Help
Options
Didn't know it?
click below
click below
Knew it?
click below
click below
Don't Know
Remaining cards (0)
Know
retry
shuffle
restart
0:00
Embed Code - If you would like this activity on your web page, copy the script below and paste it into your web page.
Normal Size Small Size show me how
Normal Size Small Size show me how
econ 570 midterm 2
| Term | Definition |
|---|---|
| What does 'import statsmodels.api as sm' do? | It imports the main statsmodels library which includes models like OLS and ANOVA. |
| What does 'import statsmodels.formula.api as smf' do? | It allows you to create statistical models using formulas like 'y ~ x1 + x2'. |
| What is the syntax to specify a model using smf? | smf.ols('y ~ x1 + x2', data=df) |
| What type of model does smf.ols() create? | An ordinary least squares (OLS) linear regression model. |
| How do you estimate the coefficients of a model? | Use .fit() after defining the model, e.g., results = model.fit() |
| How do you display regression results in full detail? | Use print(results.summary()) |
| How do you view only the estimated coefficients? | results.params |
| How do you get confidence intervals for coefficients? | results.conf_int() |
| What does it mean if the confidence interval includes 0? | The coefficient is not statistically significant. |
| How do you get the R-squared value from a regression? | results.rsquared |
| What is the formula syntax for creating a regression model? | DependentVariable ~ IndependentVariable(s) |
| Example of syntax for regression with multiple controls | sleep_model = smf.ols('sleep ~ twork + educ + age', data=sleep) |
| What function allows you to estimate model parameters? | .fit() |
| Example of fitting and printing a model | results = sleep_model.fit(); print(results.summary()) |
| What does np.log() do in regression modeling? | It transforms variables to measure elasticity or proportional changes. |
| Example of log transformation in model | smf.ols('sleep ~ twork + educ + np.log(age)', data=sleep) |
| Why use np.log(age) instead of age? | Because age effects usually diminish over time and log makes the relationship more linear. |
| How do you include interaction effects between variables? | Use the * operator, e.g., smf.ols('sleep ~ totwrk + educ*age', data=sleep) |
| What does a negative educ:age coefficient indicate? | As age increases, the effect of education on sleep decreases. |
| How do you include a categorical fixed effect? | Add C(variable), e.g., smf.ols('sleep ~ totwrk + educ + age + C(gender)', data=sleep) |
| How do you detect heteroskedasticity? | Check if residuals are unevenly distributed. |
| How do you correct for heteroskedasticity? | Use fit(cov_type='HC3') |
| Example of heteroskedasticity correction | prac_results_h3 = smf.ols('cigs ~ np.log(income)', data=smoke).fit(cov_type='HC3') |
| What function is used to make scatter plots? | ax.scatter(x, y, marker='o', alpha=0.5) |
| What does sns.despine(ax=ax) do? | Removes the top and right spines from the plot for cleaner visuals. |
| What is the correct syntax for a scatter plot of log(income) vs residuals? | ax.scatter(np.log(smoke['income']), prac_results.resid, marker='o', alpha=0.5) |
| What is wrong with this syntax: ax.scatter(np.log(smoke[income]), prac_results.resid)? | The variable 'income' must be quoted: smoke['income'] not smoke[income]. |
| How do you sort data before plotting a regression line? | sleep.sort_values('twork', inplace=True) |
| How do you plot a regression line manually? | Use ax.plot(x, y, color='black') after creating x and y values. |
| Example of creating regression line points | x = [sleep['twork'].min(), sleep['twork'].max()]; y = [p.Intercept + p.twork*i for i in x] |
| What does results.get_prediction() return? | Predicted values and statistics including confidence intervals. |
| How do you extract confidence intervals from prediction results? | summary_frame = predictions.summary_frame(alpha=0.05) |
| How do you plot confidence intervals manually? | ax.fill_between(x, ci_lower, ci_upper, color='blue', alpha=0.2, edgecolor='none') |
| Why not use seaborn for regression confidence intervals? | Seaborn doesn’t show model equations or specific statistical results. |
| What import is used for instrumental variable regression? | from linearmodels.iv.model import IV2SLS |
| What is the purpose of regression discontinuity? | To measure the difference in outcomes around a cutoff or discontinuity. |
| What does Difference in Differences (DiD) estimate? | Causal effects between two groups and two periods. |
| What is the Parallel Trends Assumption? | The treated and control groups would have moved similarly over time if no treatment occurred. |
| How to tell if parallel trends hold? | Pre-treatment coefficients are close to 0 and error bars cross 0. |
| What does it mean if error bars cross 0? | The difference is not statistically significant. |
| What library is used for time series analysis? | import statsmodels.tsa.api as tsa |
| What does tsa.adfuller() test for? | Stationarity of a time series. |
| What is the null hypothesis of the Dickey-Fuller test? | That the series is non-stationary. |
| How do you interpret a p-value < 0.05 in Dickey-Fuller? | Reject the null hypothesis → the series is stationary. |
| What function estimates autoregressive models? | tsa.AutoReg() |
| Example of fitting an AR(2) model | ri_unemp = tsa.AutoReg(ri_data, lags=2).fit() |
| How do you forecast future values in a time series? | Use .forecast(steps) |
| Example of forecasting eight future periods | predictions = ri_unemp.forecast(8) |
| How do you plot observed vs forecasted data? | Plot lines with different colors for observed (blue) and forecasted (red). |
| What does Difference in Differences require? | At least two groups (treated/control) and two time periods (before/after). |
| What are event studies used for? | To explore pre- and post-treatment trends over time. |
| What function creates dummy variables? | pd.get_dummies() |
| Why drop a reference category after get_dummies()? | To avoid multicollinearity. |
| Example of dropping a reference column | sw06.drop(columns='Di_mi', inplace=True) |
| What function merges coefficient estimates and confidence intervals? | pd.concat([results.conf_int(), results.params], axis=1) |
| How do you plot event study coefficients with error bars? | ax.errorbar(x, y, yerr, color='red', capsize=2) |
| What function draws shaded confidence intervals? | ax.fill_between(x, lower_ci, upper_ci, color='blue', alpha=0.1) |
| What is wrong with this syntax: ax.fill_between(sleep['twork'], ci_lower, ci_upper, color=”CO”, alpha=0.2)? | Invalid color code; use a valid color name like 'blue' or 'gray'. |
| What function creates horizontal and vertical reference lines? | ax.axhline() and ax.axvline() |
| What is wrong with this syntax: ax.axvline(0, linestyle=’dashed’, color’=blue’)? | Missing '=' after color → should be color='blue'. |
| What does sns.despine(ax=ax) achieve? | Removes unnecessary chart borders for cleaner visuals. |
| What does .set_xlabel() and .set_ylabel() do? | Label the x-axis and y-axis respectively. |
| What is the syntax to label axes? | ax.set_xlabel('label', fontsize=14); ax.set_ylabel('label', fontsize=14) |
| What does the parameter alpha control in scatter plots? | The transparency of points. |
| What happens if you forget to quote column names in a formula? | It causes a NameError because the variable is undefined. |
| What function combines two dataframes vertically? | pd.concat([df1, df2]) |
| What does .summary_frame(alpha=0.05) return? | A DataFrame containing mean predictions, confidence intervals, and standard errors. |
| What is the purpose of cov_type='HC3'? | To correct standard errors for heteroskedasticity. |
| What function removes columns from a DataFrame? | .drop(columns='column_name', inplace=True) |
| What function returns unique values from an index? | .unique() |
| How do you create an annual frequency DateTimeIndex? | pd.DatetimeIndex(df.index.values, freq=df.index.inferred_freq, name='date') |
| What is the dependent variable in sleep_model = smf.ols('sleep ~ twork + educ + age', data=sleep)? | sleep |
| What is the correct syntax for adding interaction terms? | variable1 * variable2 |
| What does smf.ols stand for? | Ordinary Least Squares regression using formula API. |
| What is the default alpha level for 95 | |
| confidence intervals? | 0.05 |
| What function gives fitted residuals from a model? | results.resid |
| What does results.params['x1'] represent? | The estimated coefficient for variable x1. |
| How can you interpret a negative coefficient? | A one-unit increase in x decreases y by the coefficient value. |
| What causes multicollinearity? | Including two or more highly correlated variables (like both educ and educ*age). |
| What library provides seaborn plotting functions? | import seaborn as sns |
| How do you interpret the intercept in an OLS model? | Expected value of y when all x variables are 0. |
| What is the difference between sm and smf? | sm provides core statistical models; smf allows model creation using formulas. |
| What function gives fitted values from regression? | results.fittedvalues |
| What does OLS stand for? | Ordinary Least Squares. |
| What does the ‘~’ operator mean in a formula? | It separates dependent and independent variables. |
Created by:
tmanderson5