COMM 101 Lecture Notes - Lecture 17: Web Scraping, Internal Validity, Dependent And Independent Variables

• School: U of M
• Department: Communication Studies
• Course: COMM 101
• Professor: Julia Lippman

Document Summary

What kind of diet is healthiest: three types of explanations: Something else causes people in an area to be happy: this is spuriousness, epidemiological studies are great at ruling out possible explanations that are wrong, but not as good in establishing causality. Take away #1: epidemiological studies reveal societal patterns. I. e. count all the times anyone makes a reference to the following words near the word. Job (good, like, happy, enjoy) ((bad, unhappy, dislike, hate)) See how these two groups relate to get a measure of job satisfaction. You de ne a series of things for the computer to search for. How many times those things are found provides your measure. Strengths: quicker than hand coding, generally inexpensive, very reliable. Weaknesses: can only code manifest content, cannot code nuance, irony, tone, etc. , validity varies: data-driven coding: researcher tells the computer what the dependent variable is, computer searches for what things might predict that dependent variable.