Type of data FETP India Describing
Competency to be gained from this lecture Identify the different types of data to use appropriate methods to describe their distribution
Key issues Qualitative data Quantitative data Distribution
Data: A definition Set of related numbers Raw material for statistics Example: Temperature of a patient over time Date of onset of patients Data
Epidemiological process We want to describe a population We collect data We analyze data into information “Data reduction” We interpret the information We use the information for decision making Data
Types of data Qualitative data No magnitude / size Classified by counting the units that have the same attribute Types: Binary Nominal Ordinal Quantitative data Qualitative
Qualitative, binary data The variable can only take two values 1, 0 Yes, No Example: Sex Male, female Female sex Yes, No Qualitative
REC SEX M 2 M 3 M 4 F 5 M 6 F 7 F 8 M 9 M 10 M 11 F 12 M 13 M 14 M 15 F 16 F 17 F 18 M 19 M 20 M 21 F 22 M 23 M 24 F 25 M 26 M 27 M 28 F 29 M 30 M SexFrequencyProportion Female1033.3% Male2066.7% Total % Frequency distribution for a qualitative binary variable Qualitative
Using a pie chart to display qualitative binary variable Distribution of cases by sex Qualitative
Qualitative, nominal data The variable can take more than two values Any value The information fits into one of the categories The categories cannot be ranked Example: Nationality Language spoken Blood group Qualitative
REC NATION JORDAN 2 YEMEN 3 IRAN 4 JORDAN 5 YEMEN 6 JORDAN 7 YEMEN 8 TCHAD 9 SUDAN 10 IRAN 11 YEMEN 12 IRAN 13 JORDAN 14 SUDAN 15 IRAN 16 SUDAN 17 JORDAN 18 SUDAN 19 IRAN 20 YEMEN 21 SUDAN 22 YEMEN 23 SUDAN 24 IRAN 25 YEMEN 26 YEMEN 27 YEMEN 28 SUDAN 29 YEMEN 30 SUDAN Frequency distribution for a qualitative nominal variable CountryFrequencyProportion Yemen1136.7% Sudan826.7% Iran620.0% Jordan516.6% Total %
Using a horizontal bar chart to display qualitative nominal variable Distribution of cases by nationality Qualitative
Qualitative, ordinal data The variable can only take a number of value than can be ranked through some gradient Example: Severity Mild, moderate, severe Vaccination status Unvaccinated, partially vaccinated, fully vaccinated Qualitative
REC Status Clinical status: 1: Mild; 2 : Moderate; 3 : Severe Frequency distribution for a qualitative ordinal variable SeverityFrequencyProportion Mild1343.3% Moderate1136.7% Severe620.0% Total %
Using a vertical bar chart to display qualitative ordinal variable Distribution of cases by severity Qualitative
Key issues Qualitative data Quantitative data We are not simply counting We are also measuring Discrete Continuous Quantitative
Quantitative, discrete data Values are distinct and separated Normally, values have no decimals Example: Number of sexual partners Parity Number of persons who died from measles Quantitative
REC CHILDREN Frequency distribution for a quantitative, discrete data ChildrenFrequencyProportion % % % % % % % Total %
Distribution of households by number of children Using a histogram to display a discrete quantitative variable Quantitative
Quantitative, continuous data Continuous variable Can assume continuous uninterrupted range of values Values may have decimals Example: Weight Height Hb level What about temperature? Quantitative
REC WEIGHT WeightTally markFrequency 10-19III IIIII IIIII IIIII II III III I I I I1 Frequency distribution for a continuous quantitative variable: The tally mark
REC WEIGHT WeightFrequencyProportion % % % % % % % % % % % Total % Frequency distribution for a continuous quantitative variable, after aggregation
Using a histogram to display a frequency distribution for a continuous quantitative variable, after aggregation Distribution of cases by weight Quantitative
Series of 100 values of a quantiative variable Quantitative
ValuesFrequency Total100 Tabular and graphic representation of a distribution Distribution
Position Dispersion Describing a distribution
Summary Qualitative Binary NominalOrdinal SexNationalityStatus MYemenMild MJordanModerate FYemenSevere MJordanMild FSudanModerate FYemenMild MSudanModerate MIranSevere FJordanSevere M IranMild FYemenModerate FSudanModerate M IranMild MYemenSevere MJordanSevere FJordanModerate M IranMild FSudanMild MYemenMild Quantitative Discrete Continuous ChildrenWeight
Data type in computer software Type of dataType of variable in Epi-Info software Qualitative Binary Yes / No Nominal Integer (Code/numbers) Ordinal Integer (Code/numbers) Quantitative Discrete Integer Continuous Decimal Avoid free field variables difficult to analyze
Exercise Consider the class Describe the frequency distribution of the following variable: Sex State of origin Involvement in surveillance to date (None, partial, full time) Completed numbers of years in service Height in cm
Take home messages Qualitative data can be binary, nominal or ordinal Quantitative data can be discrete or continuous Distribution can be described with a table or a graph