# Estimating Monotonic Effects with brms

## Introduction

This vignette is about monotonic effects, a special way of handling discrete predictors that are on an ordinal or higher scale (Bürkner & Charpentier, in review). A predictor, which we want to model as monotonic (i.e., having a monotonically increasing or decreasing relationship with the response), must either be integer valued or an ordered factor. As opposed to a continuous predictor, predictor categories (or integers) are not assumed to be equidistant with respect to their effect on the response variable. Instead, the distance between adjacent predictor categories (or integers) is estimated from the data and may vary across categories. This is realized by parameterizing as follows: One parameter, $$b$$, takes care of the direction and size of the effect similar to an ordinary regression parameter. If the monotonic effect is used in a linear model, $$b$$ can be interpreted as the expected average difference between two adjacent categories of the ordinal predictor. An additional parameter vector, $$\zeta$$, estimates the normalized distances between consecutive predictor categories which thus defines the shape of the monotonic effect. For a single monotonic predictor, $$x$$, the linear predictor term of observation $$n$$ looks as follows:

$\eta_n = b D \sum_{i = 1}^{x_n} \zeta_i$

The parameter $$b$$ can take on any real value, while $$\zeta$$ is a simplex, which means that it satisfies $$\zeta_i \in [0,1]$$ and $$\sum_{i = 1}^D \zeta_i = 1$$ with $$D$$ being the number of elements of $$\zeta$$. Equivalently, $$D$$ is the number of categories (or highest integer in the data) minus 1, since we start counting categories from zero to simplify the notation.

## A Simple Monotonic Model

A main application of monotonic effects are ordinal predictors that can be modeled this way without falsely treating them either as continuous or as unordered categorical predictors. In Psychology, for instance, this kind of data is omnipresent in the form of Likert scale items, which are often treated as being continuous for convenience without ever testing this assumption. As an example, suppose we are interested in the relationship of yearly income (in $) and life satisfaction measured on an arbitrary scale from 0 to 100. Usually, people are not asked for the exact income. Instead, they are asked to rank themselves in one of certain classes, say: ‘below 20k’, ‘between 20k and 40k’, ‘between 40k and 100k’ and ‘above 100k’. We use some simulated data for illustration purposes. income_options <- c("below_20", "20_to_40", "40_to_100", "greater_100") income <- factor(sample(income_options, 100, TRUE), levels = income_options, ordered = TRUE) mean_ls <- c(30, 60, 70, 75) ls <- mean_ls[income] + rnorm(100, sd = 7) dat <- data.frame(income, ls) We now proceed with analyzing the data modeling income as a monotonic effect. fit1 <- brm(ls ~ mo(income), data = dat) The summary methods yield summary(fit1)  Family: gaussian Links: mu = identity; sigma = identity Formula: ls ~ mo(income) Data: dat (Number of observations: 100) Draws: 4 chains, each with iter = 2000; warmup = 1000; thin = 1; total post-warmup draws = 4000 Population-Level Effects: Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS Intercept 30.07 1.41 27.37 32.96 1.00 2402 2095 moincome 15.58 0.64 14.30 16.84 1.00 2369 2014 Simplex Parameters: Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS moincome1 0.63 0.04 0.56 0.69 1.00 2883 2504 moincome1 0.26 0.05 0.16 0.35 1.00 3753 2625 moincome1 0.12 0.04 0.03 0.20 1.00 3307 1870 Family Specific Parameters: Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS sigma 7.20 0.52 6.28 8.29 1.00 3586 2822 Draws were sampled using sampling(NUTS). For each parameter, Bulk_ESS and Tail_ESS are effective sample size measures, and Rhat is the potential scale reduction factor on split chains (at convergence, Rhat = 1). plot(fit1, variable = "simo", regex = TRUE) plot(conditional_effects(fit1)) The distributions of the simplex parameter of income, as shown in the plot method, demonstrate that the largest difference (about 70% of the difference between minimum and maximum category) is between the first two categories. Now, let’s compare of monotonic model with two common alternative models. (a) Assume income to be continuous: dat$income_num <- as.numeric(dat$income) fit2 <- brm(ls ~ income_num, data = dat) summary(fit2)  Family: gaussian Links: mu = identity; sigma = identity Formula: ls ~ income_num Data: dat (Number of observations: 100) Draws: 4 chains, each with iter = 2000; warmup = 1000; thin = 1; total post-warmup draws = 4000 Population-Level Effects: Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS Intercept 21.06 2.21 16.87 25.50 1.00 3849 2522 income_num 15.07 0.81 13.46 16.66 1.00 3617 2775 Family Specific Parameters: Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS sigma 9.53 0.68 8.31 10.97 1.00 3563 2674 Draws were sampled using sampling(NUTS). For each parameter, Bulk_ESS and Tail_ESS are effective sample size measures, and Rhat is the potential scale reduction factor on split chains (at convergence, Rhat = 1). or (b) Assume income to be an unordered factor: contrasts(dat$income) <- contr.treatment(4)
fit3 <- brm(ls ~ income, data = dat)
summary(fit3)
 Family: gaussian
Links: mu = identity; sigma = identity
Formula: ls ~ income
Data: dat (Number of observations: 100)
Draws: 4 chains, each with iter = 2000; warmup = 1000; thin = 1;
total post-warmup draws = 4000

Population-Level Effects:
Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
Intercept    29.86      1.36    27.12    32.44 1.00     3229     2556
income2      29.42      1.98    25.56    33.29 1.00     3630     2883
income3      41.56      2.28    37.17    46.01 1.00     3691     3167
income4      47.03      1.90    43.36    50.78 1.00     3638     3274

Family Specific Parameters:
Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sigma     7.19      0.53     6.23     8.29 1.00     4066     2891

Draws were sampled using sampling(NUTS). For each parameter, Bulk_ESS
and Tail_ESS are effective sample size measures, and Rhat is the potential
scale reduction factor on split chains (at convergence, Rhat = 1).

We can easily compare the fit of the three models using leave-one-out cross-validation.

loo(fit1, fit2, fit3)
Output of model 'fit1':

Computed from 4000 by 100 log-likelihood matrix

Estimate   SE
elpd_loo   -341.0  6.5
p_loo         4.6  0.6
looic       682.0 13.0
------
Monte Carlo SE of elpd_loo is 0.0.

All Pareto k estimates are good (k < 0.5).
See help('pareto-k-diagnostic') for details.

Output of model 'fit2':

Computed from 4000 by 100 log-likelihood matrix

Estimate   SE
elpd_loo   -368.1  5.5
p_loo         2.6  0.3
looic       736.3 11.1
------
Monte Carlo SE of elpd_loo is 0.0.

All Pareto k estimates are good (k < 0.5).
See help('pareto-k-diagnostic') for details.

Output of model 'fit3':

Computed from 4000 by 100 log-likelihood matrix

Estimate   SE
elpd_loo   -341.0  6.5
p_loo         4.6  0.6
looic       682.1 13.1
------
Monte Carlo SE of elpd_loo is 0.0.

All Pareto k estimates are good (k < 0.5).
See help('pareto-k-diagnostic') for details.

Model comparisons:
elpd_diff se_diff
fit1   0.0       0.0
fit3   0.0       0.2
fit2 -27.1       6.4  

The monotonic model fits better than the continuous model, which is not surprising given that the relationship between income and ls is non-linear. The monotonic and the unordered factor model have almost identical fit in this example, but this may not be the case for other data sets.

## Setting Prior Distributions

In the previous monotonic model, we have implicitly assumed that all differences between adjacent categories were a-priori the same, or formulated correctly, had the same prior distribution. In the following, we want to show how to change this assumption. The canonical prior distribution of a simplex parameter is the Dirichlet distribution, a multivariate generalization of the beta distribution. It is non-zero for all valid simplexes (i.e., $$\zeta_i \in [0,1]$$ and $$\sum_{i = 1}^D \zeta_i = 1$$) and zero otherwise. The Dirichlet prior has a single parameter $$\alpha$$ of the same length as $$\zeta$$. The higher $$\alpha_i$$ the higher the a-priori probability of higher values of $$\zeta_i$$. Suppose that, before looking at the data, we expected that the same amount of additional money matters more for people who generally have less money. This translates into a higher a-priori values of $$\zeta_1$$ (difference between ‘below_20’ and ‘20_to_40’) and hence into higher values of $$\alpha_1$$. We choose $$\alpha_1 = 2$$ and $$\alpha_2 = \alpha_3 = 1$$, the latter being the default value of $$\alpha$$. To fit the model we write:

prior4 <- prior(dirichlet(c(2, 1, 1)), class = "simo", coef = "moincome1")
fit4 <- brm(ls ~ mo(income), data = dat,
prior = prior4, sample_prior = TRUE)

The 1 at the end of "moincome1" may appear strange when first working with monotonic effects. However, it is necessary as one monotonic term may be associated with multiple simplex parameters, if interactions of multiple monotonic variables are included in the model.

summary(fit4)
 Family: gaussian
Links: mu = identity; sigma = identity
Formula: ls ~ mo(income)
Data: dat (Number of observations: 100)
Draws: 4 chains, each with iter = 2000; warmup = 1000; thin = 1;
total post-warmup draws = 4000

Population-Level Effects:
Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
Intercept    30.03      1.34    27.44    32.64 1.00     2753     2504
moincome     15.60      0.62    14.42    16.81 1.00     2605     2654

Simplex Parameters:
Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
moincome1     0.63      0.04     0.55     0.70 1.00     3103     2535
moincome1     0.26      0.05     0.16     0.36 1.00     3991     2867
moincome1     0.12      0.04     0.03     0.20 1.00     3256     1928

Family Specific Parameters:
Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sigma     7.19      0.52     6.27     8.30 1.00     3358     2580

Draws were sampled using sampling(NUTS). For each parameter, Bulk_ESS
and Tail_ESS are effective sample size measures, and Rhat is the potential
scale reduction factor on split chains (at convergence, Rhat = 1).

We have used sample_prior = TRUE to also obtain draws from the prior distribution of simo_moincome1 so that we can visualized it.

plot(fit4, variable = "prior_simo", regex = TRUE, N = 3) As is visible in the plots, simo_moincome1 was a-priori on average twice as high as simo_moincome1 and simo_moincome1 as a result of setting $$\alpha_1$$ to 2.

dat$age <- rnorm(100, mean = 40, sd = 10) We are not only interested in the main effect of age but also in the interaction of income and age. Interactions with monotonic variables can be specified in the usual way using the * operator: fit5 <- brm(ls ~ mo(income)*age, data = dat) summary(fit5)  Family: gaussian Links: mu = identity; sigma = identity Formula: ls ~ mo(income) * age Data: dat (Number of observations: 100) Draws: 4 chains, each with iter = 2000; warmup = 1000; thin = 1; total post-warmup draws = 4000 Population-Level Effects: Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS Intercept 32.66 5.61 22.12 44.02 1.00 1020 1806 age -0.06 0.14 -0.34 0.20 1.00 1009 1838 moincome 14.02 2.54 8.87 18.90 1.00 865 1560 moincome:age 0.04 0.06 -0.08 0.17 1.00 884 1704 Simplex Parameters: Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS moincome1 0.63 0.07 0.49 0.79 1.00 1486 1487 moincome1 0.26 0.07 0.11 0.40 1.00 2199 1764 moincome1 0.11 0.06 0.01 0.23 1.00 2316 1481 moincome:age1 0.41 0.25 0.02 0.88 1.00 1749 2108 moincome:age1 0.30 0.22 0.01 0.82 1.00 2201 2309 moincome:age1 0.29 0.21 0.01 0.79 1.00 2436 2134 Family Specific Parameters: Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS sigma 7.26 0.53 6.32 8.41 1.00 2901 2318 Draws were sampled using sampling(NUTS). For each parameter, Bulk_ESS and Tail_ESS are effective sample size measures, and Rhat is the potential scale reduction factor on split chains (at convergence, Rhat = 1). conditional_effects(fit5, "income:age") ## Modelling Monotonic Group-Level Effects Suppose that the 100 people in our sample data were drawn from 10 different cities; 10 people per city. Thus, we add an identifier for city to the data and add some city-related variation to ls. dat$city <- rep(1:10, each = 10)
var_city <- rnorm(10, sd = 10)
dat$ls <- dat$ls + var_city[dat\$city]

With the following code, we fit a multilevel model assuming the intercept and the effect of income to vary by city:

fit6 <- brm(ls ~ mo(income)*age + (mo(income) | city), data = dat)
summary(fit6)
 Family: gaussian
Links: mu = identity; sigma = identity
Formula: ls ~ mo(income) * age + (mo(income) | city)
Data: dat (Number of observations: 100)
Draws: 4 chains, each with iter = 2000; warmup = 1000; thin = 1;
total post-warmup draws = 4000

Group-Level Effects:
~city (Number of levels: 10)
Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)              13.81      4.12     8.11    24.12 1.00     1472     2288
sd(moincome)                0.79      0.68     0.03     2.54 1.00     2661     2578
cor(Intercept,moincome)    -0.15      0.56    -0.97     0.91 1.00     4806     2379

Population-Level Effects:
Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
Intercept       34.43      7.80    19.49    50.48 1.00     1362     2136
age             -0.10      0.15    -0.42     0.18 1.00     1879     1939
moincome        13.14      2.88     7.29    18.58 1.00     1563     1876
moincome:age     0.06      0.07    -0.08     0.20 1.00     1566     1866

Simplex Parameters:
Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
moincome1         0.62      0.09     0.42     0.81 1.00     1824     1609
moincome1         0.25      0.09     0.06     0.44 1.00     2238     1490
moincome1         0.12      0.07     0.01     0.27 1.00     2537     1800
moincome:age1     0.41      0.24     0.02     0.87 1.00     3161     2725
moincome:age1     0.30      0.22     0.01     0.79 1.00     3589     3112
moincome:age1     0.30      0.22     0.01     0.80 1.00     3776     3005

Family Specific Parameters:
Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sigma     7.37      0.58     6.33     8.59 1.00     5177     2950

Draws were sampled using sampling(NUTS). For each parameter, Bulk_ESS
and Tail_ESS are effective sample size measures, and Rhat is the potential
scale reduction factor on split chains (at convergence, Rhat = 1).

reveals that the effect of income varies only little across cities. For the present data, this is not overly surprising given that, in the data simulations, we assumed income to have the same effect across cities.