Statistics: Data Types and Collection

• Dokuz Eylul University, Statistics Department provides statistics and probability material
• Prof.Dr. Esin FIRUZAN developed the related statistics engineering materials
• The subject matter includes data types, variable types, and measurement scales

From Data to Information

• Data is processed through statistical procedures to yield information

Scientific Data Collection

• Gathering data to reflect population characteristics requires scientific methods
• These methods involve probability, experimental design, and sampling theory
• Data collection is the process of gathering data that belongs to the selected sample with scientific methods to reflect the characteristics of populations.

Measurement Tools for Data Collection

• Statistical data collection methods include surveys, polls, interviews, Delphi techniques, and focus groups
• Data collection splits into Primary and Secondary collection methods
• Primary data collection methods include surveys, polls, interviews, Delphi techniques, and focus groups
• Secondary data collection splits into Financial Reports, Sales Reports, Government Reports, Mission Statements etc

Constructing Measurement Tools

• Steps include identifying, constructing, and testing the collection tool
• Validity and Reliability Tests are essential

Measuring Variables

• Measuring assigns numbers to phenomena based on specific rules
• Done to indicate the degree to which they have features of interests
• Measurement signifies the action of measuring and obtains information concerning variables

Variables and Data Defined

• A variable is a characteristic or property of an individual population unit that may vary
• 'Datum' is a singular item, person, or table
• 'Data' refers to a collection of items or a group of persons

Data and Variables

• Data=Variable which means it is a set of numbers representing measurements of a numerical characteristic of a group
• Terms describe nonnumerical characteristics associated with a group
• Data constitutes observations, like measurements, genders, and survey responses
• The experimental unit is the object on which the response and factors are observed or measured

Qualitative vs. Quantitative Data

• Variables divide into Qualitative or Quantitative
• Qualitative data cannot be measured on a natural numerical scale and are descriptive
• Quantitative data is measured numerically

###Qualitative Data

• Qualitative data are measurements that cant be measured on a natural numerical scale
• Qualitative data specifies data qualitatively meaning it is a specification of type only
• It classifies data into groups
• Variables like "Gender" can have answers like "Male/Female" and labels like "1/0"
• Colors (Answers: Red/Blue/Green, Labels: 1/2/3),
• Education (Answers: Primary/Middle/High, Labels: 1/2/3)
• Sickness (Answers: Sick/Not Sick, Labels 1/0).

Quantitative Data

• Quantitative data records measurements on a numerical scale
• Obtained through counting or measuring, objectively
• Expressed numerically
• This data offers specifics of quantity
• Quantitative data is convenient for mathmatics
• Examples include:
• Age (e.g., 15 years, 2 months, 3 weeks, 2 days, 5 minutes)
• Number of siblings (e.g., 3 brothers, 2 sisters)
• Amount of oxygen in the blood (e.g., 92, 93, 95)
• Heart Rate (e.g., 65, 67, 68, 55, 45)

Discrete vs Continuous Quantitative Data

• Quantitative splits into Discrete Variable or Continous Varibale
• Discrete variables have countable values in a finite amount of time and can only take integer values
• Number of patients over a day is a good example.
• Continuous variables can take any value corresponding to the points contained in one or more intervals.
• Examples are Elapsed time between 2 patients or examing product quality.

Measurement Scales

• Scales determine transactions based on variable values
• Measurement levels involve four groups, organized by measurement properties: nominal, ordinal, interval, and ratio scales

Nominal Scales

• Qualitative data is split into Nominal scales
• Nominal scales are a measurement classifying discrete categores
• These categories are represented as letters or numbers

Qualitative Data and Nominal Scales

• Dichotomous data: Yes/No, Male/Female
• More than two categories data: Color, Race, Wheat Type, fast food type, Blood Group, place of birth
• Others include, Postal Codes , TC Identity number and Occupations
• Nominal scales can calculate Frequency, Total Frequency, Relative Frequency

Ordinal Scales

• Data where the ordering of values makes sense

Qualitative data and Ordinal Scale

• Common with 5-point Likert Scale: Very Bad/Bad/Neutral/Good/Not Good
• First, second and third place in horse is an example
• Other examples are Education Level, and Socio-economic level as well as Patient’s recovery status, such as never healed/little healed/neither good nor bad/healed/fully healed
• Measurment types include Frequency, Total Frequency, and Relative Frequency

Interval Scale

• Have an order, the difference between options are equal, there is no absolute zero point.
• Can be used with Time, temperature, IQ score

Quantitative Data and Interval Scales

• Quantitative data consists of having some order
• The difference between options should be equal
• No absolute zero point
• Examples include time & temperature
• Calculations involve addition & subtraction
• Interval cannot compare the intervals

Ratio Scales

• A meaningful zero point is additional to nominal, ordinal and interval scales properties
• Has a meaningful absolute zero point in addition to nominal, ordinal and interval scales

Quantitative data and ratio scales

• Has a meaningful zero point and has all forms of mathmatical calculations
• They exist for weight, price, waiting time, height, rate, and income
• Measurments can involve weight or price

Data Type Summary

• Scales of measurment top to bottom
• Ratio Scales,
• Interval Scales,
• Ordinal Scales
• Nominal Scales
• Quantitative Types
• Ratio Scales,
• Interval Scales
• Ordinal Scales
• Qualitative Types
• Ordinal Scales
• Nominal Scales

Measurement Types

• Each has increasingly complex properties
• "Weight" is measurable, with a ratio scale (e.g., 70.8 kg)
• With an interval scale, 65kg-75kg would be a ratio scale measure
• With an ordinal scale it might be be Very heavy, heavy, not heavy
• With a nominal scale it is simply be Thin or Fat