×
Log in to StudySoup
Get Full Access to MSU - PLS 201 - Study Guide - Midterm
Join StudySoup for FREE
Get Full Access to MSU - PLS 201 - Study Guide - Midterm

Already have an account? Login here
×
Reset your password

MSU / Philosophy / POLS 201 / What is an empirical research design?

What is an empirical research design?

What is an empirical research design?

Description

School: Michigan State University
Department: Philosophy
Course: Introduction to Methods of Political Analysis
Term: Spring 2018
Tags: Reliability, Internal Validity, measurements, and nominal
Cost: 50
Name: PLS 201 midterm study guide
Description: This study guide covers material from the lectures, but does not include poll everywhere questions.
Uploaded: 02/24/2018
8 Pages 62 Views 5 Unlocks
Reviews


PLS 201 Midterm 


What is an empirical research design?



Reading notes: Kida’s introduction

➔ Kida’s 6-pack of mistakes:

◆ We prefer stories to stats.

◆ We seek to confirm.

◆ We rarely appreciate the role of chance and coincidence in life. ◆ We can misperceive the world.

◆ We oversimplify.

◆ We have faulty memories that are biased.

➔ What we believe affects what we decide.

➔ To make better decisions, we should strive to be better thinkers. ➔ Relying on anecdotal evidence to form our beliefs and decisions is risky because it means we ignore other relevant evidence like stats which often provides us with the most reliable info.

➔ Memory is constructive

◆ current beliefs, expectations, environment can influence our memory of past events.


What is a philosophy as a science?



Factors that influence our beliefs & decisions:

➔ Gut reaction

➔ Society and culture

➔ Family If you want to learn more check out What is the start of puberty dependent on?

➔ Scientific news or research

Confirmation bias: our tendency to seek out, interpret or favor info that confirms our current beliefs and views about the world

Theory: a well-substantiated but tentative explanation about the causes of some phenomenon of interest

Data: meaningful observations that help us support or refute our theory

Pseudoscience vs. science

➔ Pseudoscience: claims that are presented as science but lack sufficient supporting evidence or plausibility

➔ What makes something pseudoscience:


What is pseudoscience?



◆ Preconceived notion of what to believe

◆ Search for evidence to support it

◆ Ignore evidence that would falsify it

◆ Disregard alternative explanations the phenomenon

◆ Lack of tightly controlled experiments

◆ Employ little skepticism

➔ We often believe in pseudoscience because we find it intriguing and comforting, it resembles real science, and we see it everywhere. ➔ Science seeks to falsify, not to verify.

Inductive vs. deductive reasoning

➔ Deductive: thinking that starts out with a general statement or hypothesis and examines the possibilities to reach a specific, logical conclusion. Don't forget about the age old question of What causes microcytic anemia?

◆ Based on theory

◆ Helps us to explicitly rule out what is not the cause

◆ Ex: “All men are mortal. Harry is a man. Therefore Harry is mortal.” ➔ Inductive: makes broad generalizations from specific observations

◆ Can tell us what might be the cause

◆ Based on observations

◆ Ex: “The coin I pulled from my bag is a penny. A third coin from the bag is a penny. Therefore, all coins in the bag are pennies.”

Philosophy of science

➔ Hume: a major problem w science is that:

◆ It is performed by imperfect people

◆ Scientists are not often open to criticism of their work.

◆ Scientists dismiss evidence that does not support their theory. ◆ It presupposes a uniformity of nature.

➔ Karl Popper: believed that falsifying a theory is a deductive process. That we need deductive standards.

◆ Once you falsify a theory, you should throw it out and start over. Only keep the ones that are not falsified by evidence.

➔ Feyerabend: thought Popper was wrong- that deductive rules and rigor limit science, and that science doesn’t always behave that way.

➔ Getting rid of ideas that have been falsified or those that cannot be falsified robs us of innovative ideas.

➔ Kuhn: believed we need standards but that we didn’t need to start over when we made a mistake, just to update and test again. We also discuss several other topics like What is totalitarianism?

◆ middle ground bw Feyerabend and Popper.

◆ Believed that science “requires a decision process which permits rational ppl to disagree”

➔ Kuhn’s 5 scientific standards:

◆ Accuracy

◆ Consistency

◆ Scope (generalizability)

◆ Simplicity

◆ Fruitfulness (creation of new knowledge)

➔ Key components of science:

◆ Collaborative and public (replicative)

◆ Knowledge is cumulative

◆ Falsification, not confirmation

◆ Preference for simplicity

◆ Ideas should be testable and specific, not vague

◆ Generalizable

◆ Employs both inductive and deductive reasoning

➔ Paradigm: a set of fundamental assumptions scientists work under

Scientific research process

➔ (Alternative) hypothesis: testable statement about the explicit relationship we expect to see between variables We also discuss several other topics like What happened to beringia?

➔ Null hypothesis: what we don’t expect between our variables, what we expect to see if our theory was wrong

➔ Variable: a measurement of a characteristic that is not constant ◆ Independent (IV), dependent (DV)

Statistics

➔ Help us summarize complex amounts of info and data

➔ Descriptive stats: consist of numerical and graphical methods for summarizing info. Uses 2 tools:

◆ Measures of central tendency: a numerical value that describes a typical observation in your data. I.e. mean, median, mode. gpa.

◆ Measures of dispersion: the amount of variance or uncertainty around the central tendency. I.e. standard deviation, variance

● Variance is sensitive to outliers.

● Standard dev will always be greater than or equal to zero. Represented in a bell-shaped curve w 68%, 95%

● Understanding a variable’s dispersion is important because: ○ it is the most interesting and distinctive feature

○ not all variables have the same distribution

○ Measures of central tendency may parallel another

variable’s, but the distributions may differ

○ It may help determine the best method to use in data

analysis

➔ Inferential stats: consist of methods for drawing conclusions about a population based on info from a sample Don't forget about the age old question of What is a theory?

➔ Correlation: a simple, descriptive stat that quantifies the relationship bw 2 variables and measures a linear relationship

➔ Descriptive inferences: statements that describe something we cannot see based on what we can see

◆ Ex: You students are all young and healthy. Therefore I infer that MSU students are mostly young and healthy.

Measurement

➔ Measurement: the use of numbers to represent theoretical concepts ➔ 2 standards are needed for good measurement: reliability, validity ◆ Reliability: extent to which a measure consistently measures a concept ● Gets rid of random errors

◆ Validity: extent to which a measure records the true value of the intended characteristic and nothing more

● Gets rid of systematic errors

● Face validity: the measure makes sense

● Construct validity: it corresponds with past measurements

◆ Ex: “politically independent” is a reliable variable but not valid because it represents a lot of different groups.

➔ Operationalization: turning concepts into measures

➔ Levels of measurements

◆ Nominal level: variables that show the difference between values and have no order

● Ex: democrat, republican

◆ Ordinal level: variables that show the diff between values and have an inherent order Don't forget about the age old question of What types of religious specialists have anthropologists classified?

● Ex: scale of strongly disagree to strongly agree

◆ Interval level: variables that show the exact numerical differences between values and have an order

● Ex: years

◆ Ratio level: variables that show the exact numerical diff between values, have an order and absolute zero is a meaningful value

● Ex: thermometer- 0 degrees K

➔ Spuriousness: when 2 or more variables are not causally related to one another but are wrongly inferred that they are, usually due to a 3rd omitted variable

◆ Ex: claim- those with big shoe size can read really well! Age is omitted variable.

➔ Absolute numbers used when someone wants to make a statistic seem bigger than it truly is

➔ Relative numbers used when someone want to make a stat seem smaller ➔ Selection bias: selection of data that causes the sample to be unrepresentative of the population

Empirical research designs

➔ Include experiments, field observations, surveys, case studies, quasi-experiments ➔ Reveal whether one phenomenon chronologically precedes the other ➔ Experiments: when the researcher controls the key independent variable and randomly assigns values to participants

◆ Considered to be the “gold standard” design bc of randomization ● Randomization addresses issues of omitted variables and

spuriousness, and rules out alternative explanations whether you

think of them or not

◆ Has high internal validity (ability to rule out confusing variables and isolate the cause), but low external validity (results can’t be generalized) ➔ Observational studies: experiments that capture data that already exists ◆ Types of observational designs

● Cross-sectional: has multiple units, focuses on one pt in time ● Time series: has one unit w/ multiple points in time

● Time series, cross-sectional: multiple units with multiple points in time

◆ Has high external validity, but low internal validity

Ethical issues in science

➔ Institutional Review Board (IRB): committee that reviews and approves research involving human subjects

◆ Purpose: to ensure that all human subject research is conducted in accordance with all federal, institutional and ethical guidelines

Page Expired
5off
It looks like your free minutes have expired! Lucky for you we have all the content you need, just sign up here