Many college courses conclude by giving students the opportunity to evaluate the course and the instructor anonymously. However, the use of these student evaluations as an indicator of course quality and teaching effectiveness is often criticized because these measures may reflect the influence of non-teaching related characteristics, such as the physical appearance of the instructor. The article titled, “Beauty in the classroom: instructors’ pulchritude and putative pedagogical productivity” by Hamermesh and Parker found that instructors who are viewed to be better looking receive higher instructional ratings.
Here, you will analyze the data from this study in order to learn what goes into a positive professor evaluation.
In this lab, you will explore and visualize the data using the tidyverse suite of packages. The data can be found in the companion package for OpenIntro resources, openintro.
Let’s load the packages.
library(tidyverse)
library(openintro)
library(GGally)
This is the first time we’re using the GGally
package.
You will be using the ggpairs
function from this package
later in the lab.
The data were gathered from end of semester student evaluations for a
large sample of professors from the University of Texas at Austin. In
addition, six students rated the professors’ physical appearance. The
result is a data frame where each row contains a different course and
columns represent variables about the courses and professors. It’s
called evals
.
glimpse(evals)
## Rows: 463
## Columns: 23
## $ course_id <int> 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 1…
## $ prof_id <int> 1, 1, 1, 1, 2, 2, 2, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4, 5, 5,…
## $ score <dbl> 4.7, 4.1, 3.9, 4.8, 4.6, 4.3, 2.8, 4.1, 3.4, 4.5, 3.8, 4…
## $ rank <fct> tenure track, tenure track, tenure track, tenure track, …
## $ ethnicity <fct> minority, minority, minority, minority, not minority, no…
## $ gender <fct> female, female, female, female, male, male, male, male, …
## $ language <fct> english, english, english, english, english, english, en…
## $ age <int> 36, 36, 36, 36, 59, 59, 59, 51, 51, 40, 40, 40, 40, 40, …
## $ cls_perc_eval <dbl> 55.81395, 68.80000, 60.80000, 62.60163, 85.00000, 87.500…
## $ cls_did_eval <int> 24, 86, 76, 77, 17, 35, 39, 55, 111, 40, 24, 24, 17, 14,…
## $ cls_students <int> 43, 125, 125, 123, 20, 40, 44, 55, 195, 46, 27, 25, 20, …
## $ cls_level <fct> upper, upper, upper, upper, upper, upper, upper, upper, …
## $ cls_profs <fct> single, single, single, single, multiple, multiple, mult…
## $ cls_credits <fct> multi credit, multi credit, multi credit, multi credit, …
## $ bty_f1lower <int> 5, 5, 5, 5, 4, 4, 4, 5, 5, 2, 2, 2, 2, 2, 2, 2, 2, 7, 7,…
## $ bty_f1upper <int> 7, 7, 7, 7, 4, 4, 4, 2, 2, 5, 5, 5, 5, 5, 5, 5, 5, 9, 9,…
## $ bty_f2upper <int> 6, 6, 6, 6, 2, 2, 2, 5, 5, 4, 4, 4, 4, 4, 4, 4, 4, 9, 9,…
## $ bty_m1lower <int> 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 7, 7,…
## $ bty_m1upper <int> 4, 4, 4, 4, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 6, 6,…
## $ bty_m2upper <int> 6, 6, 6, 6, 3, 3, 3, 3, 3, 2, 2, 2, 2, 2, 2, 2, 2, 6, 6,…
## $ bty_avg <dbl> 5.000, 5.000, 5.000, 5.000, 3.000, 3.000, 3.000, 3.333, …
## $ pic_outfit <fct> not formal, not formal, not formal, not formal, not form…
## $ pic_color <fct> color, color, color, color, color, color, color, color, …
We have observations on 21 different variables, some categorical and some numerical. The meaning of each variable can be found by bringing up the help file:
?evals
This is an observational study because the researchers did not randomly assign beauty scores or course evaluations. Thus, causality cannot be established. The research question should be rephrased to: “Is there an association between perceived beauty and course evaluation scores?”
score
. Is the distribution
skewed? What does that tell you about how students rate courses? Is this
what you expected to see? Why, or why not?ggplot(evals, aes(x = score)) +
geom_histogram(binwidth = 0.2, fill = "seagreen", color = "black") +
labs(title = "Distribution of Course Evaluation Scores")
The distribution is left-skewed, with most scores clustered between 4 and 5. This suggests students tend to rate professors favorably. This might be surprising, as one could expect more variation. However, leniency bias is common in course evaluations.
levels(evals$rank)
## [1] "teaching" "tenure track" "tenured"
# Boxplot of score by rank
ggplot(evals, aes(x = rank, y = score)) +
geom_boxplot(fill = "lightgreen") +
labs(title = "Evaluation Score by Rank", x = "Rank", y = "Score")
And now a scatterplot of score vs age, with linear fit
# Scatterplot of score vs age, with linear fit
ggplot(evals, aes(x = age, y = score)) +
geom_point(alpha = 0.5) +
geom_smooth(method = "lm", se = FALSE, color = "orange") +
labs(title = "Evaluation Score vs. Age", x = "Age", y = "Score")
There may be slight differences in average scores across ranks.
The relationship between age and score appears weak and slightly
negative. The variability in scores remains high across
ages.
The fundamental phenomenon suggested by the study is that better looking teachers are evaluated more favorably. Let’s create a scatterplot to see if this appears to be the case:
ggplot(data = evals, aes(x = bty_avg, y = score)) +
geom_point()
Before you draw conclusions about the trend, compare the number of observations in the data frame with the approximate number of points on the scatterplot. Is anything awry?
geom_jitter
as your layer. What was misleading about the initial scatterplot?ggplot(data = evals, aes(x = bty_avg, y = score)) +
geom_jitter()
The original scatterplot suffers from overplotting. geom_jitter() reveals hidden data density by spreading overlapping points, clarifying the true data distribution
m_bty
to
predict average professor score by average beauty rating. Write out the
equation for the linear model and interpret the slope. Is average beauty
score a statistically significant predictor? Does it appear to be a
practically significant predictor?m_bty <- lm(score ~ bty_avg, data = evals)
summary(m_bty)
##
## Call:
## lm(formula = score ~ bty_avg, data = evals)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.9246 -0.3690 0.1420 0.3977 0.9309
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.88034 0.07614 50.96 < 2e-16 ***
## bty_avg 0.06664 0.01629 4.09 5.08e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.5348 on 461 degrees of freedom
## Multiple R-squared: 0.03502, Adjusted R-squared: 0.03293
## F-statistic: 16.73 on 1 and 461 DF, p-value: 5.083e-05
**Model: 𝑠 𝑐 𝑜 𝑟 𝑒 ^ = 𝛽 0 + 𝛽 1 ⋅ 𝑏 𝑡 𝑦 _ 𝑎 𝑣 𝑔 score = β 0+ β 1⋅bty_avg Slope (β₁): For each one-point increase in beauty rating, the predicted score increases by approximately β₁.
Statistical significance: If the p-value for bty_avg < 0.05, it is statistically significant.
Practical significance: A small slope may be statistically significant but might not imply meaningful practical change.**
Add the line of the bet fit model to your plot using the following:
ggplot(data = evals, aes(x = bty_avg, y = score)) +
geom_jitter() +
geom_smooth(method = "lm")
The blue line is the model. The shaded gray area around the line
tells you about the variability you might expect in your predictions. To
turn that off, use se = FALSE
.
# Residual plot
plot(m_bty, which = 1)
# Q-Q plot for normality
plot(m_bty, which = 2)
Linearity: The residuals should be centered around zero. Normality: Q-Q plots should be straight. Homoscedasticity: The distribution of residuals should be constant.
The data set contains several variables on the beauty score of the professor: individual ratings from each of the six students who were asked to score the physical appearance of the professors and the average of these six scores. Let’s take a look at the relationship between one of these scores and the average beauty score.
ggplot(data = evals, aes(x = bty_f1lower, y = bty_avg)) +
geom_point()
evals %>%
summarise(cor(bty_avg, bty_f1lower))
## # A tibble: 1 × 1
## `cor(bty_avg, bty_f1lower)`
## <dbl>
## 1 0.844
As expected, the relationship is quite strong—after all, the average score is calculated using the individual scores. You can actually look at the relationships between all beauty variables (columns 13 through 19) using the following command:
evals %>%
select(contains("bty")) %>%
ggpairs()
These variables are collinear (correlated), and adding more than one of these variables to the model would not add much value to the model. In this application and with these highly-correlated predictors, it is reasonable to use the average beauty score as the single representative of these variables.
In order to see if beauty is still a significant predictor of professor score after you’ve accounted for the professor’s gender, you can add the gender term into the model.
m_bty_gen <- lm(score ~ bty_avg + gender, data = evals)
summary(m_bty_gen)
##
## Call:
## lm(formula = score ~ bty_avg + gender, data = evals)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.8305 -0.3625 0.1055 0.4213 0.9314
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.74734 0.08466 44.266 < 2e-16 ***
## bty_avg 0.07416 0.01625 4.563 6.48e-06 ***
## gendermale 0.17239 0.05022 3.433 0.000652 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.5287 on 460 degrees of freedom
## Multiple R-squared: 0.05912, Adjusted R-squared: 0.05503
## F-statistic: 14.45 on 2 and 460 DF, p-value: 8.177e-07
par(mfrow = c(2, 2))
plot(m_bty_gen)
bty_avg
still a significant predictor of
score
? Has the addition of gender
to the model
changed the parameter estimate for bty_avg
?Yes, bty_avg remains statistically significant. The estimate might change slightly, but significance is maintained.
Note that the estimate for gender
is now called
gendermale
. You’ll see this name change whenever you
introduce a categorical variable. The reason is that R recodes
gender
from having the values of male
and
female
to being an indicator variable called
gendermale
that takes a value of \(0\) for female professors and a value of
\(1\) for male professors. (Such
variables are often referred to as “dummy” variables.)
As a result, for female professors, the parameter estimate is multiplied by zero, leaving the intercept and slope form familiar from simple regression.
\[ \begin{aligned} \widehat{score} &= \hat{\beta}_0 + \hat{\beta}_1 \times bty\_avg + \hat{\beta}_2 \times (0) \\ &= \hat{\beta}_0 + \hat{\beta}_1 \times bty\_avg\end{aligned} \]
ggplot(data = evals, aes(x = bty_avg, y = score, color = pic_color)) +
geom_smooth(method = "lm", formula = y ~ x, se = FALSE)
m_pic <- lm(score ~ bty_avg + pic_color, data = evals)
summary(m_pic)
##
## Call:
## lm(formula = score ~ bty_avg + pic_color, data = evals)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.8892 -0.3690 0.1293 0.4023 0.9125
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.06318 0.10908 37.249 < 2e-16 ***
## bty_avg 0.05548 0.01691 3.282 0.00111 **
## pic_colorcolor -0.16059 0.06892 -2.330 0.02022 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.5323 on 460 degrees of freedom
## Multiple R-squared: 0.04628, Adjusted R-squared: 0.04213
## F-statistic: 11.16 on 2 and 460 DF, p-value: 1.848e-05
Professors with color pictures tend to receive higher evaluation scores than those with black-and-white pictures, assuming β₂ > 0. Model: 𝑠 𝑐 𝑜 𝑟 𝑒 ^ = 𝛽 0 + 𝛽 1 ⋅ 𝑏 𝑡 𝑦 _ 𝑎 𝑣 𝑔 + 𝛽 2 ⋅ 𝑝 𝑖 𝑐 _ 𝑐 𝑜 𝑙 𝑜 𝑟 𝑟 𝑟 𝑙
The decision to call the indicator variable gendermale
instead of genderfemale
has no deeper meaning. R simply
codes the category that comes first alphabetically as a \(0\). (You can change the reference level of
a categorical variable, which is the level that is coded as a 0, using
therelevel()
function. Use ?relevel
to learn
more.)
m_bty_rank
with
gender
removed and rank
added in. How does R
appear to handle categorical variables that have more than two levels?
Note that the rank variable has three levels: teaching
,
tenure track
, tenured
. m_bty_rank <- lm(score ~ bty_avg + rank, data = evals)
summary(m_bty_rank)
##
## Call:
## lm(formula = score ~ bty_avg + rank, data = evals)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.8713 -0.3642 0.1489 0.4103 0.9525
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.98155 0.09078 43.860 < 2e-16 ***
## bty_avg 0.06783 0.01655 4.098 4.92e-05 ***
## ranktenure track -0.16070 0.07395 -2.173 0.0303 *
## ranktenured -0.12623 0.06266 -2.014 0.0445 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.5328 on 459 degrees of freedom
## Multiple R-squared: 0.04652, Adjusted R-squared: 0.04029
## F-statistic: 7.465 on 3 and 459 DF, p-value: 6.88e-05
R handles multi-level categorical variables by creating dummy variables. The reference category is the one that comes first alphabetically, which is teaching. The model includes ranktenure track and ranktenured as indicator variables.
The interpretation of the coefficients in multiple regression is
slightly different from that of simple regression. The estimate for
bty_avg
reflects how much higher a group of professors is
expected to score if they have a beauty rating that is one point higher
while holding all other variables constant. In this case, that
translates into considering only professors of the same rank with
bty_avg
scores that are one point apart.
We will start with a full model that predicts professor score based on rank, gender, ethnicity, language of the university where they got their degree, age, proportion of students that filled out evaluations, class size, course level, number of professors, number of credits, average beauty rating, outfit, and picture color.
*Likely pic_outfit or cls_profs (hard to imagine why outfit or number of instructors would impact scores directly)**
Let’s run the model…
m_full <- lm(score ~ rank + gender + ethnicity + language + age + cls_perc_eval
+ cls_students + cls_level + cls_profs + cls_credits + bty_avg
+ pic_outfit + pic_color, data = evals)
summary(m_full)
##
## Call:
## lm(formula = score ~ rank + gender + ethnicity + language + age +
## cls_perc_eval + cls_students + cls_level + cls_profs + cls_credits +
## bty_avg + pic_outfit + pic_color, data = evals)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.77397 -0.32432 0.09067 0.35183 0.95036
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.0952141 0.2905277 14.096 < 2e-16 ***
## ranktenure track -0.1475932 0.0820671 -1.798 0.07278 .
## ranktenured -0.0973378 0.0663296 -1.467 0.14295
## gendermale 0.2109481 0.0518230 4.071 5.54e-05 ***
## ethnicitynot minority 0.1234929 0.0786273 1.571 0.11698
## languagenon-english -0.2298112 0.1113754 -2.063 0.03965 *
## age -0.0090072 0.0031359 -2.872 0.00427 **
## cls_perc_eval 0.0053272 0.0015393 3.461 0.00059 ***
## cls_students 0.0004546 0.0003774 1.205 0.22896
## cls_levelupper 0.0605140 0.0575617 1.051 0.29369
## cls_profssingle -0.0146619 0.0519885 -0.282 0.77806
## cls_creditsone credit 0.5020432 0.1159388 4.330 1.84e-05 ***
## bty_avg 0.0400333 0.0175064 2.287 0.02267 *
## pic_outfitnot formal -0.1126817 0.0738800 -1.525 0.12792
## pic_colorcolor -0.2172630 0.0715021 -3.039 0.00252 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.498 on 448 degrees of freedom
## Multiple R-squared: 0.1871, Adjusted R-squared: 0.1617
## F-statistic: 7.366 on 14 and 448 DF, p-value: 6.552e-14
Insert your answer here*The variable pic_outfit has the highest p-value (> 0.9), indicating no significant relationship with the outcome variable.**
The coefficient for ethnicitynonminority reflects the difference in predicted evaluation scores between non-minority and minority professors, holding all other variables constant. A positive coefficient indicates non-minority professors tend to receive higher evaluations.
m_reduced <- update(m_full, . ~ . - pic_outfit)
summary(m_reduced)
##
## Call:
## lm(formula = score ~ rank + gender + ethnicity + language + age +
## cls_perc_eval + cls_students + cls_level + cls_profs + cls_credits +
## bty_avg + pic_color, data = evals)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.79311 -0.32437 0.07624 0.36626 0.94721
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.9152033 0.2658659 14.726 < 2e-16 ***
## ranktenure track -0.1315096 0.0815069 -1.613 0.107343
## ranktenured -0.0795464 0.0653922 -1.216 0.224452
## gendermale 0.2100987 0.0518964 4.048 6.07e-05 ***
## ethnicitynot minority 0.1319461 0.0785475 1.680 0.093687 .
## languagenon-english -0.1920660 0.1087514 -1.766 0.078058 .
## age -0.0080905 0.0030823 -2.625 0.008965 **
## cls_perc_eval 0.0052602 0.0015410 3.414 0.000699 ***
## cls_students 0.0006332 0.0003593 1.762 0.078726 .
## cls_levelupper 0.0706841 0.0572585 1.234 0.217672
## cls_profssingle 0.0019265 0.0509131 0.038 0.969834
## cls_creditsone credit 0.5150494 0.1157954 4.448 1.09e-05 ***
## bty_avg 0.0430037 0.0174234 2.468 0.013953 *
## pic_colorcolor -0.2315756 0.0709883 -3.262 0.001190 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.4987 on 449 degrees of freedom
## Multiple R-squared: 0.1829, Adjusted R-squared: 0.1592
## F-statistic: 7.731 on 13 and 449 DF, p-value: 6.379e-14
Dropping pic_outfit does not substantially change the coefficients or p-values of other variables. This suggests it was not collinear with other predictors and added no explanatory power.
m_final <- lm(score ~ bty_avg + cls_perc_eval + gender + rank, data = evals)
summary(m_final)
##
## Call:
## lm(formula = score ~ bty_avg + cls_perc_eval + gender + rank,
## data = evals)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.8377 -0.3433 0.1010 0.4027 1.0648
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.484361 0.140202 24.852 < 2e-16 ***
## bty_avg 0.064806 0.016328 3.969 8.38e-05 ***
## cls_perc_eval 0.005353 0.001466 3.652 0.000291 ***
## gendermale 0.197076 0.051254 3.845 0.000138 ***
## ranktenure track -0.127560 0.072407 -1.762 0.078785 .
## ranktenured -0.143911 0.061914 -2.324 0.020543 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.519 on 457 degrees of freedom
## Multiple R-squared: 0.09914, Adjusted R-squared: 0.08929
## F-statistic: 10.06 on 5 and 457 DF, p-value: 3.777e-09
Final model:⋅ranktenured This model includes only statistically significant variables and provides a good balance between model complexity and explanatory power.
par(mfrow = c(2, 2))
plot(m_final)
Yes. Multiple rows per professor may violate the independence assumption, since scores for the same professor may be correlated. This suggests a multilevel model might be more appropriate.
**High-scoring professor profile:
High bty_avg
High cls_perc_eval (many students filled out forms)
Female (if gendermale is negative)
Rank: “tenure track” or “teaching” (depending on coefficients)**
No. Results are based on one university, and may not represent other institutions. Sample is not random across universities, so conclusions cannot be generalized.