Psyc232 Lecture 4: sampling and sampling biases

Sampling refers to the process of selecting individuals from a larger population to represent the entire group.
Survivorship bias is essential to consider, as it biases the conclusions by only focusing on subjects that survived a selection process.
Abraham Wald: A key figure in statistical reasoning who identified the flaws in armor placement on WWII planes based on bullet hole patterns.

Historical instance where planes returned with visible bullet holes (survivors) led to incorrect conclusions about armor placement.
Wald emphasized that armor should be placed where there were no bullet holes since those planes did not survive.
Demonstrated a critical understanding of survivorship bias.

Sampling is a significant part of research, as it affects the validity of the study.
Improper sampling can ruin an entire research study or lead to the saving of resources through effective design.
Expected you will spend 80% of your future work on sampling-related tasks.

Also known as random sampling, where each member of the population has an equal chance of being selected.
Example: Jury selection in New Zealand based on the electoral roll.
Defined as: "Everyone in a population is available and has an equal chance of being selected."

Population is divided into distinct subgroups, and random samples are taken from each.
Example: Electoral polling, where groups such as urban vs. rural voters are represented.
This reduces bias associated with demographic representation.

Researchers deliberately choose specific individuals based on characteristics pertinent to the study.
Example: A patient with retrograde amnesia chosen for a specific psychological study based on their condition.

Utilizes individuals who are easily accessible or involved, rather than randomly selected.
A less scientific approach but commonly used in practical studies due to ease.

Aim: To determine the perceived meaning of life across cultures.
The potential sampling methods discussed included probability, stratified, purposive, and convenience.
The conclusion was that convenience sampling may be the best method if the aim is to measure a broad societal phenomenon without geographical barriers.

The targeted sampling method was used to interview principals who successfully incorporated Mataranga Maori.
Researchers knew beforehand about the schools' practices related to Mataranga Maori, emphasizing purposive sampling.

Target population: Adults in New Zealand with an initial sampling method based on the electoral roll.
Identified as probability sampling since it randomly selects participants.

Survivorship Bias: Focuses on outcomes based on those who participated and can lead researchers to draw biased conclusions about the entire population.
Generalizability: How well can findings be applied to a broader context? This is often a limitation in convenience sampling.
W.E.I.R.D Phenomenon (Western, Educated, Industrialized, Rich, Democratic): Dominates psychological research sources, creating an unbalanced representation of global populations.

Emphasis on normal distribution: Understanding how the data appears can help identify biases.
Examined sexiness ratings from speed dates with a general bias showing people tend to rate others more generously than themselves.
Negative and Positive Skew: Classification of distributions based on extremities in data ratings.

Normal Distribution: Ideally, would work with over 30 participants in psychology research to capture broader trends.
Deviations from normal distributions can highlight various biases present in the sample, questioning the legitimacy of conclusions drawn.

Consider a balance in sampling strategies: Define the research goal against budget constraints.
Recognize that perfect sampling is impractical—successful studies can be achieved even with limited sample sizes.
Ultimately, reflection on participant representation is vital to ensure insights and conclusions reflect reality and account for broader perspectives.