WEEK 3 Flashcards
Structure of a quantitative paper
title + abstract
intro (outlines RQ, theory and summarizes research design)
LIT REV (usually written about DV)
Theory (usually written about IV)
HYP (introduced at end of thoery section as H1, H2)
Research Design (specifies variables, data sources, and model specifications)
Results (typically includse regression tables and/or plots (graphs)
Discussion / conclusion
purpose of descriptive stats
Descriptive statistics inform our choices (!!!!!!) in choosing variables and construct models, but aren’t quantitative analysis by themselves as there is no statistical math associated with them.
Regression tables
a regression is a tool for understanding a phenomenon of interest as a linear function of some other combination of predictor variables.
regression coefficient
the # NOT in parenthesis
provides the expected change in DV for a 1 unit increase of IV
Standard error
the # IN parenthesis
The standard error is our estimate of the standard deviation of the coefficient.
smaller SE = more reliant coefficient
the asteriks
indicate the level of the statistical significance of a regression coefficient.
t-statistic
coefficient divided by SE
Tells you how far the coefficient is from 0
P-value
probability that the effect is due to chance
r-squared
% of variance in DV explained by the model
closer to 1 = better fit
N (observation)
how many cases were used to estimate mdoel
more =. more reliable
coefficient plots
illustrates results of regression model, but less detail
Dot in the middle of each is the coefficient, same as in tables
Bars around each dot: confidence intervals = margin of error
- wider at values with fewer observations and slimmer with value with more observations
if margin of error doesn’t overlap with 0, the relationship is statistically sign.
NOT GOOD FOR EVALUATING INTERACTIVE EFFECTS, need marginal effects plot
marginal effects plots
The outcome on this graph is a predicted value of the dependent variable
line running through the plot is the correlation line – its calculated using the coefficient of a regression table
gray area around the line is the confidence interval which you can think of as a margin of error
what do tables show
Coefficients, standard errors, test statistics, and significance of each IV
Markers of significance and goodness of fit for entire model
The intercept (value of “y” when each “x” equals 0) and its errors and significance
histogram
Shows the distribution of independent variables, and how many observations are at each value
They can skew in one direction or another: We see a clear leftward skew, meaning that the bulk of the values are concentrated around a certain point – closer to 0 than to 100
purpose of descriptive stats
To build research design, we need to know the distribution of our variables
Can also use descriptive stats to identify cases that support or contradict hypotheses
“univariate hypothesis testing” is not used in modern social science
Descriptive stats alone are used in Research Design or in case studies
types of variables
Categorical – non-ordered categories: “religious denomination” or “political party”
Ordinal – ordered categories: “not at all, not so much, somewhat, very”
Continuous – can take on any number: GDP, market concentration, scope, etc.
distribution of data
How many instances of each value occur in your data?
Measured using standard deviation, visualized using histograms
The result shows us the sampling distribution
When data does not follow a normal distribution, we have ways of adjusting
Key descriptive stats
Mean – the average of the data – sum of values divided by number of observations
Median – the point in the middle of data, 50% of observations above and below
Minimum and maximum values – the lowest and highest values of a variable
Standard Deviation – how spread out the data is
standard deviation
For each value of a variable, we subtract the mean
We sum up the square of the differences – this is a common theme in statistics
Then we divide the result by observations minus one – degrees of freedom
