab

37 SAS

38 SAS COST Complicated pricing model $8,500 first year license fee CON Very very expensive Not user friendly Steap learning curve Relatively poor graphics capabilities PRO Widely accepted as the leader in statistical analysis and modeling Widely used in the industry and academia Very flexible and very powerful.

39 LibreOffice Calc

40 LibreOffice Calc LibreOffice is a free and open source office suite, developed by The Document Foundation. It is descended from OpenOffice.org, from which it was forked in 2010 OpenOffice vs LibreOffice Star Sun Oracle Apache, Document Foundation OpenOffice LibreOffice

41 EpiInfo

42 EpiInfo Epi Info is public domain statistical software for epidemiology developed by Centers for Disease Control and Prevention (CDC) Epi Info has been in existence for over 20 years and is currently available for Microsoft Windows. The program allows for electronic survey creation, data entry, and analysis. Within the analysis module, analytic routines include t-tests, ANOVA, nonparametric statistics, cross tabulations and stratification with estimates of odds ratios, risk ratios, and risk differences, logistic regression (conditional and unconditional), survival analysis (Kaplan Meier and Cox proportional hazard), and analysis of complex survey data. The software is in the public domain, free, and can be downloaded from Limited support is available

43 EpiInfo PRO Consists of multiple modules to accomplish various tasks beyond just statistical analysis. ability to rapidly develop a questionnaire customize the data entry process quickly enter data into that questionnaire analyze the data COST Free CON Not a dedicated statistical package Not as powerful as commercial alternative for performing advanced analysis and modeling

44 PSPP

45 COST Free PRO PSPP Last version released in 2010 Aims as a free SPSS Not very well known nor alternative with an interface widely used that closely resembles SPSS User friendly Good enough for basic statistical analysis CON Lacks many advanced statistical tests and features that are present in SPSS

46 R

47 R R provides a wide variety of statistical and graphical techniques, including linear and nonlinear modeling, classical statistical tests, time-series analysis, classification, clustering, and others. R is easily extensible through functions and extensions, and the R community is noted for its active contributions in terms of packages. There are some important differences, but much code written for S runs unaltered. Many of R's standard functions are written in R itself, which makes it easy for users to follow the algorithmic choices made. R is highly extensible through the use of user-submitted packages for specific functions or specific areas of study. Due to its S heritage, R has stronger object-oriented programming facilities than most statistical computing languages. Extending R is also eased by its permissive lexical scoping rules.[10] According to Rexer's Annual Data Miner Survey in 2010, R has become the data mining tool used by more data miners (43%) than any other.[11] Another strength of R is static graphics, which can produce publication-quality graphs, including mathematical symbols. Dynamic and interactive graphics are available through additional packages.[12]

48 R PRO Widely used and accepted in industry and academia Very powerful and flexible Very large user base Lots of books and manuals Several User Interface Shells available COST Free / Open Source CON Not user friendly Requires steep learning curve

49 Dataset The Dataset and Story Library DASL (pronounced "dazzle") is an online library of datafiles and stories that illustrate the use of basic statistics methods. We hope to provide data from a wide variety of topics so that statistics teachers can find real-world examples that will be interesting to their students. Use DASL's powerful search engine to locate the story or datafile of interest.

50 Brain Size and Intelligence Are the size and weight of your brain indicators of your mental capacity? In this study by Willerman et al. (1991) the researchers use Magnetic Resonance Imaging (MRI) to determine the brain size of the subjects. The researchers take into account gender and body size to draw conclusions about the connection between brain size and intelligence. Methods Correlation Regression Scatterplot

51 Brain Size and Intelligence Description: Willerman et al. (1991) collected a sample of 40 right-handed Anglo introductory psychology students at a large southwestern university. Subjects took four subtests (Vocabulary, Similarities, Block Design, and Picture Completion) of the Wechsler (1981) Adult Intelligence Scale-Revised. The researchers used Magnetic Resonance Imaging (MRI) to determine the brain size of the subjects. Information about gender and body size (height and weight) are also included. The researchers withheld the weights of two subjects and the height of one subject for reasons of confidentiality. Number of cases: 40 Variable Names: Gender: Male or Female FSIQ: Full Scale IQ scores based on the four Wechsler (1981) subtests VIQ: Verbal IQ scores based on the four Wechsler (1981) subtests PIQ: Performance IQ scores based on the four Wechsler (1981) subtests Weight: body weight in pounds Height: height in inches MRI_Count: total pixel Count from the 18 MRI scans

52 Conclusion Statistical analysis is an integral part of any study and publication While commercial statistical software may cost an arm and a leg, free alternatives do exists. While some free alternatives don't measure up, others are growing and expending rapidly and may overtake commercial software in features and popularity

53 References

54

55