PSYCH 209 Lecture Notes - Lecture 11: Nonprobability Sampling, Construct Validity, Random Assignment
27 Apr 2018
School
Department
Course
Professor
4/20/18 Lecture Notes: Midterm Review
General Exam Information
• Everything we have covered in class so far (lecture & book)
o Chapters 1,2,3,5,6,7
• 30 multiple-choice questions (30 pts)
o Similar to the in-class PollEv questions and practice quiz questions from the study guide
o Not similar to the InQuizitive questions
• 3 short-answer questions (10 pts)
• Bring a scantron & pencil
Sampling
How to get a Non-Representative Sample
• All of these methods use nonprobability or nonrandom sampling
• Convenience Sampling
o Who's easy to reach
• Quota Sampling
o Let's get different types of people
• Purposive Sampling
o Let's get one type of people
• Snowball Sampling
o Recruit from one social network
• Sampling vs. Assigned to Groups
o Sampling is the group of people you are studying
o Assigning to Groups is when you assign people in your sample to different groups that you
want to test
How to get a Representative Sample
• Probability Sampling
o A class of sampling procedures that all use randomization processes for picking the
members of the sample
Validity
• Construct Validity
o Quality of the measures and manipulations
• Internal Validity
o No alternative causal explanations for the outcome
• Statistical Validity
o Statistical Conclusions are
Poll Everywhere
• Suppose you know that most middle school students who are bullied have low self-esteem.
However, you know that this may not be causal, because you are not sure which happened first:
being bullied or having low self-esteem. Which of the following criteria for causation are you
questioning?
o temporal precedence
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
How to get a non-representative sample: all of these methods use nonprobability or nonrandom sampling, convenience sampling, who"s easy to reach, quota sampling, let"s get different types of people, purposive sampling, let"s get one type of people. Snowball sampling: recruit from one social network. Assigned to groups: sampling is the group of people you are studying, assigning to groups is when you assign people in your sample to different groups that you want to test. How to get a representative sample: probability sampling, a class of sampling procedures that all use randomization processes for picking the members of the sample. Validity: construct validity, quality of the measures and manipulations. Internal validity: no alternative causal explanations for the outcome. Suppose you know that most middle school students who are bullied have low self-esteem. However, you know that this may not be causal, because you are not sure which happened first: being bullied or having low self-esteem.
