Quiz 2 Empirical Political Anaylsis

Controlled Comparison

A comparison between groups or conditions made with the value of a potentially confounding variable held constant.

Compositional Differences

Differences in the groups being compared, which can affect research outcomes and distort estimated effect of a treatment (or another independent variable).

Spurious Relationships

The IV → DV relationship does not exist after controlling for a rival explanation.

Additive Relationships

The same IV → DV relationship is observed within each category of the control variable.

Interactions

Zero-order relationship

The initial relationship observed between variables without controlling for the effects of other variables; the raw association between variables.

Partial Relationship/Partial Effect

When we are doing a controlled comparison, we are only looking

at a subset of our data

Controlled Effect

The effect or contribution of an independent variable on the dependent variable within one constant value of a control variable.

Mean Comparison Table

Table that reports how the mean values of a DV change across categories of an IV; can also show standard deviations and counts.

Graphing Patterns for Relationships

Inferential Statistics

Procedures used to assess how well a descriptive sample statistic reflects a population parameter and to test hypotheses.

Population

Entire possible group

Sample

Selected from the population but does not include all members

Census

Data on the entire population

Population Parameter

Some characteristic or property of the population we are interested in

Sample Statistic

Estimate of the population parameter that you get from looking at a sample

Random sampling error

Possibility that through randomness you inadvertently get a really weird sample

Population vs. Sample Symbols

Central Limit Theorem

Statistical principle that the distribution of the sample mean will approximate a normal distribution, regardless of how variable values are distributed in the population from which samples are drawn, for sufficiently large samples.

Cumulative Density

The proportion (or percentage) of a probability distribution that’s at or below a given value in its range of possible values. A bell-curve distribution’s cumulative density function is an S-shaped curve.

Three things that influence how well sample statistics match the population parameter

1. Sample quality

2. Sample size

3. Variance

Standard Error

How much a sample statistic will vary, on average, when it is repeatedly estimated from random samples.

Confidence Interval

Measure of confidence for their sample statistic matching the population parameter - 95% is most commonly applied

Margin of error

Interval of reasonable uncertainty about a population parameter; the distance between sample statistic and upper or lower bound of confidence interval.

Test of statistical significance

A procedure to determine whether a hypothesis about a population parameter should be rejected based on sample data.

Five Steps to hypothesis testing

• Propose a research hypothesis (which implies a null hypothesis).

• Set the significance level (usually .05).

• Estimate relevant population parameters using sample data.

• Calculate the confidence interval or P-value.

• Reach a conclusion about the null hypothesis.

Directional hypotheses

When the researcher hypothesizes that the IV has a positive or negative effect on the DV. Use one-tailed P-values with directional hypotheses.

Non-directional Hypotheses

Asserts a difference or effect but not its direction, evaluated with two-tail P-values.

Type 1 Errors

The error of rejecting a true null hypothesis, incorrectly concluding that there is a significant effect or relationship when one does not exist.

Type 2 Errors

The error of failing to reject a false null hypothesis, incorrectly concluding that there is no significant effect or relationship when one actually exists. Better to make this error than type 1

𝛂

Alpha - predetermined confidence level you require to reject the null hypothesis p < a you can reject the null hypothesis. If the CI is 95%, a is 0.05 which is a 5% chance of a type 1 error.

P-value

Probability of finding that result assuming the null is true. Ranges from 0-1. Ranges from least likely to most likely when p < a you can reject the null hypothesis

Confidence Interval

Range around mean/proportion where the population parameter can lie

Descriptive Hypothesis

Describes the proportion doesn’t speak to cause

Causal Hypothesis

Measures causal relationships

Error bar chart

Graph with line segments that represent uncertainty of estimates, typically indicating the range of values within which the true value is likely to fall.