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.
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.
## 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:
This is an observational study because the researchers simply observe and collect data on existing conditions or behaviors. They don’t manipulate any variables. Its not possible to answer the question as currently constructed. Because of the Correlation vs Causation factor, a better question would be “Is there a correlation between instructor beauty and course evaluations?”
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?The distribution of score is skewed to the left
which suggest that the students have a tendency to score the evals above
the average. I expected the opposite since from my experience students
tend to rate professors lower when providing feedback
hist(evals$score, main = "Distribution of score in evals", xlab = "Score")
abline(v = mean(evals$score), col = "red", lwd = 2)
abline(v = median(evals$score), col = "blue", lwd = 2)
legend("topleft", legend = c("Mean", "Median"), col = c("red", "blue"), lwd = 2)score, select two other variables and
describe their relationship with each other using an appropriate
visualization.The scatterplot shows a weak negative relationship between beauty average and age. There is a considerable amount of scatter in the data points.
# Visualize the relationship between bty_avg and age of professors and added a trend line
evals |>
ggplot(aes(x = bty_avg, y = age)) +
geom_point() +
geom_smooth(method = "lm") +
labs(title = "Relationship between Beauty Avg and Age of Professors", x = "Beauty Average", y = "Age")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:
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?The number of data points in the initial plot seem to be way less than actual observations. There are 463 observations in the data-frame. The jitter plot seems to be more accurate because it adds some random jitter to the data points assisting with the overlapping.
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?The equation for the linear model is score = β₀ + β₁ * bty_avg + ε. The average beauty score seems to be a statistically significant predictor.
##
## 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
Add the line of the best fit model to your plot using the following:
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.
ggplot(data = evals, aes(x = bty_avg, y = score)) +
geom_jitter() +
geom_smooth(method = "lm", se = FALSE)It seems that the variance of the residuals might not be constant across the range of fitted values. The errors for the higher predicted scores have a smaller variance compared to errors for lower predicted scores
ggplot(data = m_bty, aes(x = .fitted, y = .resid)) +
geom_point() +
geom_hline(yintercept = 0, linetype = "dashed") +
xlab("Fitted Values") +
ylab("Residuals")We can see deviations from the diagonal line in the QQ-plot. The residuals seem to curve away from the line at the tails suggesting a slightly more extreme than expected normal distribution
# Use residual plots to evaluate whether the conditions of least squares regression are resonable
qqnorm(residuals(m_bty))
qqline(residuals(m_bty))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.
## # 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:
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.
##
## 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
{r})
There seems to be a linear relationship between the gender and score and the bty_avg. As established in the previous exercises, there is a relationship between the variables.
# Determine if there is a linear relationship between bty and score
evals |>
ggplot(aes(x = score, y = bty_avg, color = gender)) +
geom_point() +
geom_smooth(method = "lm") +
labs(title = "Relationship between Beauty Avg and Score and Gender", x = "Beauty Average", y = "Score")# Determine if residuals are normally distributed
qqnorm(residuals(m_bty_gen))
qqline(residuals(m_bty_gen))# Determine if the variablility of the residuals is nearly constantby comparing the residuals to the fitted values
ggplot(data = m_bty_gen, aes(x = .fitted, y = .resid)) +
geom_point() +
geom_hline(yintercept = 0, linetype = "dashed") +
xlab("Fitted Values") +
ylab("Residuals") +
geom_jitter()bty_avg still a significant predictor of
score? Has the addition of gender to the model
changed the parameter estimate for bty_avg?bty_avg is still a significant predictor of
score. The addition of gender made the p-value
smaller so it adds significance.
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 re-codes
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)Score = C1 + C2 * bty_avg for color photos and for non-color photos Score = C1 + C2 * bty_avg - C3. This indicates that for two professors who received the same beauty rating, the picture with color tends to have a higher score.
##
## 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
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.It creates one less dummy variable than the number of
categories in the original variable. In this case, rank has
3 levels, so R creates 2 dummy variables (rank_tenure track and
ranktenured). Rank_teaching is not represented.
##
## 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
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.
I would expect the language of the university where the professor got their degree to have the highest p-value because it doesn’t seem like it would have any association with score.
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
The highest p value goes to the cls_profssingle variable that was created by the model. This makes sense because the number of professors teaching sections in the course should not have any association with the score
## # A tibble: 15 × 5
## term estimate std.error statistic p.value
## <chr> <dbl> <dbl> <dbl> <dbl>
## 1 cls_profssingle -0.0147 0.0520 -0.282 7.78e- 1
## 2 cls_levelupper 0.0605 0.0576 1.05 2.94e- 1
## 3 cls_students 0.000455 0.000377 1.20 2.29e- 1
## 4 ranktenured -0.0973 0.0663 -1.47 1.43e- 1
## 5 pic_outfitnot formal -0.113 0.0739 -1.53 1.28e- 1
## 6 ethnicitynot minority 0.123 0.0786 1.57 1.17e- 1
## 7 ranktenure track -0.148 0.0821 -1.80 7.28e- 2
## 8 languagenon-english -0.230 0.111 -2.06 3.97e- 2
## 9 bty_avg 0.0400 0.0175 2.29 2.27e- 2
## 10 age -0.00901 0.00314 -2.87 4.27e- 3
## 11 pic_colorcolor -0.217 0.0715 -3.04 2.52e- 3
## 12 cls_perc_eval 0.00533 0.00154 3.46 5.90e- 4
## 13 gendermale 0.211 0.0518 4.07 5.54e- 5
## 14 cls_creditsone credit 0.502 0.116 4.33 1.84e- 5
## 15 (Intercept) 4.10 0.291 14.1 1.32e-37
The coefficient associated with the ethnicity variable is .123 with a p value of 0.17. The coefficient indicates that on average, professors identified as part of the “not minority” group are predicted to have a score .123 higher than professors in the minority category, when all other variables in the model are held constant.The high p-value indicates that it is not statistically significant to claim that professors that are “not minority” have higher scores than professors who are minorities.
By removing the cls_prof variable the other
coefficients and p-values for the remaining variables hardly changed.
This indicates that the dropped variable likely wasn’t collinear with
the other explanatory variables.
m_full2 <- lm(score ~ rank + gender + ethnicity + language + age + cls_perc_eval
+ cls_students + cls_level + cls_credits + bty_avg
+ pic_outfit + pic_color, data = evals)
summary(m_full2)##
## Call:
## lm(formula = score ~ rank + gender + ethnicity + language + age +
## cls_perc_eval + cls_students + cls_level + cls_credits +
## bty_avg + pic_outfit + pic_color, data = evals)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.7836 -0.3257 0.0859 0.3513 0.9551
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.0872523 0.2888562 14.150 < 2e-16 ***
## ranktenure track -0.1476746 0.0819824 -1.801 0.072327 .
## ranktenured -0.0973829 0.0662614 -1.470 0.142349
## gendermale 0.2101231 0.0516873 4.065 5.66e-05 ***
## ethnicitynot minority 0.1274458 0.0772887 1.649 0.099856 .
## languagenon-english -0.2282894 0.1111305 -2.054 0.040530 *
## age -0.0089992 0.0031326 -2.873 0.004262 **
## cls_perc_eval 0.0052888 0.0015317 3.453 0.000607 ***
## cls_students 0.0004687 0.0003737 1.254 0.210384
## cls_levelupper 0.0606374 0.0575010 1.055 0.292200
## cls_creditsone credit 0.5061196 0.1149163 4.404 1.33e-05 ***
## bty_avg 0.0398629 0.0174780 2.281 0.023032 *
## pic_outfitnot formal -0.1083227 0.0721711 -1.501 0.134080
## pic_colorcolor -0.2190527 0.0711469 -3.079 0.002205 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.4974 on 449 degrees of freedom
## Multiple R-squared: 0.187, Adjusted R-squared: 0.1634
## F-statistic: 7.943 on 13 and 449 DF, p-value: 2.336e-14
## # A tibble: 14 × 5
## term estimate std.error statistic p.value
## <chr> <dbl> <dbl> <dbl> <dbl>
## 1 cls_levelupper 0.0606 0.0575 1.05 2.92e- 1
## 2 cls_students 0.000469 0.000374 1.25 2.10e- 1
## 3 ranktenured -0.0974 0.0663 -1.47 1.42e- 1
## 4 pic_outfitnot formal -0.108 0.0722 -1.50 1.34e- 1
## 5 ethnicitynot minority 0.127 0.0773 1.65 9.99e- 2
## 6 ranktenure track -0.148 0.0820 -1.80 7.23e- 2
## 7 languagenon-english -0.228 0.111 -2.05 4.05e- 2
## 8 bty_avg 0.0399 0.0175 2.28 2.30e- 2
## 9 age -0.00900 0.00313 -2.87 4.26e- 3
## 10 pic_colorcolor -0.219 0.0711 -3.08 2.21e- 3
## 11 cls_perc_eval 0.00529 0.00153 3.45 6.07e- 4
## 12 gendermale 0.210 0.0517 4.07 5.66e- 5
## 13 cls_creditsone credit 0.506 0.115 4.40 1.33e- 5
## 14 (Intercept) 4.09 0.289 14.1 7.53e-38
\[ \begin{aligned} \widehat{score} &= \hat{\beta}_0 + \hat{\beta}_1 \times language \\ &+ \hat{\beta}_2 \times ethnicity \\ &+\hat{\beta}_3 \times age \\ &+ \hat{\beta}_4 \times pic\_color \\ &+ \hat{\beta}_5 \times bty\_avg \\ &+ \hat{\beta}_6 \times cls\_perc\_eval \\ &+ \hat{\beta}_8 \times gender \\ &+ \hat{\beta}_9 \times cls\_credits \end{aligned} \]
## -------------STEP 1 -------------
## The drop statistics :
## Single term deletions
##
## Model:
## score ~ rank + gender + ethnicity + language + age + cls_perc_eval +
## cls_students + cls_level + cls_profs + cls_credits + bty_avg +
## pic_outfit + pic_color
## Df Sum of Sq RSS AIC F value Pr(>F)
## <none> 111.08 -630.90
## rank 2 0.8930 111.98 -631.19 1.8007 0.1663804
## gender 1 4.1085 115.19 -616.09 16.5694 5.544e-05 ***
## ethnicity 1 0.6117 111.70 -630.36 2.4668 0.1169791
## language 1 1.0557 112.14 -628.52 4.2576 0.0396509 *
## age 1 2.0456 113.13 -624.45 8.2499 0.0042688 **
## cls_perc_eval 1 2.9698 114.06 -620.69 11.9769 0.0005903 ***
## cls_students 1 0.3599 111.44 -631.40 1.4513 0.2289607
## cls_level 1 0.2740 111.36 -631.76 1.1052 0.2936925
## cls_profs 1 0.0197 111.11 -632.82 0.0795 0.7780566
## cls_credits 1 4.6495 115.73 -613.92 18.7510 1.839e-05 ***
## bty_avg 1 1.2967 112.38 -627.53 5.2294 0.0226744 *
## pic_outfit 1 0.5768 111.66 -630.50 2.3262 0.1279153
## pic_color 1 2.2893 113.37 -623.46 9.2328 0.0025162 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## --------
## Term dropped in step 1 : cls_profs
## --------
##
## -------------STEP 2 -------------
## The drop statistics :
## Single term deletions
##
## Model:
## score ~ rank + gender + ethnicity + language + age + cls_perc_eval +
## cls_students + cls_level + cls_credits + bty_avg + pic_outfit +
## pic_color
## Df Sum of Sq RSS AIC F value Pr(>F)
## <none> 111.11 -632.82
## rank 2 0.8939 112.00 -633.11 1.8063 0.1654529
## gender 1 4.0895 115.19 -618.08 16.5265 5.665e-05 ***
## ethnicity 1 0.6728 111.78 -632.02 2.7191 0.0998556 .
## language 1 1.0442 112.15 -630.49 4.2199 0.0405303 *
## age 1 2.0422 113.15 -626.39 8.2529 0.0042616 **
## cls_perc_eval 1 2.9502 114.06 -622.69 11.9224 0.0006072 ***
## cls_students 1 0.3893 111.49 -633.20 1.5733 0.2103843
## cls_level 1 0.2752 111.38 -633.67 1.1121 0.2922000
## cls_credits 1 4.7999 115.90 -615.24 19.3974 1.329e-05 ***
## bty_avg 1 1.2872 112.39 -629.49 5.2018 0.0230315 *
## pic_outfit 1 0.5574 111.66 -632.50 2.2527 0.1340803
## pic_color 1 2.3457 113.45 -625.15 9.4795 0.0022052 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## --------
## Term dropped in step 2 : cls_level
## --------
##
## -------------STEP 3 -------------
## The drop statistics :
## Single term deletions
##
## Model:
## score ~ rank + gender + ethnicity + language + age + cls_perc_eval +
## cls_students + cls_credits + bty_avg + pic_outfit + pic_color
## Df Sum of Sq RSS AIC F value Pr(>F)
## <none> 111.38 -633.67
## rank 2 0.8140 112.19 -634.30 1.6443 0.1943109
## gender 1 3.8989 115.28 -619.74 15.7525 8.400e-05 ***
## ethnicity 1 0.8698 112.25 -632.07 3.5142 0.0614914 .
## language 1 0.9015 112.28 -631.94 3.6423 0.0569659 .
## age 1 1.9342 113.31 -627.70 7.8147 0.0054040 **
## cls_perc_eval 1 3.0290 114.41 -623.25 12.2380 0.0005148 ***
## cls_students 1 0.2459 111.63 -634.65 0.9934 0.3194514
## cls_credits 1 4.5296 115.91 -617.22 18.3006 2.306e-05 ***
## bty_avg 1 1.3694 112.75 -630.02 5.5329 0.0190925 *
## pic_outfit 1 0.6618 112.04 -632.93 2.6736 0.1027219
## pic_color 1 2.0777 113.46 -627.12 8.3942 0.0039478 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## --------
## Term dropped in step 3 : cls_students
## --------
##
## -------------STEP 4 -------------
## The drop statistics :
## Single term deletions
##
## Model:
## score ~ rank + gender + ethnicity + language + age + cls_perc_eval +
## cls_credits + bty_avg + pic_outfit + pic_color
## Df Sum of Sq RSS AIC F value Pr(>F)
## <none> 111.63 -634.65
## rank 2 0.7892 112.42 -635.39 1.5943 0.2041875
## gender 1 4.0945 115.72 -619.97 16.5430 5.613e-05 ***
## ethnicity 1 0.8832 112.51 -633.00 3.5683 0.0595353 .
## language 1 1.0338 112.66 -632.38 4.1769 0.0415580 *
## age 1 2.1163 113.74 -627.96 8.5504 0.0036285 **
## cls_perc_eval 1 2.7922 114.42 -625.21 11.2814 0.0008493 ***
## cls_credits 1 4.5163 116.14 -618.29 18.2472 2.368e-05 ***
## bty_avg 1 1.5783 113.20 -630.15 6.3770 0.0119026 *
## pic_outfit 1 0.9700 112.60 -632.65 3.9191 0.0483466 *
## pic_color 1 1.9477 113.57 -628.64 7.8693 0.0052456 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## --------
## Term dropped in step 4 : rank
## --------
##
## -------------STEP 5 -------------
## The drop statistics :
## Single term deletions
##
## Model:
## score ~ gender + ethnicity + language + age + cls_perc_eval +
## cls_credits + bty_avg + pic_outfit + pic_color
## Df Sum of Sq RSS AIC F value Pr(>F)
## <none> 112.42 -635.39
## gender 1 4.0577 116.47 -620.97 16.3513 6.184e-05 ***
## ethnicity 1 1.1790 113.59 -632.56 4.7508 0.0297983 *
## language 1 1.3403 113.75 -631.90 5.4010 0.0205673 *
## age 1 1.6847 114.10 -630.50 6.7888 0.0094749 **
## cls_perc_eval 1 2.9147 115.33 -625.54 11.7455 0.0006655 ***
## cls_credits 1 6.1208 118.54 -612.84 24.6649 9.681e-07 ***
## bty_avg 1 1.8553 114.27 -629.81 7.4762 0.0064972 **
## pic_outfit 1 0.7141 113.13 -634.46 2.8775 0.0905102 .
## pic_color 1 1.7841 114.20 -630.10 7.1895 0.0076009 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## --------
## Term dropped in step 5 : pic_outfit
## --------
##
## -------------STEP 6 -------------
## The drop statistics :
## Single term deletions
##
## Model:
## score ~ gender + ethnicity + language + age + cls_perc_eval +
## cls_credits + bty_avg + pic_color
## Df Sum of Sq RSS AIC F value Pr(>F)
## <none> 113.13 -634.46
## gender 1 4.2526 117.38 -619.37 17.0660 4.299e-05 ***
## ethnicity 1 1.2393 114.37 -631.41 4.9734 0.026228 *
## language 1 0.9862 114.11 -632.44 3.9576 0.047259 *
## age 1 1.3350 114.46 -631.03 5.3574 0.021080 *
## cls_perc_eval 1 2.6225 115.75 -625.85 10.5243 0.001265 **
## cls_credits 1 5.8690 119.00 -613.04 23.5530 1.674e-06 ***
## bty_avg 1 2.2663 115.39 -627.28 9.0950 0.002707 **
## pic_color 1 1.9951 115.12 -628.36 8.0068 0.004867 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Call:
## lm(formula = score ~ gender + ethnicity + language + age + cls_perc_eval +
## cls_credits + bty_avg + pic_color, data = evals)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.85320 -0.32394 0.09984 0.37930 0.93610
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.771922 0.232053 16.255 < 2e-16 ***
## gendermale 0.207112 0.050135 4.131 4.30e-05 ***
## ethnicitynot minority 0.167872 0.075275 2.230 0.02623 *
## languagenon-english -0.206178 0.103639 -1.989 0.04726 *
## age -0.006046 0.002612 -2.315 0.02108 *
## cls_perc_eval 0.004656 0.001435 3.244 0.00127 **
## cls_creditsone credit 0.505306 0.104119 4.853 1.67e-06 ***
## bty_avg 0.051069 0.016934 3.016 0.00271 **
## pic_colorcolor -0.190579 0.067351 -2.830 0.00487 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.4992 on 454 degrees of freedom
## Multiple R-squared: 0.1722, Adjusted R-squared: 0.1576
## F-statistic: 11.8 on 8 and 454 DF, p-value: 2.58e-15
The standard regression assumptions include the following about residuals/errors:
The error has a normal distribution (normality assumption).
The errors have mean zero.
The errors have same but unknown variance (homoscedasticity assumption).
The error are independent of each other (independent errors assumption).
# Import olsrr package
if(!require("olsrr")) {install.packages("olsrr"); library("olsrr")}
par(mfrow = c(2,2 ))
# Residual QQ Plot - Graph for detecting violation of normality assumption.
ols_plot_resid_qq(m_best)
# Residual vs Fitted Values Plot - Graph for detecting violation of linearity assumption.
ols_plot_resid_fit(m_best)
# Residual Histogram - Graph for detecting violation of normality assumption.
ols_plot_resid_hist(m_best)This could potentially add independence errors. Courses taught by the same professor could affect the scores of all their courses
Professor Characteristics:
Gender: The analysis suggests a slight positive association between being male and receiving a higher score.
Ethnicity: The analysis shows a slight positive association between not belonging to a minority group and receiving a higher score.
Age: There’s a weak negative association between age and score, suggesting a slight decrease in score with increasing age. However, the magnitude of this effect is likely small.
Course Characteristics:
Language: Courses taught in English are predicted to have slightly higher scores compared to those taught in non-English languages.
Course Evaluations: Courses with a higher percentage of students giving positive evaluations (cls_perc_eval) tend to have higher scores. This suggests that student perception of the course plays a significant role in the evaluation.
Course Credits: One-credit courses are predicted to have higher scores compared to courses with a different number of credits. .
Other Considerations:
No I would not be comfortable generalizing my conclusions because the data used in the model comes from a single university. This can lead to biased conclusions. Also there can be confounding variables such as class size, course difficulty, department etc.