conveRt to R: the short course

# conveRt to R: the short course
### Chris Hanretty
### January 2020

---

# Unit 3: basic regression modelling

---

# Reminder

```r
library("tidyverse")
```

otherwise things will not work
]

---

# Reading in the data (reminder)

```r
dat <- read.csv("vdem_short.csv")
```

If this doesn't work, you can try the following command, which will allow you to select the file using your operating system.

```r
filepath <- file.choose()
dat <- read.csv(filepath)
```

---

# Description of the data

- **country_name**: the name of the country
 - **country_text_id**: the ISO 3166 3 letter code for the country
 - **year**: the year in which the observation was made
 - **v2x_polyarchy**: a continuous measure of electoral democracy on a
   0-1 scale from the V-Dem project
 - **e_boix_regime**: a dichotomous measure of democracy from Carles
   Boix.
 - **e_migdppc**: GDP per capita in US dollars
 - **e_regionpol_6C**: region the country is located in.

---

# OLS regression

To run a regression, we need to get to know a new function (`lm`) and the syntax for formulae.

It's best to demonstrate with an example.

```r
mod <- lm(v2x_polyarchy ~ e_migdppc + e_regionpol_6C, 
          data = dat)
```

Here, we've run an OLS regression predicting levels of democracy using GDP per capita and the region each country is located in.

R knows to find our variables in the data frame `dat`.

We store the results in the object `mod`.

---

# OLS output

To see the results of our groundbreaking regression, we use `summary`.

```r
summary(mod)
```

```
## 
## Call:
## lm(formula = v2x_polyarchy ~ e_migdppc + e_regionpol_6C, data = dat)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -0.48511 -0.10073 -0.00284  0.11781  0.45167 
## 
## Coefficients:
##                                                  Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                                     4.132e-01  4.525e-02   9.131 3.48e-16 ***
## e_migdppc                                       1.246e-06  1.143e-06   1.090  0.27757    
## e_regionpol_6CEastern Europe and Central Asia   1.418e-01  5.533e-02   2.563  0.01132 *  
## e_regionpol_6CLatin America and the Caribbean   2.405e-01  5.842e-02   4.118 6.20e-05 ***
## e_regionpol_6CMiddle East and Northern Africa  -1.764e-01  6.217e-02  -2.838  0.00515 ** 
## e_regionpol_6CSub-Saharan Africa                3.958e-02  5.226e-02   0.757  0.45001    
## e_regionpol_6CWestern Europe and North America  4.242e-01  6.488e-02   6.538 8.53e-10 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.1931 on 155 degrees of freedom
##   (16 observations deleted due to missingness)
## Multiple R-squared:  0.4802,	Adjusted R-squared:   0.46 
## F-statistic: 23.86 on 6 and 155 DF,  p-value: < 2.2e-16
```
]

---

# Recoding for legibility

In the last unit we talked about recoding factors.

Let's recode for legibility now.

```r
dat <- dat %>% 
  mutate(e_migdppc = e_migdppc / 1000,
         Region = recode(e_regionpol_6C,
                         "Asia and Pacific" = "Asia-Pacific",
                         "Eastern Europe and Central Asia" = "E Europe / C Asia",
                         "Latin America and the Caribbean" = "Lat. Am and Caribbean",
                         "Middle East and Northern Africa" = "Middle East / N Africa",
                         "Sub-Saharan Africa" = "Sub-Saharan Africa",
                         "Western Europe and North America" = "W Europe / N America"))

mod <- lm(v2x_polyarchy ~ e_migdppc + Region, 
          data = dat)
```

---

```r
summary(mod)
```

```
## 
## Call:
## lm(formula = v2x_polyarchy ~ e_migdppc + Region, data = dat)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -0.48511 -0.10073 -0.00284  0.11781  0.45167 
## 
## Coefficients:
##                               Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                   0.413218   0.045254   9.131 3.48e-16 ***
## e_migdppc                     0.001246   0.001143   1.090  0.27757    
## RegionE Europe / C Asia       0.141831   0.055335   2.563  0.01132 *  
## RegionLat. Am and Caribbean   0.240547   0.058417   4.118 6.20e-05 ***
## RegionMiddle East / N Africa -0.176413   0.062168  -2.838  0.00515 ** 
## RegionSub-Saharan Africa      0.039580   0.052265   0.757  0.45001    
## RegionW Europe / N America    0.424199   0.064880   6.538 8.53e-10 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.1931 on 155 degrees of freedom
##   (16 observations deleted due to missingness)
## Multiple R-squared:  0.4802,	Adjusted R-squared:   0.46 
## F-statistic: 23.86 on 6 and 155 DF,  p-value: < 2.2e-16
```

---

# Helpers for modelling functions

**`summary(...)`** is one of a number of helper functions which works on a large number of R model objects.

Others include:

- **`fitted(...)`**, which will give you predicted values for the data you used to estimate the model
 - **`resid(...)`**, which will give you your residuals
 - **`vcov(...)`**, which will give you the variance-covariance matrix (should you want it)
 - **`predict(..., newdata = ...)`**, which will either give you the same as `fitted(...)`, or which you can be used to generate out-of-sample predictions if you supply `newdata`.
 - **`logLik(...)`** and **`AIC(...)`** for measures of model fit.
 
---

# Example predictions

```r
newdata <- data.frame(e_migdppc = quantile(dat$e_migdppc,
                                           probs = seq(0, 1, length.out = 6),
                                           na.rm = TRUE),
                      Region = "E Europe / C Asia")

newdata <- newdata %>% 
  mutate(preds = predict(mod, newdata = newdata))
head(newdata)
```

```
##   e_migdppc            Region     preds
## 1    0.6470 E Europe / C Asia 0.5558549
## 2    2.5618 E Europe / C Asia 0.5582403
## 3    7.3072 E Europe / C Asia 0.5641521
## 4   14.7970 E Europe / C Asia 0.5734827
## 5   30.6208 E Europe / C Asia 0.5931958
## 6  123.0570 E Europe / C Asia 0.7083514
```

---

# Terms within formulae

There was no statistically significant relationship between GDP per
capita and democracy in the regression we just ran.

Suppose we think that the relationship between GDP per capita and democracy is not linear, but linear-in-the-log of GDP per capita.

We could revise our model in two ways. We could:

- create a new variable, a log-transformed version of our GDP per capita variable (which we already know how to do)
 - include the `log(...)` operator directly in our regression formula.
 
This second route is preferable, because (for example) you would not have to create the same variable if you were calling predict on a new data frame.

```r
mod2 <- lm(v2x_polyarchy ~ log(e_migdppc) + Region, 
          data = dat)
```

---

# Other helpful operators in formulae

- **`scale(...)`**: standardize your variable so that you get standardized coefficients
 - **`poly(..., 2)`**: get a(n orthogonalized) polynomial of  degree two in there (i.e., a squared term. If you don't want orthogonalization, add `poly(..., 2, raw=  TRUE)`)
 - **`a*b`**: include `a`, `b` and the interaction between them.
 - **`a:b`**: include the interaction between `a` and `b`, but not the constituent terms ([don't do this](http://mattgolder.com/interactions)).
 - **`(a + b + c)^2`**: include all the pairwise interactions between `a`, `b` and `c`.
 
---

# Getting your regression out there

We'll use a package written by former colleague David Hugh-Jones.

The package is called `huxtable`. We can install it now, together with some other helper packages.

```r
install.packages(c("huxtable", "broom", "officer", "flextable"))
```

```r
library(huxtable)
```

```
## 
## Attaching package: 'huxtable'
```

```
## The following object is masked from 'package:dplyr':
## 
##     add_rownames
```

```
## The following object is masked from 'package:purrr':
## 
##     every
```

```
## The following object is masked from 'package:ggplot2':
## 
##     theme_grey
```

---

# Getting your regression out there

.pull-left[
.small[
We can call the `huxreg` function by just passing our model object(s), separating them with a comma.

```r
huxreg(mod, mod2)
```
]
]
.pull-right[
.small[
<table class="huxtable" style="border-collapse: collapse; margin-bottom: 2em; margin-top: 2em; width: 50%; margin-left: auto; margin-right: auto;  ">
<col><col><col><tr>
<td style="vertical-align: top; text-align: center; white-space: nowrap; border-style: solid solid solid solid; border-width: 0.8pt 0pt 0pt 0pt; padding: 4pt 4pt 4pt 4pt;"></td>
<td style="vertical-align: top; text-align: center; white-space: nowrap; border-style: solid solid solid solid; border-width: 0.8pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt;">(1)</td>
<td style="vertical-align: top; text-align: center; white-space: nowrap; border-style: solid solid solid solid; border-width: 0.8pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt;">(2)</td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(Intercept)</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.413 ***</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.308 ***</td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;"></td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.045)   </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.055)   </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">e_migdppc</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.001    </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">        </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;"></td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.001)   </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">        </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">RegionE Europe / C Asia</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.142 *  </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.112 *  </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;"></td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.055)   </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.054)   </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">RegionLat. Am and Caribbean</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.241 ***</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.223 ***</td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;"></td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.058)   </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.057)   </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">RegionMiddle East / N Africa</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">-0.176 ** </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">-0.213 ***</td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;"></td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.062)   </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.061)   </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">RegionSub-Saharan Africa</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.040    </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.094    </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;"></td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.052)   </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.053)   </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">RegionW Europe / N America</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.424 ***</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.356 ***</td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;"></td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.065)   </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.064)   </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">log(e_migdppc)</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">        </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.061 ** </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;"></td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt;">        </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt;">(0.018)   </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">N</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">162        </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">162        </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">R2</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.480    </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.511    </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">logLik</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">40.117    </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">45.135    </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; border-style: solid solid solid solid; border-width: 0pt 0pt 0.8pt 0pt; padding: 4pt 4pt 4pt 4pt;">AIC</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0pt 0pt 0.8pt 0pt; padding: 4pt 4pt 4pt 4pt;">-64.233    </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0pt 0pt 0.8pt 0pt; padding: 4pt 4pt 4pt 4pt;">-74.270    </td>
</tr>
<tr>
<td colspan="3" style="vertical-align: top; text-align: left; white-space: normal; padding: 4pt 4pt 4pt 4pt;"> *** p < 0.001;  ** p < 0.01;  * p < 0.05.</td>
</tr>
</table>

]
]
---

# HTML output

When we call `huxreg` it will by default give us plain text output.

That might not be enough for us.

One other output format is HTML. The following command will open a document in a web browser.

```r
quick_html(huxreg(mod, mod2))
```

Other valuable commands include `quick_docx(...)` and `quick_xlsx(...)` for export to Word and Excel respectively.

`quick_pdf(...)` may also work if you have LaTeX installed. If you do not know whether you have LaTeX installed, then you do not have LaTeX installed.

---

# Making it pretty

Our variable names are not intuitive.

We could create variables with different names, but that would be restrictive.

A better option is to pass `huxreg` a list of changes we want to make.

```r
huxreg(mod, mod2,
       coefs = c('GDP per capita, $000s' = 'e_migdppc',
                 'log(GDP per capita, $000s' = 'log(e_migdppc)',
                 'RegionE Europe / C Asia' = 'E Europe / C Asia',
                 'RegionLat. Am and Caribbean' = 'Latin America and the Caribbean',
                 'RegionMiddle East / N Africa' = 'Middle East / N Africa',
                 'RegionSub-Saharan Africa' = 'Sub-Saharan Africa',
                 'RegionW Europe / N America' = 'W Europe / N America'))
```

Yes, this is the opposite order to `recode`'s `oldthing = newthing`. **Sorry...**

---

<table class="huxtable" style="border-collapse: collapse; margin-bottom: 2em; margin-top: 2em; width: 50%; margin-left: auto; margin-right: auto;  ">
<col><col><col><tr>
<td style="vertical-align: top; text-align: center; white-space: nowrap; border-style: solid solid solid solid; border-width: 0.8pt 0pt 0pt 0pt; padding: 4pt 4pt 4pt 4pt;"></td>
<td style="vertical-align: top; text-align: center; white-space: nowrap; border-style: solid solid solid solid; border-width: 0.8pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt;">(1)</td>
<td style="vertical-align: top; text-align: center; white-space: nowrap; border-style: solid solid solid solid; border-width: 0.8pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt;">(2)</td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">GDP per capita, $000s</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.001    </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">        </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;"></td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.001)   </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">        </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">log(GDP per capita, $000s</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">        </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.061 ** </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;"></td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">        </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.018)   </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">E Europe / C Asia</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.142 *  </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.112 *  </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;"></td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.055)   </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.054)   </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">Latin America and the Caribbean</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.241 ***</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.223 ***</td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;"></td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.058)   </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.057)   </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">Middle East / N Africa</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">-0.176 ** </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">-0.213 ***</td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;"></td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.062)   </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.061)   </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">Sub-Saharan Africa</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.040    </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.094    </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;"></td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.052)   </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">(0.053)   </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">W Europe / N America</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.424 ***</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.356 ***</td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;"></td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt;">(0.065)   </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt;">(0.064)   </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">N</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">162        </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">162        </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">R2</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.480    </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">0.511    </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">logLik</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">40.117    </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; padding: 4pt 4pt 4pt 4pt;">45.135    </td>
</tr>
<tr>
<td style="vertical-align: top; text-align: left; white-space: nowrap; border-style: solid solid solid solid; border-width: 0pt 0pt 0.8pt 0pt; padding: 4pt 4pt 4pt 4pt;">AIC</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0pt 0pt 0.8pt 0pt; padding: 4pt 4pt 4pt 4pt;">-64.233    </td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0pt 0pt 0.8pt 0pt; padding: 4pt 4pt 4pt 4pt;">-74.270    </td>
</tr>
<tr>
<td colspan="3" style="vertical-align: top; text-align: left; white-space: normal; padding: 4pt 4pt 4pt 4pt;"> *** p < 0.001;  ** p < 0.01;  * p < 0.05.</td>
</tr>
</table>

---

# Changing base levels

If you're really attentive, you may have noticed that there is a missing region from our output. It's that troublesome Asia-Pacific region.

We might want to express region fx relative to a different baseline region.

We can do this by `relevel`ling our region factor.

```r
dat <- dat %>% 
  mutate(Region = relevel(Region, "W Europe / N America"))
```

We would then call `lm` again.

---

# Omitting coefficients

Suppose we thought that region fixed effects were useless control variables that no one wanted to see.

We could suppress these by just not passing them as part of the `coefs` argument.

```r
huxreg(mod, mod2,
       coefs = c('GDP per capita, $000s' = 'e_migdppc',
                 'log(GDP per capita, $000s)' = 'log(e_migdppc)'),
  note = "Regional fixed effects not reported here. ")
```

---

# Changing the summary statistics

By default, `huxtable` reports:

- the number of observations
 - (unadjusted) R-squared
 - the log-likelihood
 - the AIC value
 
for a linear model.

Suppose we wanted to include a different set of summary statistics. What could we use?

---

# Broom

`broom` is a really useful package for tidying up model outputs.

If installed, it's loaded when `huxtable` is loaded.

We can use the `glance` function to find out available summary statistics.

```r
glance(mod)
```

<table class="huxtable" style="border-collapse: collapse; margin-bottom: 2em; margin-top: 2em; width: 100%; margin-left: 0%; margin-right: auto;  ">
<col><col><col><col><col><col><col><col><col><col><col><tr>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0.4pt 0pt 0.4pt 0.4pt; padding: 4pt 4pt 4pt 4pt; font-weight: bold;">r.squared</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0.4pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; font-weight: bold;">adj.r.squared</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0.4pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; font-weight: bold;">sigma</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0.4pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; font-weight: bold;">statistic</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0.4pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; font-weight: bold;">p.value</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0.4pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; font-weight: bold;">df</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0.4pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; font-weight: bold;">logLik</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0.4pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; font-weight: bold;">AIC</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0.4pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; font-weight: bold;">BIC</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0.4pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; font-weight: bold;">deviance</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0.4pt 0.4pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; font-weight: bold;">df.residual</td>
</tr>
<tr>
<td style="vertical-align: top; text-align: right; white-space: normal; border-style: solid solid solid solid; border-width: 0pt 0pt 0.4pt 0.4pt; padding: 4pt 4pt 4pt 4pt; background-color: rgb(242, 242, 242);">0.48</td>
<td style="vertical-align: top; text-align: right; white-space: normal; border-style: solid solid solid solid; border-width: 0pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; background-color: rgb(242, 242, 242);">0.46</td>
<td style="vertical-align: top; text-align: right; white-space: normal; border-style: solid solid solid solid; border-width: 0pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; background-color: rgb(242, 242, 242);">0.193</td>
<td style="vertical-align: top; text-align: right; white-space: normal; border-style: solid solid solid solid; border-width: 0pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; background-color: rgb(242, 242, 242);">23.9</td>
<td style="vertical-align: top; text-align: right; white-space: normal; border-style: solid solid solid solid; border-width: 0pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; background-color: rgb(242, 242, 242);">7.05e-20</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; background-color: rgb(242, 242, 242);">7</td>
<td style="vertical-align: top; text-align: right; white-space: normal; border-style: solid solid solid solid; border-width: 0pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; background-color: rgb(242, 242, 242);">40.1</td>
<td style="vertical-align: top; text-align: right; white-space: normal; border-style: solid solid solid solid; border-width: 0pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; background-color: rgb(242, 242, 242);">-64.2</td>
<td style="vertical-align: top; text-align: right; white-space: normal; border-style: solid solid solid solid; border-width: 0pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; background-color: rgb(242, 242, 242);">-39.5</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0pt 0pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; background-color: rgb(242, 242, 242);">5.78</td>
<td style="vertical-align: top; text-align: right; white-space: nowrap; border-style: solid solid solid solid; border-width: 0pt 0.4pt 0.4pt 0pt; padding: 4pt 4pt 4pt 4pt; background-color: rgb(242, 242, 242);">155</td>
</tr>
</table>

and pass these to `huxreg`, adding the special statistic `nobs`:

```r
huxreg(mod, mod2,
       statistics = c('# observations' = 'nobs', 
                      '(Adjusted) R squared' = 'adj.r.squared',
                      'BIC' = 'BIC'))
```

---

# Finding out more

`huxtable` has a nice [vignette](https://cran.r-project.org/web/packages/huxtable/vignettes/huxreg.html) showing how to customize the output from regression tables.

---

# Recap

- You've learned about the R formula syntax: `depvar ~ indvar + control + ... + `
 - You've learned how to carry out linear regression using `lm`
 - You've learned about different **helper functions**
 - You've learned how to get these tables out of R into something else!

---

# What about you?

- Is there a paper in your field which uses plain old linear regression?
 - Would you be able to run that regression now in R?
 - Is there replication data available?
 
---