More Related Content
What's hot (20)
Similar to Basics of statistics (20)
More from donthuraj (20)
Recently uploaded (20)
Basics of statistics
- 1. Basics of statistics Conducted by Dept. of Biostatistics, NIMHANS From 28 to 30 Sept, 2015 "It is easy to lie with statistics, But it is hard to tell the truth without statistics." –Andrejs Dunkels
- 2. Topics covered • Introduction • Types of statistics • Definitions • Variable & Types • Variable scales • Description of data • Distribution of sample & population • Measures of center, dispersion & shape • Properties of Normal distribution • Testing of hypothesis • Types of error • Estimation of sample size • Various tests to be used • Central limit theorem • Parametric tests- t-test, ANOVA, Post Hoc, Correlation & Regression • Non Parametric tests • Tests for categorical data • Summary of tests to be used • Qualitative vs Quantitative research • Qualitative research • Software packages
- 3. Statistics • Consists of a body of methods for collecting and analyzing data. • It provides methods for- – Design- planning and carrying out research studies – Description- summarizing and exploring data – Inference- making predictions & generalizing about phenomena represented by data.
- 4. Types of statistics • 2 major types of statistics • Descriptive statistics- It consists of methods for organizing and summarizing information. – Includes- graphs, charts, tables & calculation of averages, percentiles • Inferential statistics- It consists of methods for drawing and measuring the reliability of conclusions about population based on information obtained. – Includes- point estimation, interval estimation, hypothesis testing. • Both are interrelated. Necessary to use methods of descriptive statistics to organize and summarize the information obtained before methods of inferential statistics can be used.
- 5. Population & Sample • Basic concepts in statistics. • Population- It is the collection of all individuals or items under consideration in a statistical study • Sample- It is the part of the population from which information is collected. • Population always represents the target of an investigation. We learn about population by sampling from the collection.
- 6. • Parameters- used to summarize the features of the population under investigation. • Statistic- it describes a characteristics of the sample, which can then be used to make inference about unknown parameters.
- 7. Variable & types • Variable- a characteristic that varies from one person or thing to another. • Types- Qualitative/ Quantitative, Discrete/ Continuous, Dependent/ Independent • Qualitative data- the variable which yield non numerical data. – Eg- sex, marital status, eye colour • Quantitative data- the variables that yield numerical data – Eg- height, weight, number of siblings.
- 8. • Discrete variable- the variable has only a countable number of distinct possible values. – Eg- number of car accidents, number of children • Continuous variable- the variable has divisible unit. – Eg- weight, length, temperature. • Independent variable- variable is not dependent on other variable. – Eg- age, sex. • Dependent variable- depends on the independent variable. – Eg- weight of a newborn, stress
- 9. Variable scales • Variables can also be described according to the scale on which they are defined. • Nominal scale- the categories are merely names. They do not have a natural order. – Eg- male/female, yes/no • Ordinal scale- the categories can be put in order. But the difference between the two may not be same as other two. – Eg- mild/ Moderate/ Severe.
- 10. • Interval scale- the differences between variables are comparable. The variable does not has absolute zero. – Eg- temperature, time • Ratio scale- the variable has absolute zero as well as difference between variables are comparable.. – Eg- stress using PSS, insomnia using ISI • Nominal & Ordinal scales are used to describe Qualitative data. • Interval & Ratio scales are used to describe Quantitative data.
- 11. Describing data • Qualitative data- – Frequency- number of observations falling into particular class/ category of the qualitative variable. – Frequency distribution- table listing all classes & their frequencies. – Graphical representation- Pie chart, Bar graph. – Nominal data best displayed by pie chart – Ordinal data best displayed by bar graph
- 12. • Quantitative data- – Can be presented by a frequency distribution. – If the discrete variable has a lot of different values, or if the data is a continuous variable then data can be grouped into classes/ categories. – Class interval- covers the range between maximum & minimum values. – Class limits- end points of class interval. – Class frequency- number of observations in the data that belong to each class interval. – Usually presented as a Histogram or a Bar graph.
- 13. Population & Sample distribution • Population distribution- frequency distribution of the population. • Sample distribution- frequency distribution of the sample. • Sample distribution is a blurry photo of the population distribution. • As the sample size ↑, the sample distribution becomes closer representative of the population distribution. • Sample of population distribution can be summarized by describing its shape (based on the graph). • It can be Symmetric or Nonsymmetric/ Skewed to left/ right based on its tail.
- 14. Properties of Numerical data & Measures Central tendency Mean Median Mode Dispersion Range Interquartile Range Standard Deviation Shape Skewness Kurtosis
- 15. Measures of center • Central tendency- In any distribution, majority of the observations pile up, or cluster around in a particular region. – Includes- Mean, Median & Mode. • Mean- sum of observed values in a data divided by the number of observations • Median- observation in the data set that divides the data set into half. • Mode- value of the data set which occurs with greatest frequency • Mean & Median can be applied only to Quantitative data • Mode can be used either to Qualitative or Quantitative data.
- 16. What to choose? • Qualitative variable- Mode. • Quantitative with symmetric distribution- Mean. • Quantitative with skewed distribution- Median. • Outlier- observation that falls far from the rest of the data. Mean gets highly influenced by the outlier. • We use sample mean, median & mode to estimate the population mean, median & mode.
- 17. Measures of dispersion • Dispersion- It is the spread/ variability of values about the measures of central tendency. They quantify the variability of the distribution. • Measures include- – Range – Sample interquartile range – Standard deviation • Mostly used for quantitative data • Range- difference between the largest observed value in the data set and the smallest one. – So, while considering range great deal of information is ignored.
- 18. • Interquartile range- difference between the first & third quartiles of the variable. – Percentile- divides the observed values into hundredths/ 100 equal parts. – Deciles- divides the observed values into tenths/ 10 equal parts – Quartiles- divides the observed values into 4 equal parts. Q1 divides the bottom 25% of observed values from top 75%... • Standard deviation- it is a kind of average of the absolute deviation of observed values from the mean of the variable. – It is defined using the sample mean & values get strongly affected by few extreme observations.
- 19. Shape • Skewness- Lack of symmetry in distribution. It can be interpreted from frequency polygon. • Properties- – Mean, median & mode fall at different points. – Quartiles are not equidistant from median. – Curve is not symmetrical but stretched more to one side. • Distribution may be positively or negatively skewed. Limits for coefficient of skewness is ± 3. • Kurtosis- convexity of a curve. – Gives an idea about the flatness/ peakedness of the curve.
- 20. Normal distribution • Bell shaped symmetric distribution. • Why is it important? – Many things are normally distributed, or very close to it. – It is easy to work with mathematically – Most inferential statistical methods make use of properties of the normal distribution. • Mean = Median = Mode • 68.2% of the values lie within 1SD. • 95.4% of the values lie within 2SD. • 99.7% of the values lie within 3SD.
- 21. Tests to check normal distribution 1. Checking measures of Central tendency, Skewness & Kurtosis. 2. Graphical evaluation- normal plot, frequency polygon. 3. Statistical tests- – Kolmogorov-Smirnov test – Shapiro-Wilk test – Lilliefor’s test – Pearson’s chi-squared test • Shapiro-Wilk has the best power for a given significance. • If not normally distributed?- correction by transformation of the data- log transformation, square root transformation.
- 22. Hypothesis testing • Aim of doing a study is to check whether the data agree with certain predictions. These predictions are called hypothesis. • Hypothesis arise from the theory that drives the research. • Significance test- it is a way of statistically testing a hypothesis by comparing the data values. – It consists of two hypothesis- Null (H0) & Alternative hypothesis (H1). – Null hypothesis is usually a statement that the parameter has value corresponding to, in some sense, no effect. – Alternative hypothesis is a hypothesis contradicts null hypothesis. – Hypothesis are formulated before collecting the data.
- 23. • Significance test analyzes the strength of sample evidence against the null hypothesis. • The test is conducted to investigate whether the data contradicts the null hypothesis, suggesting alternative hypothesis is true. • Test statistics- statistic calculated from the sample data to test the null hypothesis. • p-value- is the probability, if H0 were true, that the test statistic would fall in this collection of values. The smaller the p-value, the more strongly the data contradicts H0. • When p-value ≤ 0.05, data sufficiently contradicts H0.
- 24. Types of error • Type I/ α error- Rejecting true null hypothesis. – We may conclude that difference is significant, when in fact there is no real difference. – It is popularly known as p-value. Maximum p-value allowed is called as level of significance. Being serious p-value is kept low, mostly less than 5% or p<0.05. • Type II/ β error- Accepting false null hypothesis. – We may conclude that difference is not significant, when in fact there is real difference. – It is also called as Power of the test & indicates sensitivity of the test. • Not possible to reduce both type I & II, So α error is fixed at a tolerable limit & β error is minimized by ↑ sample size.
- 25. Estimation of Sample size • Small sample- fails to detect clinically important effects (lack of Power) • Large sample- identify differences which has no clinical relevance. • Calculation is based on (not included formulas)- – Estimation of mean – Estimation of proportions – Comparison in two means – Comparison in two proportions • Checklist- level of significance, power, study design, statistical procedure. • Minimum sample size required for statistical analysis- 50.
- 26. Basic theorem in statistics • Central limit theorem- – States that the distribution of the sum/ average of a large number of independent, identically distributed variables will be approximately normal. • Why is this important? – Basis of many statistical procedures.
- 27. Parametric tests • These are statistical tests that makes assumptions about the parameters (defining properties). • Assumptions made are- – Data follows normal distribution. – Sample size is large enough for Central limit theorem to lead to normality of averages. – Data is not normal, but can be transformed. • Some situations where data does not follow normal distribution- – Outcome is an ordinal variable. – Presence of definite outliers – Outcome has clear limits of demarcation.
- 28. Tests to be used Scale type Permissible statistics Nominal Mode Chi-Square test Ordinal Mode/ Median Interval Mean, Standard Deviation t-test, ANOVA, Post hoc, Correlation, Regression, Ratio One sample t-Test Independent t- test Dependent t- test Compares the sample mean with the population mean Compares the means of two independent samples Compares the means of paired samples (before-after, pre-post)
- 29. ANOVA • t- Test- difference between 2 means. – If there are more than 2 means, then doing t test increases the α & β error. Which creates a serious flaw. • So when there are >2 means to be compared we use ANOVA. • Types- – One way- study effects of one factors. – Two way- study effects of multiple factors. • Assumptions of ANOVA- Normality, Linearity. • ANCOVA- It is a blend of ANOVA & Regression. In other words, measures how much 2 variables change together & how strong is the relationship.
- 30. Post Hoc • Latin phrase, means- “after this” or “after the event” • Why do Post hoc tests? – ANOVA tells whether there is an overall difference between groups, but it does not tell which specific group differed. – Post hoc tests tell where the difference occurred between groups. • Different Post hoc tests- – Bonferroni – Fisher’s least significant difference (LSD) – Tukey’s honestly significant difference (HSD) – Scheffe post hoc tests
- 31. Correlation & Regression • Correlation- denotes association between 2 quantitative variables. – Assume that the association is linear (i.e.., one variable ↑/ ↓ a fixed amount for a unit ↑/ ↓ in the other). – Degree of association is measured by a correlation coefficient, r. – r is measured on a scale from -1 through 0 to +1. – When both variables ↑, then r is + & when 1 variable ↑ and other decreases, then r is -. • Graphically- Scatter diagrams, usually independent variable is plotted against x-axis & dependent against y-axis. • Limitation- it does not say anything about Cause & Effect relationship. – Beware of spurious/ non sense correlation.
- 32. • Correlation- – Strength/ degree of association. • Regression- – Nature of association (eg- if x & y related, it means if x changes by certain amount then y changes on an average by certain amount). – Expresses the linear relationship between variables. – Regression coefficient- β – Types- Linear, Non linear, Stepwise • Regression coefficient gives a better summary of the relationship between the two variables than Correlation coefficient.
- 33. Non Parametric tests • Also called as “Distribution free tests”, because they are based on fewer assumptions. • Advantages- – When data does not follow normal distribution. – When the average is better represented by median. – Sample size is small. – Presence of outliers. – Relatively simple to conduct
- 34. Tests Characters Parametric test Non Parametric test Testing mean, a hypothesized value One sample t test Sign test Comparison of means of 2 groups Independent t test Mann Whitney U test Means of related samples Paired t test Wilcoxon Signed rank test Comparison of means of > 2 groups ANOVA Kruskal Wallis test Comparison of means of > 2 related groups Repeated measures of ANOVA Friedman’s test Assessing the relationship between 2 quantitative variables Pearson’s correlation Spearman’s correlation
- 35. Chi-Square test • Used for analysis of categorical data. • Other tests- Fisher exact probability test, McNemar’s test. • Requirements of Chi-Square- – Sample should be independent – Sample size should be reasonably large (n >40) – Expected cell frequency should not be < 5. • Yate’s correction- if expected cell frequency is < 5 • Fisher exact probability test- used when sample size is small (n < 20) • McNemar’s test- used when there are two related samples or there are repeated measurements
- 36. RR & OR • Relative Risk (RR)- – It is the ratio of incidence rate among exposed to the incidence rate among not exposed. – used in RCTs & Cohort studies – Values- <1 - risk of disease is less among exposed – >1 – risk of disease is more among exposed – =1 – equal risk among exposed & non exposed • Odds Ratio (OR)- – Ratio of odds of exposure among the cases to odds of exposure among controls. Used for rare diseases/ events – Used in case control & retrospective studies (no meaning in calculating the risk of getting the disease) – Values- >1- more among cases, <1- more among controls
- 37. Qualitative v/s Quantitative Qualitative research • Seeks to confirm hypothesis • Highly structured methods used • Uses closed ended, numerical methods of collecting data • Study design is fixed & subject to statistical assumptions Quantitative research • Seeks to explore phenomena • Semi-structured methods used • Uses open ended, textual methods • Study design is flexible, iterative & subject to textual analysis
- 38. Qualitative research • Provides complex descriptions & information about issues such as contradictory behavior, belief, opinions, emotions & relationships. • Methods used are- – Phenomenology – Ethnography – Grounded theory • Designs used- – Case studies – Comparative designs – Snapshots – Retrospective & Longitudinal studies
- 39. Statistical software packages Quantitative research • SPSS by IBM • R by R Foundation • GenStat by VSN International • Mathematica by Wolfram research • Minitab, MATLAB, Nmath Stats etc.., Qualitative research • ATLASti • NVIVO • MAXQDA • NUDist • ANTHTOPAC
- 40. "An approximate answer to the right problem is worth a good deal, more than an exact answer to an approximate problem." -- John Tukey