manuscript.Rmd 35.2 KB
Newer Older
1
---
linushof's avatar
linushof committed
2
3
title: 'Sampling Strategies in Decisions from Experience'
author: "Linus Hof, Thorsten Pachur, Veronika Zilker"
4
5
6
7
8
9
bibliography: sampling-strategies-in-dfe.bib
output:
  html_document:
    code_folding: hide
    toc: yes
    toc_float: yes
linushof's avatar
linushof committed
10
    number_sections: no
linushof's avatar
linushof committed
11
12
13
  pdf_document:
    toc: yes
csl: apa.csl
linushof's avatar
linushof committed
14
15
16
editor_options: 
  markdown: 
    wrap: sentence
17
18
---

19
20
```{r}
# load packages
linushof's avatar
linushof committed
21
22
pacman::p_load(repro,
               tidyverse,
linushof's avatar
linushof committed
23
24
               knitr, 
               viridis)
25
26
```

27
# Author Note
28

29
This document was created from the commit with the hash `r repro::current_hash()`. 
30

31
32
- Add information on how to reproduce the project.
- Add contact.
linushof's avatar
linushof committed
33

34
# Abstract
linushof's avatar
linushof committed
35

36
37
A probability theoretic definition of prospects and a rough stochastic model for decisions from experience is proposed.
It is demonstrated how the model can be used a) to explicate assumptions about the sampling and decision strategies that agents may apply and b) to derive function forms and parameter values that describe the resulting decision behavior. 
38
Synthetic choice data is simulated and modeled in cumulative prospect theory to test these predictions. 
39

linushof's avatar
linushof committed
40
# Introduction
41

42
43
...

44
## Sampling in Decisions from Experience 
linushof's avatar
linushof committed
45

linushof's avatar
linushof committed
46
In research on the decision theory, a standard paradigm is the choice between at least two (monetary) prospects.
47
48
Let a prospect be a probability space $(\Omega, \mathscr{F}, P)$.
$\Omega$ is the sample space 
linushof's avatar
linushof committed
49

50
$$\begin{equation}
linushof's avatar
linushof committed
51
\Omega = \{\omega_1, ..., \omega_n\}
52
\end{equation}$$ 
linushof's avatar
linushof committed
53

linushof's avatar
linushof committed
54
containing a finite set of possible outcomes $\omega$, monetary gains and/or losses respectively. 
55
$\mathscr{F}$ is the set of all possible subsets of $\Omega$: 
linushof's avatar
linushof committed
56

57
$$\begin{equation}
linushof's avatar
linushof committed
58
\mathscr{F} = \{A_1, A_2, ...\} = \mathscr{P}(\Omega) 
59
\; .
60
\end{equation}$$
linushof's avatar
linushof committed
61

62
$P$ is a probability mass function  
63
64

$$\begin{equation}
65
P: \mathscr{F} \mapsto [0,1] 
66
67
\end{equation}$$

linushof's avatar
linushof committed
68
that assigns each outcome $\omega$ a probability $0 < p(\omega) \leq 1$ with $P(\Omega) = 1$ [ @kolmogorovFoundationsTheoryProbability1950, pp. 2-3].
69

70
In such a choice paradigm, agents are asked to evaluate the prospects and build a preference for either one of them. 
linushof's avatar
linushof committed
71
It is common to make a distinction between two variants of this evaluation process [cf. @hertwigDescriptionexperienceGapRisky2009]. 
72
For decisions from description (DfD), agents are provided a full symbolic description of the prospects.
linushof's avatar
linushof committed
73
74
75
For decisions from experience [DfE; e.g., @hertwigDecisionsExperienceEffect2004], prospects are not described but must be explored by the means of sampling. 

To provide a formal definition of sampling in risky choice, we make use of the mathematical concept of a random variable and start by referring to a prospect as *"risky"* in the case where $p(\omega) \neq 1$ for all $\omega \in \Omega$.
76
Here, risky describes the fact that if agents would choose a prospect and any of its outcomes in $\Omega$ must occur, none of these outcomes will occur with certainty. 
linushof's avatar
linushof committed
77
It is acceptable to speak of the occurrence of $\omega$ as a realization of a random variable $X$ defined on a prospect iff the following conditions (1) and (2) are met: 
78

linushof's avatar
linushof committed
79
(1) $X$ is a measurable function $$\begin{equation} X: (\Omega, \mathscr{F})  \mapsto (\Omega', \mathscr{F'}) \; , \end{equation}$$ where $\Omega'$ is a set of real numbered values $X$ can take and $\mathscr{F'}$ is a set of subsets of $\Omega'$. I.e., $\Omega$ maps into $\Omega'$ such that correspondingly each subset $A' \in \mathscr{F'}$ has a pre-image $X^{-1}A' \in \mathscr{F}$, which is the set $\{\omega \in \Omega: X(\omega) \in A'\}$ [@kolmogorovFoundationsTheoryProbability1950, p. 21].
80

linushof's avatar
linushof committed
81
(2) The mapping is such that $X(\omega) = x \equiv \omega$. 
82

linushof's avatar
linushof committed
83
In (2), $x \equiv \omega$ means that the realization of a random variable $X(\omega) = x$ is numerically equivalent to its pre-image $\omega$.  
84
85
Given conditions (1) and (2), we denote any observation of $\omega$ as a *"single sample"*, or realization, of a random variable defined on a prospect and the act of generating a sequence of single samples in discrete time as *"sequential sampling"*. 
Note that, since random variables defined on the same prospect are independent and identically distributed (iid), the weak law of the large number applies to the relative frequency of occurrence of an outcome $\omega$ in a sequence of single samples originating from the same prospect [cf. @bernoulliArsConjectandiOpus1713].
linushof's avatar
linushof committed
86
Thus, long sample sequences in principle allow to obtain the same information about a prospect by sampling as by symbolic description.
87

linushof's avatar
linushof committed
88
Consider now a choice between prospects $1, ..., k$.
89
To construct a stochastic model for DfE, we assume that agents base their decision on the information related to these prospects and define a decision variable as a function of the latter:
90
91

$$\begin{equation}
linushof's avatar
linushof committed
92
93
D:= f((\Omega, \mathscr{F}, P)_1, ..., (\Omega, \mathscr{F}, P)_k)
\;.
94
95
\end{equation}$$

96
Now, since in DfE no symbolic descriptions of the prospects are provided, the model must be restricted to the case where decisions are based on sequences of single samples originating from the respective prospects:
97

98
$$\begin{equation}
99
D := f(X_{i1}, ..., X_{ik}) 
100
\; ,
101
102
\end{equation}$$

103
where $i = 1, ..., N$ denotes a sequence of length $N$ of random variables that are iid.  
104

105
106
Concerning the form of $f$ and the measures it utilizes, it is quite proper to say that they reflect our assumptions about the exact kind of information agents process and the way they do and that these choices should be informed by psychological theory and empirical protocols. 
In the following section, it is demonstrated how a stochastic model can be used to explicate assumptions about the sampling and decision strategies that agents may apply in DfE.  
107

108
## A Stochastic Model Capturing Differences in Sampling and Decision Strategies  
109

linushof's avatar
linushof committed
110
111
Hills and Hertwig [-@hillsInformationSearchDecisions2010] discussed a potential link between sampling and decision strategies in DfE. 
Specifically, the authors suppose that if single samples originating from different prospects are generated in direct succession (piecewise sampling), the evaluation of prospects is based on multiple ordinal comparisons of single samples (round-wise decisions).
112
113
114
115
116
117
118
119
120
121
122
123
124
125
In contrast, if single samples originating from the same prospect are generated in direct succession (comprehensive sampling), it is supposed that the evaluation of prospects is based on a single ordinal comparison of long sequences of single samples (summary decisions) [@hillsInformationSearchDecisions2010, Figure 1 for a graphical summary]. 
We now consider choices between two prospects to build a stochastic model that captures these assumptions.

Let $X$ and $Y$ be random variables defined on the prospects $(\Omega, \mathscr{F}, P)_X$ and $(\Omega, \mathscr{F}, P)_Y$. 

Hills and Hertwig [-@hillsInformationSearchDecisions2010] suggest that any two sequences $X_i$ and $Y_i$ are compared by their means. 
Let $C = \mathbb{R}$ be the set of all possible outcomes of such a mean comparison for given lengths $N_X$ and $N_Y$ of the sample sequences and 

$$\begin{equation}
\mathscr{C} = \left\{ \{c \in C: \overline{X}_{N_X} - \overline{Y}_{N_Y} > 0\}, \{ c \in C: \overline{X}_{N_X} - \overline{Y}_{N_Y} \leq 0\}
\right\}
\end{equation}$$

be a set of subsets of $C$, indicating that comparisons of prospects on the ordinal (rather than on the metric) scale are of primary interest. 
126

linushof's avatar
linushof committed
127
By definition, the decision variable $D$ should quantify the accumulated evidence for one prospect over the other, which is described in units of won comparisons.
128

129
Hence, $f$ should map the possible outcomes of a comparison of quantitative measures related to $X$ and $Y$, hereafter the sampling space $S = \mathbb{R}$, to a measure space $S' = \{0,1\}$, indicating the possible outcomes of a single comparison:
130

131
132
133
134
135
136
$$\begin{equation}
D:= f: S \mapsto S' 
\; .
\end{equation}$$

Since Hills and Hertwig [-@hillsInformationSearchDecisions2010] assume that comparisons of prospects are based on sample means, $S$ is the set
137
138

$$\begin{equation}
139
140
141
S = 
  \left\{
    \frac{\frac{1}{N_X} \sum\limits_{i=1}^{N_X} x_i}
linushof's avatar
linushof committed
142
    {\frac{1}{N_Y} \sum\limits_{i=1}^{N_Y} y_i} 
143
  \right\} 
144
145
  = 
  \left\{
linushof's avatar
linushof committed
146
    \frac{\overline{X}_{N_X}} {\overline{Y}_{N_Y}}
147
  \right\}
148
  \; ,
149
150
\end{equation}$$

linushof's avatar
linushof committed
151
where $N_X$ and $N_Y$ denotes the number of single samples within a comparison. 
152
To indicate that the comparison of prospects on the ordinal scale is of primary interest, we define 
153

154
$$\begin{equation}
linushof's avatar
linushof committed
155
\mathscr{D} = \left\{\frac{\overline{X}_{N_X}}{\overline{Y}_{N_Y}} > 0, \frac{\overline{X}_{N_X}}{\overline{Y}_{N_Y}} \leq 0 \right\} 
156
\end{equation}$$
157

158
as a set of subsets of $S$, i.e., the *event space*, and the decision variable as the measure 
159

160
$$\begin{equation}
161
D:= f: (S, \mathscr{D}) \mapsto S'
162
\end{equation}$$
linushof's avatar
linushof committed
163

164
with the mapping
165

166
$$\begin{equation}
167
168
D:= 
  \left(
linushof's avatar
linushof committed
169
    \frac{\overline{X}_{N_X}} {\overline{Y}_{N_Y}}
170
171
172
173
  \right) 
  \in S : 
  f
  \left(
linushof's avatar
linushof committed
174
    \frac{\overline{X}_{N_X}} {\overline{Y}_{N_Y}}
175
176
  \right) 
  =
177
  \begin{cases}
linushof's avatar
linushof committed
178
179
    1 & \text{if} & \frac{\overline{X}_{N_X}}{\overline{Y}_{N_Y}} > 0 \in \mathscr{D} \\
    0 & \text{else}. 
180
  \end{cases} 
181
\end{equation}$$
182

183
184
185
186
187
188
189
190
It can be shown that for the case $N_X = N_Y = 1$, $D$ is a random variable that follows the Bernoulli distribution

$$\begin{equation}
D \sim B\left( p \left( \frac{\overline{X}_{N_X = 1}}{\overline{Y}_{N_Y = 1}} > 0\right), n\right) \; ,
\end{equation}$$

where $n$ is the number of comparisons (see *Proof 1* in Appendix). 

191
## Predicting Choice Behavior in DfE 
linushof's avatar
linushof committed
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222

Hills and Hertwig [-@hillsInformationSearchDecisions2010] proposed the two different sampling strategies in combination with the respective decision strategies, i.e., piecewise sampling and round-wise comparison vs. comprehensive sampling and summary comparison, as an explanation for different choice patterns in DfE. 
How does the current version of the SSM support this proposition? 
Given prospects $X$ and $Y$, the sample spaces $S = \left\{\frac{\overline{X}_{N_X}} {\overline{Y}_{N_Y}}\right\}^{\mathbb{N}}$ can be varied by changes on three parameters, i.e., the number of comparisons $\mathbb{N}$ and the sample sizes $N_X$ and $N_Y$ on which these comparisons are based.
First, only considering the pure cases formulated by the above authors, the following restrictions are put the parameters:

$$\begin{equation}
\mathbb{N}
  =
  \begin{cases}
    1 & \text{if} & \text{Summary} \\ 
    \geq 1 & \text{if} & \text{Round-wise}  
  \end{cases}\\
\end{equation}$$
 
and 

$$\begin{equation}
N_X \, \text{and} \, N_Y
  =
  \begin{cases}
    \geq 1 & \text{if} & \text{Summary} \\ 
    1 & \text{if} & \text{Round-wise}  
  \end{cases}\\
\end{equation}$$

For the summary strategy, the following prediction is obtained: 

Given that 

$$\begin{equation}
223
224
P\left(\lim_{N_X \to \infty} \overline{X}_{N_X} = E(X) \right) = 
P\left(\lim_{N_Y \to \infty} \overline{Y}_{N_Y} = E(Y) \right) = 
linushof's avatar
linushof committed
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
1 \; ,
\end{equation}$$

we obtain that

$$\begin{equation}
\left(
    \frac{\overline{X}_{N_X}} {\overline{Y}_{N_Y}}
\right)
\in S : 
P\left(\lim_{N_X \to \infty} \lim_{N_Y \to \infty} \frac{\overline{X}_{N_X}} {\overline{Y}_{N_Y}} = 
\frac{E(X)} {E(Y)}
\right ) =
1 \; ,
\end{equation}$$

I.e., for the summary strategy, we assume that for increasing sample sizes $N_X$ and $N_Y$, the prospect with the larger expected value is chosen almost surely.  

For the round-wise strategy, the following prediction is obtained:

245
Given that $N_X$ and $N_Y$ are set to 1, $D$ follows the binomial distribution 
linushof's avatar
linushof committed
246

247
248
249
250
251
252
253
254
255
256
257
258
259
260
$$\begin{equation}
B(k \, | \, p_X, \mathbb{N}) \; ,
\end{equation}$$

where $p$ is the probability that a single sample of prospect $X$ is larger than a single sample of prospect $Y$, $\mathbb{N}$ is the number of comparisons and $k$ is the number of times $x \in X$ is larger than $y \in Y$. 

*Proof.*

For $N_X = N_Y = 1$, the sample space is 

$$\begin{equation}
\left\{\frac{\overline{X}_{N_X = 1}} {\overline{Y}_{N_Y = 1}} \right\}^{\mathbb{N}} =
\left\{\frac{x_i \in \Omega'_X} {y_j \in \Omega'_Y}\right\}^{\mathbb{N}}
\end{equation}$$
261

linushof's avatar
linushof committed
262
263
# Method

linushof's avatar
linushof committed
264
## Test set
265

linushof's avatar
linushof committed
266
267
268
269
270
271
272
273
274
Under each condition, i.e., strategy-parameter combinations, all gambles are played by 100 synthetic agents.
We test a set of gambles, in which one of the prospects contains a safe outcome and the other two risky outcomes (*safe-risky gambles*).
Therefore, 60 gambles from an initial set of 10,000 are sampled.
Both outcomes and probabilities are drawn from uniform distributions, ranging from 0 to 20 for outcomes and from .01 to .99 for probabilities of the lower risky outcomes $p_{low}$.
The probabilities of the higher risky outcomes are $1-p_{low}$, respectively.
To omit dominant prospects, safe outcomes fall between both risky outcomes.
The table below contains the test set of 60 gambles.
Sampling of gambles was stratified, randomly drawing an equal number of 20 gambles with no, an attractive, and an unattractive rare outcome.
Risky outcomes are considered *"rare"* if their probability is $p < .2$ and *"attractive"* (*"unattractive"*) if they are higher (lower) than the safe outcome.
275

linushof's avatar
linushof committed
276
277
278
```{r message=FALSE}
gambles <- read_csv("data/gambles/sr_subset.csv")
gambles %>% kable()
279
280
```

linushof's avatar
linushof committed
281
## Model Parameters
282

linushof's avatar
linushof committed
283
**Switching probability** $s$ is the probability with which agents draw the following single sample from the prospect they did not get their most recent single sample from.
linushof's avatar
linushof committed
284
$s$ is varied between .1 to 1 in increments of .1.
285

linushof's avatar
linushof committed
286
The **boundary type** is either the minimum value any prospect's sample statistic must reach (absolute) or the minimum value for the difference of these statistics (relative).
linushof's avatar
linushof committed
287
Sample statistics are sums over outcomes (comprehensive strategy) and sums over wins (piecewise strategy), respectively.
288

linushof's avatar
linushof committed
289
290
For comprehensive integration, the **boundary value** $a$ is varied between 15 to 75 in increments of 15.
For piecewise integration $a$ is varied between 1 to 5 in increments of 1.
291

linushof's avatar
linushof committed
292
```{r message=FALSE}
293
294
295
296
297
298
299
300
# read choice data 
cols <- list(.default = col_double(),
             strategy = col_factor(),
             boundary = col_factor(),
             gamble = col_factor(),
             rare = col_factor(),
             agent = col_factor(),
             choice = col_factor())
linushof's avatar
linushof committed
301
choices <- read_csv("data/choices/choices.csv", col_types = cols)
302
303
```

linushof's avatar
linushof committed
304
In sum, 2 (strategies) x 60 (gambles) x 100 (agents) x 100 (parameter combinations) = `r nrow(choices)` choices are simulated.
linushof's avatar
linushof committed
305

linushof's avatar
linushof committed
306
# Results
307

linushof's avatar
linushof committed
308
309
Because we are not interested in deviations from normative choice due to sampling artifacts (e.g., ceiling effects produced by low boundaries), we remove trials in which only one prospect was attended.
In addition, we use relative frequencies of sampled outcomes rather than 'a priori' probabilities to compare actual against normative choice behavior.
310
311

```{r}
linushof's avatar
linushof committed
312
313
314
# remove choices where prospects were not attended
choices <- choices %>%
  filter(!(is.na(a_ev_exp) | is.na(b_ev_exp)))
315
316
```

linushof's avatar
linushof committed
317
318
319
320
321
```{r eval = FALSE}
# remove choices where not all outcomes were sampled
choices <- choices %>% 
  filter(!(is.na(a_ev_exp) | is.na(b_ev_exp) | a_p1_exp == 0 | a_p2_exp == 0))
```
linushof's avatar
linushof committed
322

linushof's avatar
linushof committed
323
Removing the respective trials, we are left with `r nrow(choices)` choices.
linushof's avatar
linushof committed
324

linushof's avatar
linushof committed
325
## Sample Size
linushof's avatar
linushof committed
326

linushof's avatar
linushof committed
327
328
329
330
331
332
```{r message=FALSE}
samples <- choices %>% 
  group_by(strategy, s, boundary, a) %>% 
  summarise(n_med = median(n_sample))
samples_piecewise <- samples %>% filter(strategy == "piecewise")
samples_comprehensive <- samples %>% filter(strategy == "comprehensive")
333
334
```

linushof's avatar
linushof committed
335
The median sample sizes generated by different parameter combinations ranged from `r min(samples_piecewise$n_med)` to `r max(samples_piecewise$n_med)` for piecewise integration and `r min(samples_comprehensive$n_med)` to `r max(samples_comprehensive$n_med)` for comprehensive integration.
336

linushof's avatar
linushof committed
337
### Boundary type and boundary value (a)
338

linushof's avatar
linushof committed
339
As evidence is accumulated sequentially, relative boundaries and large boundary values naturally lead to larger sample sizes, irrespective of the integration strategy.
linushof's avatar
linushof committed
340

linushof's avatar
linushof committed
341
342
```{r message=FALSE}
group_med <- samples_piecewise %>%
linushof's avatar
linushof committed
343
  group_by(boundary, a) %>% 
linushof's avatar
linushof committed
344
  summarise(group_med = median(n_med)) # to get the median across all s values
linushof's avatar
linushof committed
345

linushof's avatar
linushof committed
346
347
samples_piecewise %>%
  ggplot(aes(a, n_med, color = a)) + 
linushof's avatar
linushof committed
348
  geom_jitter(alpha = .5, size = 2) +
linushof's avatar
linushof committed
349
350
351
  geom_point(data = group_med, aes(y = group_med), size = 3) +
  facet_wrap(~boundary) + 
  scale_color_viridis() + 
352
  labs(title = "Piecewise Integration",
linushof's avatar
linushof committed
353
       x ="a", 
linushof's avatar
linushof committed
354
       y="Sample Size", 
linushof's avatar
linushof committed
355
       col="a") + 
linushof's avatar
linushof committed
356
  theme_minimal()
linushof's avatar
linushof committed
357
```
linushof's avatar
linushof committed
358

linushof's avatar
linushof committed
359
360
```{r message=FALSE}
group_med <- samples_comprehensive %>%
linushof's avatar
linushof committed
361
  group_by(boundary, a) %>% 
linushof's avatar
linushof committed
362
  summarise(group_med = median(n_med)) 
linushof's avatar
linushof committed
363

linushof's avatar
linushof committed
364
365
samples_comprehensive %>%
  ggplot(aes(a, n_med, color = a)) + 
linushof's avatar
linushof committed
366
  geom_jitter(alpha = .5, size = 2) +
linushof's avatar
linushof committed
367
368
369
  geom_point(data = group_med, aes(y = group_med), size = 3) +
  facet_wrap(~boundary) + 
  scale_color_viridis() + 
370
  labs(title = "Comprehensive Integration",
linushof's avatar
linushof committed
371
       x ="a", 
linushof's avatar
linushof committed
372
       y="Sample Size", 
linushof's avatar
linushof committed
373
       col="a") + 
linushof's avatar
linushof committed
374
  theme_minimal()
375
376
```

linushof's avatar
linushof committed
377
### Switching probability (s)
378

linushof's avatar
linushof committed
379
380
381
For piecewise integration, there is an inverse relationship between switching probability and sample size.
I.e., the lower s, the less frequent prospects are compared and thus, boundaries are only approached with larger sample sizes.
This effect is particularly pronounced for low probabilities such that the increase in sample size accelerates as switching probability decreases.
linushof's avatar
linushof committed
382

linushof's avatar
linushof committed
383
384
```{r message=FALSE}
group_med <- samples_piecewise %>%
linushof's avatar
linushof committed
385
  group_by(boundary, s) %>% 
linushof's avatar
linushof committed
386
  summarise(group_med = median(n_med)) # to get the median across all a values
linushof's avatar
linushof committed
387

linushof's avatar
linushof committed
388
389
390
391
392
393
samples_piecewise %>%
  ggplot(aes(s, n_med, color = s)) + 
  geom_jitter(alpha = .5, size = 2) +
  geom_point(data = group_med, aes(y = group_med), size = 3) +
  facet_wrap(~boundary) + 
  scale_color_viridis() + 
394
  labs(title = "Piecewise Integration",
linushof's avatar
linushof committed
395
       x ="s", 
linushof's avatar
linushof committed
396
       y="Sample Size", 
linushof's avatar
linushof committed
397
       col="s") + 
linushof's avatar
linushof committed
398
399
400
  theme_minimal()
```

linushof's avatar
linushof committed
401
402
403
For comprehensive integration, boundary types differ in the effects of switching probability.
For absolute boundaries, switching probability has no apparent effect on sample size as the distance of a given prospect to its absolute boundary is not changed by switching to (and sampling from) the other prospect.
For relative boundaries, however, samples sizes increase with switching probability.
linushof's avatar
linushof committed
404

linushof's avatar
linushof committed
405
406
```{r message=FALSE}
group_med <- samples_comprehensive %>%
linushof's avatar
linushof committed
407
  group_by(boundary, s) %>% 
linushof's avatar
linushof committed
408
  summarise(group_med = median(n_med)) # to get the median across all a values
linushof's avatar
linushof committed
409

linushof's avatar
linushof committed
410
411
412
413
414
415
samples_comprehensive %>%
  ggplot(aes(s, n_med, color = s)) + 
  geom_jitter(alpha = .5, size = 2) +
  geom_point(data = group_med, aes(y = group_med), size = 3) +
  facet_wrap(~boundary) + 
  scale_color_viridis() + 
416
  labs(title = "Comprehensive Integration",
linushof's avatar
linushof committed
417
418
419
       x ="s",
       y = "Sample Size", 
       col="s") + 
linushof's avatar
linushof committed
420
421
422
  theme_minimal()
```

linushof's avatar
linushof committed
423
## Choice Behavior
linushof's avatar
linushof committed
424

linushof's avatar
linushof committed
425
Below, in extension to Hills and Hertwig [-@hillsInformationSearchDecisions2010], the interplay of integration strategies, gamble features, and model parameters in their effects on choice behavior in general and their contribution to underweighting of rare events in particular is investigated.
linushof's avatar
linushof committed
426
427
428
429
430
431
432
433
434
435
436
437
We apply two definitions of underweighting of rare events: Considering false response rates, we define underweighting such that the rarity of an attractive (unattractive) outcome leads to choose the safe (risky) prospect although the risky (safe) prospect has a higher expected value.

```{r message=FALSE}
fr_rates <- choices %>% 
  mutate(ev_ratio_exp = round(a_ev_exp/b_ev_exp, 2), 
         norm = case_when(ev_ratio_exp > 1 ~ "A", ev_ratio_exp < 1 ~ "B")) %>% 
  filter(!is.na(norm)) %>% # exclude trials with normative indifferent options
  group_by(strategy, s, boundary, a, rare, norm, choice) %>% # group correct and incorrect responses
  summarise(n = n()) %>% # absolute numbers 
  mutate(rate = round(n/sum(n), 2), # response rates 
         type = case_when(norm == "A" & choice == "B" ~ "false safe", norm == "B" & choice == "A" ~ "false risky")) %>% 
  filter(!is.na(type)) # remove correct responses
linushof's avatar
linushof committed
438
439
```

linushof's avatar
linushof committed
440
Considering the parameters of Prelec's [-@prelecProbabilityWeightingFunction1998] implementation of the weighting function [CPT; cf. @tverskyAdvancesProspectTheory1992], underweighting is reflected by decisions weights estimated to be smaller than the corresponding objective probabilities.
linushof's avatar
linushof committed
441

linushof's avatar
linushof committed
442
### False Response Rates
linushof's avatar
linushof committed
443

linushof's avatar
linushof committed
444
445
446
```{r message=FALSE}
fr_rates_piecewise <- fr_rates %>% filter(strategy == "piecewise")
fr_rates_comprehensive <- fr_rates %>% filter(strategy == "comprehensive")
linushof's avatar
linushof committed
447
```
448

linushof's avatar
linushof committed
449
The false response rates generated by different parameter combinations ranged from `r min(fr_rates_piecewise$rate)` to `r max(fr_rates_piecewise$rate)` for piecewise integration and from `r min(fr_rates_comprehensive$rate)` to `r max(fr_rates_comprehensive$rate)` for comprehensive integration.
linushof's avatar
linushof committed
450
However, false response rates vary considerably as a function of rare events, indicating that their presence and attractiveness are large determinants of false response rates.
linushof's avatar
linushof committed
451

linushof's avatar
linushof committed
452
453
454
455
456
457
```{r message=FALSE}
fr_rates %>% 
  group_by(strategy, boundary, rare) %>% 
  summarise(min = min(rate),
            max = max(rate)) %>% 
  kable()
linushof's avatar
linushof committed
458
459
```

linushof's avatar
linushof committed
460
The heatmaps below show the false response rates for all strategy-parameter combinations.
linushof's avatar
linushof committed
461
462
Consistent with our - somewhat rough - definition of underweighting, the rate of false risky responses is generally higher, if the unattractive outcome of the risky prospect is rare (top panel).
Conversely, if the attractive outcome of the risky prospect is rare, the rate of false safe responses is generally higher (bottom panel).
linushof's avatar
linushof committed
463
As indicated by the larger range of false response rates, the effects of rare events are considerably larger for piecewise integration.
464

linushof's avatar
linushof committed
465
466
467
468
469
470
471
472
473
474
475
476
477
478
```{r message=FALSE}
fr_rates %>% 
  filter(strategy == "piecewise", boundary == "absolute") %>% 
  ggplot(aes(a, s, fill = rate)) + 
  facet_grid(type ~ fct_relevel(rare, "attractive", "none", "unattractive"), switch = "y") +
  geom_tile(colour="white", size=0.25) + 
  scale_x_continuous(expand=c(0,0), breaks = seq(1, 5, 1)) +
  scale_y_continuous(expand=c(0,0), breaks = seq(.1, 1, .1)) +
  scale_fill_viridis() + 
  labs(title = "Piecewise Integration | Absolute Boundary",
       x = "a", 
       y= "s", 
       fill = "% False Responses") + 
  theme_minimal() 
479
480
```

linushof's avatar
linushof committed
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
```{r message=FALSE}
fr_rates %>% 
  filter(strategy == "piecewise", boundary == "relative") %>% 
  ggplot(aes(a, s, fill = rate)) + 
  facet_grid(type ~ fct_relevel(rare, "attractive", "none", "unattractive"), switch = "y") +
  geom_tile(colour="white", size=0.25) + 
  scale_x_continuous(expand=c(0,0), breaks = seq(1, 5, 1)) +
  scale_y_continuous(expand=c(0,0), breaks = seq(.1, 1, .1)) +
  scale_fill_viridis() + 
  labs(title = "Piecewise Integration | Relative Boundary",
       x = "a", 
       y= "s", 
       fill = "% False Responses") + 
  theme_minimal() 
```
linushof's avatar
linushof committed
496

linushof's avatar
linushof committed
497
498
```{r message=FALSE}
fr_rates %>% 
linushof's avatar
linushof committed
499
  filter(strategy == "comprehensive", boundary == "absolute") %>% 
linushof's avatar
linushof committed
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
  ggplot(aes(a, s, fill = rate)) + 
  facet_grid(type ~ fct_relevel(rare, "attractive", "none", "unattractive"), switch = "y") +
  geom_tile(colour="white", size=0.25) + 
  scale_x_continuous(expand=c(0,0), breaks = seq(15, 75, 15)) +
  scale_y_continuous(expand=c(0,0), breaks = seq(.1, 1, .1)) +
  scale_fill_viridis() + 
  labs(title = "Comprehensive Integration | Absolute Boundary",
       x = "a", 
       y= "s", 
       fill = "% False Responses") + 
  theme_minimal() 
```

```{r message=FALSE}
fr_rates %>% 
linushof's avatar
linushof committed
515
  filter(strategy == "comprehensive", boundary == "relative") %>% 
linushof's avatar
linushof committed
516
517
518
519
520
521
522
523
524
525
526
  ggplot(aes(a, s, fill = rate)) + 
  facet_grid(type ~ fct_relevel(rare, "attractive", "none", "unattractive"), switch = "y") +
  geom_tile(colour="white", size=0.25) + 
  scale_x_continuous(expand=c(0,0), breaks = seq(15, 75, 15)) +
  scale_y_continuous(expand=c(0,0), breaks = seq(.1, 1, .1)) +
  scale_fill_viridis() + 
  labs(title = "Comprehensive Integration | Relative Boundary",
       x = "a", 
       y= "s", 
       fill = "% False Responses") + 
  theme_minimal() 
527
528
```

linushof's avatar
linushof committed
529
#### Switching Probability (s) and Boundary Value (a)
linushof's avatar
linushof committed
530

linushof's avatar
linushof committed
531
As for both piecewise and comprehensive integration the differences between boundary types are rather minor and of magnitude than of qualitative pattern, the remaining analyses of false response rates are summarized across absolute and relative boundaries.
linushof's avatar
linushof committed
532

linushof's avatar
linushof committed
533
Below, the $s$ and $a$ parameter are considered as additional sources of variation in the false response pattern above and beyond the interplay of integration strategies and the rarity and attractiveness of outcomes.
linushof's avatar
linushof committed
534

linushof's avatar
linushof committed
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
```{r message=FALSE}
fr_rates %>% 
  filter(strategy == "piecewise") %>% 
  ggplot(aes(s, rate, color = a)) + 
  facet_grid(type ~ fct_relevel(rare, "attractive", "none", "unattractive"), switch = "y") +
  geom_jitter(size = 2) + 
  scale_x_continuous(breaks = seq(0, 1, .1)) +
  scale_y_continuous(breaks = seq(0, 1, .1)) +
  scale_color_viridis() + 
  labs(title = "Piecewise Integration",
       x = "s", 
       y= "% False Responses", 
       color = "a") + 
  theme_minimal() 
```
550

linushof's avatar
linushof committed
551
552
```{r message=FALSE}
fr_rates %>% 
linushof's avatar
linushof committed
553
  filter(strategy == "comprehensive") %>% 
linushof's avatar
linushof committed
554
555
556
557
558
559
  ggplot(aes(s, rate, color = a)) + 
  facet_grid(type ~ fct_relevel(rare, "attractive", "none", "unattractive"), switch = "y") +
  geom_jitter(size = 2) + 
  scale_x_continuous(breaks = seq(0, 1, .1)) +
  scale_y_continuous(breaks = seq(0, 1, .1)) +
  scale_color_viridis() + 
560
  labs(title = "Comprehensive Integration",
linushof's avatar
linushof committed
561
562
563
564
       x = "s", 
       y= "% False Responses", 
       color = "a") + 
  theme_minimal() 
565
566
```

linushof's avatar
linushof committed
567
For piecewise integration, switching probability is naturally related to the size of the samples on which the round-wise comparisons of prospects are based on, with low values of $s$ indicating large samples and vice versa.
linushof's avatar
linushof committed
568
Accordingly, switching probability is positively related to false response rates.
linushof's avatar
linushof committed
569
570
I.e., the larger the switching probability, the smaller the round-wise sample size and the probability of experiencing a rare event within a given round.
Because round-wise comparisons are independent of each other and binomial distributions within a given round are skewed for small samples and outcome probabilities [@kolmogorovFoundationsTheoryProbability1950], increasing boundary values do not reverse but rather amplify this relation.
571

linushof's avatar
linushof committed
572
573
574
For comprehensive integration, switching probability is negatively related to false response rates, i.e., an increase in $s$ is associated with decreasing false response rates.
This relation, however, may be the result of an artificial interaction between the $s$ and $a$ parameter.
Precisely, in the current algorithmic implementation of sampling with a comprehensive integration mechanism, decreasing switching probabilities cause comparisons of prospects based on increasingly unequal sample sizes immediately after switching prospects.
linushof's avatar
linushof committed
575
Consequentially, reaching (low) boundaries is rather a function of switching probability and associated sample sizes than of actual evidence for a given prospect over the other.
576

linushof's avatar
linushof committed
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
### Cumulative Prospect Theory

In the following, we examine the possible relations between the parameters of the *choice-generating* sampling models and the *choice-describing* cumulative prospect theory.

For each distinct strategy-parameter combination, we ran 20 chains of 40,000 iterations each, after a warm-up period of 1000 samples.
To reduce potential autocorrelation during the sampling process, we only kept every 20th sample (thinning).

```{r}
# read CPT data
cols <- list(.default = col_double(),
             strategy = col_factor(),
             boundary = col_factor(),
             parameter = col_factor())
estimates <- read_csv("data/estimates/estimates_cpt_pooled.csv", col_types = cols)
```

#### Convergence

```{r}
gel_92 <- max(estimates$Rhat) # get largest scale reduction factor (Gelman & Rubin, 1992) 
```

The potential scale reduction factor $\hat{R}$ was $n \leq$ `r round(gel_92, 3)` for all estimates, indicating good convergence.

601
#### Piecewise Integration
linushof's avatar
linushof committed
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635

```{r}
# generate subset of all strategy-parameter combinations (rows) and their parameters (columns)
curves_cpt <- estimates %>% 
  select(strategy, s, boundary, a, parameter, mean) %>% 
  pivot_wider(names_from = parameter, values_from = mean)
```

##### Weighting function w(p)

We start by plotting the weighting curves for all parameter combinations under piecewise integration.

```{r}

cpt_curves_piecewise <- curves_cpt %>% 
  filter(strategy == "piecewise") %>% 
  expand_grid(p = seq(0, 1, .1)) %>% # add vector of objective probabilities
  mutate(w = round(exp(-delta*(-log(p))^gamma), 2)) # compute decision weights (cf. Prelec, 1998)

# all strategy-parameter combinations 

cpt_curves_piecewise %>% 
  ggplot(aes(p, w)) + 
  geom_path(size = .5) +
  geom_abline(intercept = 0, slope = 1, color = "red", size = 1) +
  labs(title = "Piecewise Integration: Weighting functions",
       x = "p", 
       y= "w(p)") + 
  theme_minimal() 
```

```{r}
cpt_curves_piecewise %>% 
  ggplot(aes(p, w)) + 
636
  geom_path() +
linushof's avatar
linushof committed
637
638
  geom_abline(intercept = 0, slope = 1, color = "red", size = 1) +
  facet_wrap(~a) + 
639
640
641
642
643
  labs(title = "Piecewise Integration: Weighting functions",
       x = "p",
       y= "w(p)",
       color = "Switching Probability") + 
  scale_color_viridis() +
linushof's avatar
linushof committed
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
  theme_minimal() 
```

```{r}
cpt_curves_piecewise %>% 
  ggplot(aes(p, w, color = s)) + 
  geom_path() +
  geom_abline(intercept = 0, slope = 1, color = "red", size = 1) +
  labs(title = "Piecewise Integration: Weighting functions",
       x = "p", 
       y= "w(p)", 
       color = "Switching Probability") + 
  scale_color_viridis() +
  theme_minimal() 
```

```{r}
cpt_curves_piecewise %>% 
  ggplot(aes(p, w, color = s)) + 
  geom_path() +
  geom_abline(intercept = 0, slope = 1, color = "red", size = 1) +
  facet_wrap(~a) + 
  labs(title = "Piecewise Integration: Weighting functions",
       x = "p",
       y= "w(p)",
       color = "Switching Probability") + 
  scale_color_viridis() +
  theme_minimal() 
```

##### Value function v(x)

```{r}

cpt_curves_piecewise <- curves_cpt %>% 
  filter(strategy == "piecewise") %>% 
  expand_grid(x = seq(0, 20, 2)) %>% # add vector of objective outcomes
  mutate(v = round(x^alpha, 2)) # compute decision weights (cf. Prelec, 1998)

# all strategy-parameter combinations 

cpt_curves_piecewise %>% 
  ggplot(aes(x, v)) + 
  geom_path(size = .5) +
  geom_abline(intercept = 0, slope = 1, color = "red", size = 1) +
  labs(title = "Piecewise Integration: Value functions",
       x = "p", 
       y= "w(p)") + 
  theme_minimal() 
```

```{r}
cpt_curves_piecewise %>% 
  ggplot(aes(x, v, color = s)) + 
  geom_path(size = .5) +
  geom_abline(intercept = 0, slope = 1, color = "red", size = 1) +
  labs(title = "Piecewise Integration: Value functions",
       x = "p", 
       y= "w(p)") + 
  scale_color_viridis() + 
  theme_minimal() 
```

```{r}
cpt_curves_piecewise %>% 
  ggplot(aes(x, v, color = s)) + 
  geom_path(size = .5) +
  geom_abline(intercept = 0, slope = 1, color = "red", size = 1) +
  facet_wrap(~a) + 
  labs(title = "Piecewise Integration: Value functions",
       x = "p", 
       y= "w(p)") + 
  scale_color_viridis() + 
  theme_minimal() 
```

720
#### Comprehensive Integration
linushof's avatar
linushof committed
721
722
723
724

##### Weighting function w(p)

We start by plotting the weighting curves for all parameter combinations under piecewise integration.
linushof's avatar
linushof committed
725

linushof's avatar
linushof committed
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
```{r}

cpt_curves_comprehensive <- curves_cpt %>% 
  filter(strategy == "comprehensive") %>% 
  expand_grid(p = seq(0, 1, .1)) %>% # add vector of objective probabilities
  mutate(w = round(exp(-delta*(-log(p))^gamma), 2)) # compute decision weights (cf. Prelec, 1998)

# all strategy-parameter combinations 

cpt_curves_comprehensive %>% 
  ggplot(aes(p, w)) + 
  geom_path(size = .5) +
  geom_abline(intercept = 0, slope = 1, color = "red", size = 1) +
  labs(title = "Comprehensive Integration: Weighting functions",
       x = "p", 
       y= "w(p)") + 
  theme_minimal() 
```

```{r}
cpt_curves_comprehensive %>% 
  ggplot(aes(p, w)) + 
  geom_path(size = .5) +
  geom_abline(intercept = 0, slope = 1, color = "red", size = 1) +
  labs(title = "Comprehensive Integration: Weighting functions",
       x = "p", 
752
753
       y= "w(p)") + 
  facet_wrap(~a) + 
linushof's avatar
linushof committed
754
755
756
757
758
  theme_minimal() 
```

```{r}
cpt_curves_comprehensive %>% 
759
760
  ggplot(aes(p, w, color = s)) + 
  geom_path() +
linushof's avatar
linushof committed
761
762
763
  geom_abline(intercept = 0, slope = 1, color = "red", size = 1) +
  labs(title = "Comprehensive Integration: Weighting functions",
       x = "p", 
764
765
766
       y= "w(p)", 
       color = "Switching Probability") + 
  scale_color_viridis() +
linushof's avatar
linushof committed
767
768
769
770
771
772
773
774
  theme_minimal() 
```

```{r}
cpt_curves_comprehensive %>% 
  ggplot(aes(p, w, color = s)) + 
  geom_path() +
  geom_abline(intercept = 0, slope = 1, color = "red", size = 1) +
775
  facet_wrap(~a) + 
linushof's avatar
linushof committed
776
  labs(title = "Comprehensive Integration: Weighting functions",
777
778
       x = "p",
       y= "w(p)",
linushof's avatar
linushof committed
779
780
781
782
783
784
785
       color = "Switching Probability") + 
  scale_color_viridis() +
  theme_minimal() 
```

```{r}
cpt_curves_comprehensive %>% 
786
  filter(s >= .7) %>% 
linushof's avatar
linushof committed
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
  ggplot(aes(p, w, color = s)) + 
  geom_path() +
  geom_abline(intercept = 0, slope = 1, color = "red", size = 1) +
  facet_wrap(~a) + 
  labs(title = "Comprehensive Integration: Weighting functions",
       x = "p",
       y= "w(p)",
       color = "Switching Probability") + 
  scale_color_viridis() +
  theme_minimal() 
```

##### Value function v(x)

```{r}

cpt_curves_comprehensive <- curves_cpt %>% 
  filter(strategy == "comprehensive") %>% 
  expand_grid(x = seq(0, 20, 2)) %>% # add vector of objective outcomes
  mutate(v = round(x^alpha, 2)) # compute decision weights (cf. Prelec, 1998)


# all strategy-parameter combinations 

cpt_curves_comprehensive %>% 
  ggplot(aes(x, v)) + 
  geom_path(size = .5) +
  geom_abline(intercept = 0, slope = 1, color = "red", size = 1) +
  labs(title = "Comprehensive Integration: Value functions",
       x = "p", 
       y= "w(p)") + 
  theme_minimal() 
```

```{r}
cpt_curves_comprehensive %>% 
  ggplot(aes(x, v)) + 
  geom_path(size = .5) +
  geom_abline(intercept = 0, slope = 1, color = "red", size = 1) +
  facet_wrap(~a) + 
  labs(title = "Comprehensive Integration: Value functions",
       x = "p", 
       y= "w(p)") + 
  theme_minimal() 
```

```{r}
cpt_curves_comprehensive %>% 
  ggplot(aes(x, v, color = s)) + 
  geom_path(size = .5) +
  geom_abline(intercept = 0, slope = 1, color = "red", size = 1) +
  labs(title = "Comprehensive Integration: Value functions",
       x = "p", 
       y= "w(p)") + 
  scale_color_viridis() + 
  theme_minimal() 
```

```{r}
cpt_curves_comprehensive %>% 
  ggplot(aes(x, v, color = s)) + 
  geom_path(size = .5) +
  geom_abline(intercept = 0, slope = 1, color = "red", size = 1) +
  facet_wrap(~a) + 
  labs(title = "Comprehensive Integration: Value functions",
       x = "p", 
       y= "w(p)") + 
  scale_color_viridis() + 
  theme_minimal() 
```
linushof's avatar
linushof committed
857

858
859
860
861
# Discussion 

# Conclusion

862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
# Appendix 

Let $X_n$ and $Y_m$ be independent and discrete random variables of the sequences

$$\begin{equation}
X_1, ..., X_n, ..., X_{N_X}
\end{equation}$$

and 

$$\begin{equation}
Y_1, ..., Y_m, ..., Y_{N_Y} \; .
\end{equation}$$

Then

$$\begin{equation}
P\left(\frac{\overline{X}_{N_X}}{\overline{Y}_{N_Y}} > 0 \right) = 
P\left(\frac{\frac{1}{N_X}\sum\limits_{n=1}^{N_X} (X(\omega_i) = A'_X \in \mathscr{F'}_X)_n} 
{\frac{1}{N_Y}\sum\limits_{m=1}^{N_Y} (Y(\omega_j) = A'_Y \in \mathscr{F'}_Y)_m} > 0 \right)
\end{equation}$$

is the probability that the quotient of the mean of both sequences takes on a value larger $0$.
Given the sample sizes $N_X = N_Y = 1$, the equation reduces to 

$$\begin{equation}
P\left(\frac{X(\omega_i \in \Omega_X) = A'_X \in \mathscr{F'}_X}{Y(\omega_j \in \Omega_Y) = A'_Y \in \mathscr{F'}_Y} > 0 \right) \; ,
\end{equation}$$

which is the sum across all joint probabilities $p(\omega_i \cap \omega_j)$ for which the above inequation holds:.


$$\begin{equation}
D:= 
  f
  \left(
    \frac{\overline{X}_{N_X}} {\overline{Y}_{N_Y}}
  \right) 
  =
  \begin{cases}
    1 & \text{if} & \frac{\overline{X}_{N_X}}{\overline{Y}_{N_Y}} > 0 \in \mathscr{D} \\
    0 & \text{else}. 
  \end{cases} 
\end{equation}$$

linushof's avatar
linushof committed
907
# References
908