Chapter 2 Flashcards
Goals of Reading and Evaluating Scientific Research
-Emphasis on critical thinking
-Objectivity
Quality Scientific Research
-Based on objective, valid and reliable measurements
-Is generalizable
-Uses techniques to reduce bias
-Is public
-Can be replicated
Objective Measurements
Consistent measurements across instruments and observers
Validity
Degree to which an instrument or procedure measures intended measurements
Reliability
Consistency in measurements across multiple observations or points in time
Tips for Generalizability
-Use large sample size
-Random vs convenient
-Natural environment (Want to be as natural as possible)
Examples of Bias
-Researcher Bias
-Subject / Participant Bias
-Hawthorne Effect (change in behaviour due to being observed)
Reducing Bias
-Anonymity, confidentiality, placebo
-Using ‘blind’ procedures
Single Blind Study
Participants don’t know purpose or treatment/condition they have
Double Blind Study
Both participants and experimenters don’t know exact treatment/ conditions
Publicity
Publishing in academic journals
Replication Crisis
-Older studies are hard to replicate and/or give differing results
-publication bias (pressure for exciting results, reduced motivation for replication of research)
Poor Scientific Research Traits
- Untestable hypothesis (unfalsifiable)
- Anecdotal/reliant on subjective experience
- Biased selection on data (selective data reporting)
- Reliance on authority, not scientific evidence
- Reliance on common sense, not scientific evidence
Descriptive Research
-Qualitative Research: Examines issue/behaviour without numerical measurement
-Quantitative Research: Examines issue/behaviour with numerical measurement/statistics
Case Studies
An in depth report about the details of a specific case
Naturalistic Observation
Unobtrusively observing and examining behaviour in its natural environment
Surveys and Questionnaires
-face to face interviews
-phone surveys
-paper and pencil tests
-web based questionnaire
-allows assessment of attitudes, opinions, beliefs and abilities
-Note: participant bias, validity (questions giving answers to right things), questionable truthfulness
Correlation Research
-Measuring degree of association between 2+ variables
-positive (🔼🔼 🔽🔽)
-negative (🔼🔽)
-magnitude ranges from -1(negative correlation) to 1(positive correlation)
-0 = no correlation
Problems of Correlation Research
-Third variable (Something else could be the cause)
-Spurring correlations (coincidental correlation)
Experimental Research Design
-Random assignment for test groups (sample divided into groups with equal likelihood for placement in each experimental condition)
-increased experimental control
Types of Experimental Research
-Between-Subjects design
-Within-Subjects design
Between-Subjects Design
Compares performance/score between different groups
-Treatment group vs non-treatment group at the same time
Within-Subjects Design
The same participants experience all treatments
-Treatment 1 vs Treatment 2 at different times
Quasi-Experimental Research Designs
-not random assignment
-groups being compared are selected based on pre-determined characteristics
Strengths and Limits of Each Research Design
Case Studies-Detailed, but not generalizable (single subject focus)
Naturalistic Observation- Detailed description of subjects in natural environment, but poor control over influential variables
Surveys/Questionnaires- Quick, convenient, large sample size, poor control and uncertain honesty
Correlation Studies- Shows strength of relationship between variables, but no cause and effect
Experiment- Tests for cause and effect, good control, but risk of not being general to real-world situations
Research Ethics Board (REB)
Committee of researchers and officials at an institution responsible for protection of human research participants
-risks must be outweighed by benefits
-no great harm for participants, or prevention for harm
Examples of Bad Conduct
-Little Albert
-Milgram’s Obedience Studies
-Stanford Prison Experiment
Ethics
-Informed consent
-Debriefing must occur after study, especially with deception
-Anonymity and confidentiality
-Collection, storage, reporting of data
-Animal welfare (check with animal ethics board)
Hypothesis Test
A statistical method of evaluating whether differences are meaningful
2 Possible Outcomes to Experiments
-null hypothesis: no significant difference
-experimental/alternative hypothesis: significant difference
Statistical Significance
If probability (p) of results being chance is 5% or greater (p > 0.05), results are not significant
Otherwise, results are significant
