Sampling Bias & Techniques

MDM4U1 Data Managament

Sample Bias

Sample does not represent the population

Response Bias

Respondents may not be honest to avoid embarrassment

Measurement Bias

The measurement tool/technique changes outcome

Non-Response Bias

Where the pinions of the responders & non responders will be very different

Leading Question

Asking another question that then influences the answer

Simple Random

Sample chosen by a random generator. Everyone has equal chance of being selected & selected are independent

Systemic Random

From a random starting point, sample is chosen sequentially & selected at random intervals

Stratified Random

Population is divided by common characteristics & proportional amounts of people are chosen from that group, representing the whole group.

Cluster Random

Population is divided into groups of all members of randomly selected groups respond.

Multi-Stage

Population is divided into different levels & a random sample is chosen at each level

Voluntary

Population is invited to participate voluntarily.

Convenience

Researcher selects most accessible participants.