COMM 102 Lecture Notes - Lecture 5: Random Assignment, Millennials
12 Jun 2018
School
Department
Course
Professor
SCIENTIFIC METHODS IN MEDIA EFFECTS RESEARCH
● Experiment
● Surevy
● Content Analysis
I. EXPERIMENTAL METHOD
● Gathering data under controlled conditions
○ Violent TV→ Aggressive behavior?
● Randomization (random assignment)
○ Every participant has the SAME odds of being placed in the control group or
experimental group
■ Makes the groups in an experiment equivalent pre-experiment
● Manipulation of independent variable
○ Lets us see whether the changes in IV lead to corresponding changes in DV
PROS & CONS OF EXPERIMENTAL RESEARCH
● PROS→
○ Research has a lot of control
○ Can be used to infer causality***
■ Randomizing groups takes AWAY the concern of spurious results
■ Control allows us to show that CAUSE did in fact precede the EFFECT
● CONS→
○ Artificial settings
■ We WANT generalizability (aka ability to apply results to all of society), but that is
hard to do because experiments are performed in very controlled, unnatural
settings
○ Can only test short-term effects
II. SURVEY
● Investigating people by asking questions directly
○ Standardized questionnaire
● Contact methods of survey
○ Face-to-face interviewing
○ Telephone, mail, email, web surveys
○ Fax & text-message polls
● Selection of respondents
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
Every participant has the same odds of being placed in the control group or experimental group. Makes the groups in an experiment equivalent pre-experiment. Lets us see whether the changes in iv lead to corresponding changes in dv. Randomizing groups takes away the concern of spurious results. Control allows us to show that cause did in fact precede the effect. We want generalizability (aka ability to apply results to all of society), but that is hard to do because experiments are performed in very controlled, unnatural settings. Population whole population of people we want info from. Sample the smaller group we actually use for the info. Everyone has an equal chance of being selected into the sample. Surveying a single sample of some population at one time. Surveying at multiple times: trend survey, cohort survey, panel survey, trend survey. Survey different groups of individuals at multiple times. Interviewing high school students in 1980, 1990, 2000.
