COM CM 321 Lecture Notes - Lecture 8: Sampling Error, Sampling Frame, Observational Error
26 Oct 2019
School
Course
Professor
CM321
➔ Oct. 23rd, 2019
Error
- Error is anything and everything that interferes with our true understanding of reality
- Validity = the degree to which measurements/collected data and reality correspond
- In measurement and study design, one of the biggest problems is to identify and control
error
-
Threats to validity
- Measurement error: reliability + validity of measures
o Threatens internal validity
▪ Random error: unknown, hard to control, and everywhere
▪ Systematic error: what we should focus on avoiding
▪ Can happen in the same study
- Sampling error: generalizability and selection bias
o Threatens external validity
▪ How much conclusions of a study can be generalized
Population vs. sample
- Sampling
o Goal is to:
▪ Identify a population
▪ Survey a small selection of that population
▪ Draw accurate conclusions about the entire population based on the small
selection
o Questions to consider:
▪ Can we generalize?
▪ Is the sample representative?
o Steps:
1. Identify the population
• Population: unit of analysis (people, nations, headlines, tweets…)
2. Identify a sampling frame
• How do you identify a complete list of all possible units that could
be included?
3. Selecting sample from frame
• Means of choosing who will be participant/subject or what will be
analyzed
- Population
o A group of people or subjects under investigation
o Census: process of examining every member in a population
o Sample: subset of a population
o Steps
1. Identify the population
• Depends on research question/hypotheses
• Broader scope is often better
