Lec 7:Data Collection Methods

Random Sampling

Subjects are selected by random numbers.

Systematic

Subjects are selected using every kth number after the first subject is randomly selected from 1 through k

Stratified

Subjects are selected by dividing up the population into groups (strata, and subjects are randomly selected within groups.

Cluster

Subjects are selected using an intact group that is representative of the population

Data Collection

Systematic approach to gather relevant information from various sources

Primary Data Collection

• Collect/generate new data

• Produce raw data

• Takes time

Secondary Data Collection

• Facts from published sources (online/offline)

• Compiled from primary sources

• Takes less time

Observational Study

• Researcher observes and measures characteristics of interest

• Done in a very short time or something that takes place over long period of time

• No manipulation, just observe in natural environment

Retrospective Observational Study

• Observes data that happened in the past

Prospective Observational Study

• Starts right now and studies group for set period of time

Experiments

• Study examine effect on variable of interest by manipulating levels of another variable, while keeping all other things equal

• Typical experiment contains both treatment and control groups

Simulation

• Mathematical or physical model to recreate real-world conditions

• Done with computers

• Saves time and money

• Synthetic data

Surveys

• Uses questionnaires to gather values for responses

• F2F, online, phone, email

• Potential for bias

• Best when respondents feel anonymous

Secondary Data Collection Methods

• (Processed) data available in print or in electronic form

Variable

Characteristic (of interest) or phenomenon that can be measured or classified in research study

Nominal

Classifies data into mutually exclusive (i.e., non-overlapping) categories w/o order or ranking.

• Nominal Scale

Ordinal

Classifies data that can be ranked. Precise differences between ranks do not exist

• Ordinal Scale

Continuous

Usually result of measurement. Can take any value (analogous to real numbers)

• Ratio Scale

• Interval Scale

Discrete

• Usually result of counting something

• Distinct/certain values only (analogous to unsigned integer numbers)

  • Ratio Scale

Data

• Observed or measured values representing attributes/characteristics of variables

• Gathered facts