MED106 1b Types of Variables and Basic Summary Statistics PDF

Introduction to measurement IA: types of variables and proportions Avgis Hadjipapas Professor for Neuroscience and Research Methods [email protected]  Session LOBs LOB1: Outline the different types of variables in biom...

Introduction to measurement IA: types of variables and proportions Avgis Hadjipapas Professor for Neuroscience and Research Methods [email protected]  Session LOBs LOB1: Outline the different types of variables in biomedical research. LOB2: Determine the applicability of different summary statistics for descriptive analysis of data. Gender City Ethnicity Smoking Alcohol hyperchole status consumptio s n terolaemia M Nicosia Cypriot current no no M Limassol non-Cypriot current moderate no F Limassol Cypriot current moderate no M Nicosia Cypriot current moderate no F Larnaca Cypriot ex heavy no F Larnaca Cypriot ex moderate no F Paphos Cypriot current moderate no F Larnaca non-Cypriot current heavy yes F Nicosia Cypriot no no no F Larnaca non-Cypriot no no yes F Paphos Cypriot current heavy no M Nicosia Cypriot current heavy no M Limassol Cypriot ex heavy yes Age (years) Body Height (m) Systolic Fasting Portions weight blood blood of fruits (kg) pressure glucosse and (mmHG) (mmol/L) vegetables per day 20 60 1.70 99 4.9 2 22 66 1.77 90 5.5 5 25 69 1.60 105 5.8 10 21 80 1.80 80 6.0 3 23 48 1.49 88 4.8 6 24 65 1.55 96 5.2 8 26 68 1.66 102 5.4 12 19 54 1.60 110 5.7 0 26 51 1.59 118 6.1 7 22 89 1.51 125 6.8 4 24 52 1.55 98 5.7 3 18 70 1.79 90 5.9 3 20 98 1.85 120 7.0 1 What is the main difference between these 2 sets of variables??  Types of variables One way of categorizing (more popular in the UK) numeric categorical normal ordered binary continuous discrete categorical categorical  Types of variables (A) Numeric: describes a measurable quantity – Continuous: a variable that can take any possible value within a given range, e.g. Temperature in Celsius – Discrete: a variable that can take specific values within a given range (i.e. there are gaps in the range of variables) – typical discrete variable is a count (of integers)- e.g. number of teeth Categorical: sorted to categories based on qualitative description – Normal categorical: a variable that is divided into different categories not ordered in any way, e.g. brand of car – Ordered categorical: a variable that is divided into different categories in an ordered manner, e.g. educational level (elementary school, high school, university) – Binary: a categorical variable with only 2 categories, e.g. affected by disease (yes vs no)  Types of variables Another way of categorizing (more popular in the US) numeric categorical interval ratio nominal ordinal scale scale scale scale  Types of variables (B) Numeric: – Interval scale: a variable whose values are graded in equal increments (i.e. the distance from one value to the next is equal along the whole range of values) e.g. Temperature in Celsius (what does 0 C mean?) – Ratio scale: Same as interval scale, however a true 0 exists (true absence of quantity), e.g. Temperature in Kelvin (what does 0 Kelvin mean?) Categorical: – Nominal scale: a variable that is divided into different categories not ordered in any way, e.g. country of origin – Ordinal scale: a variable that is divided into different categories in an ordered manner, e.g. income categorized as (“low income”,”middle income”,”high income”)  Types of variables blood triglyceride levels ? dietary pattern (very healthy, moderately healthy, moderately unhealthy, very unhealthy) ? Cyprus district (Nicosia, Limassol, Larnaca, Paphos, Famagusta, Kerynia) ? number of asthma attacks per month? - Please note that type of variable very much depends on how it was defined and used in a particular study - E.g. Compare blood pressure variable defined as hypertensive (yes vs no) (binary categorical), versus Blood Pressure in mm Hg (numeric continuous) Summarizing categorical variables: proportions  Summarising categorical data  Actual (absolute) numbers are rarely informative and in general do not aid comparisons and may be misleading! Example 1 Most cars involved in accidents in New Zealand are white (Furness et al, 2003) Are white cars more likely to be involved in accidents? Example 2 There are 900,000 diabetics in Greece and 1,800,000 diabetics in Germany In which of the two countries is diabetes more frequent?  Basic summary statistics for categorical variables: proportion  We can present categorical variables by counting the number of data values in each category of interest We can organize these counts into a frequency table, which displays the numbers and proportion for each category Example Pulmonar No. of Relative Proportion (%) y cases tumour frequenc y Benign 32 0.8 80  Presenting categorical data with graphs Bar chart Pie chart  Bar Chart  A bar chart displays the distribution of a categorical variable, showing the counts for each category next to each other for easy comparison  Same widths, equal spaces between bars Proportion with chronic disease by age 65-74 75-85 ≥85 Bar Charts can be clustered..  Pie Chart  Pie charts show the total population/sample as a circle They slice the circle into pieces whose size is proportional to the fraction of the whole in each category Age of persons with Alzheimer’s Disease  Homework 1 (proportions, bar charts, pie charts) We have created a dataset containing the following variables:  Smoking status (never smoker / ex / current smoker)  Height (m) Using Excel, please attempt the following: 1. Summarize the variable ‘smoking status’ and create a bar chart 2. Summarize the variable ‘smoking status’ and create a pie chart Note: for calculating the frequency of each observation use the Excel function ‘COUNTIF’ and then convert that to a proportion by dividing with the total number of individuals  Session LOBs LOB1: Outline the different types of variables in biomedical research. LOB2: Determine the applicability of different summary statistics for descriptive analysis of data.  Further reading (optional) Petrie A. & Sabin C. Medical Statistics at a Glance, 3rd Edition, Chapters 1, 4, 5,6 [ISBN : 978-1-4051-8051-1] Kirkwood B. & Sterne J. Essential Medical Statistics, 2nd Edition, Chapters 3, 4, 15 [ISBN : 978-1-118-30096-1]