Capítulo 3 Las TIC en la Enseñanza
3.1 Las TIC en el ámbito educativo
3.1.2. El papel del docente y el alumno frente a las TIC
General introduction to studies
Daily reasoning is permeated by the search for causes and effects of life events. For example, we constantly wonder about the reasons and the consequences of other’s behavior. Tversky and Kahneman (1974) demonstrated that people generally use their knowledge and experience to make both predictive and diagnostic inferences. Predictive reasoning is the process through which people infer the likelihood of effects starting from the cause. By contrast, diagnostic reasoning is the process through which the likelihood of causes are inferred starting from the effect. For example, while reading about tax increases, people are likely to make predictive inferences about the effects on their economic situation; while reading about suicide cases, people usually make diagnostic inferences about possible causes behind such extreme acts. Finally, while reading that “olive trees are attacked by a fungus”, people may look either for causes (e.g., climate change, pesticides use) or for consequences (e.g., reduced oil production, contagion of other trees). In the latter example, it is critical to understand what is driving the reader to one or the other process.
Research on the distinction between predictive and diagnostic reasoning shows that predictive, in contrast to diagnostic, reasoning is easier because it follows the natural order of causes preceding effects. For this reason, people judge causal links stronger in predictive than in diagnostic inferences (Tversky & Kahneman, 1974). For instance, Fernbach, Darlow, and Sloman (2010), across three experiments, they demonstrated that people tend to underestimate the potential role of alternative causes when making predictive, but not when making diagnostic inferences. For instance, clinicians read a statement about a patient who was diagnosed with depression (predictive frame) or presented a symptom of lethargy (diagnostic frame). Then, they were asked to judge the
likely to present lethargy given the depression diagnosis, rather than suffering from depression given the lethargy symptom. Thus, the relation between depression and lethargy was evaluated stronger when a predictive frame was adopted. The authors stressed that, when asked to make predictive inferences, individuals may consider the cause as the sole explanation of the effect and hence perceive the link as stronger; to the contrary, by generating diagnostic inferences, they may engage in a reasoning process in which a greater number of possible causes is considered (Fernbach, et al., 2010). However, this research is confounded with respect to the order in which causes and effects were presented. When investigating the processes underlying predictive and diagnostic reasoning, Fernbach et al (2010) also changed the order in which cause and effect appeared in the experimental material without considering the order as a potential independent variable able to favor one direction (e.g., predictive, from causes to effects) over the other (e.g., diagnostic, from effects to causes). Hence, it is not clear whether frame or word order drove the observed effects, or both. Thus it remains to be understood whether the result pattern depended exclusively on the induced frame (predictive vs. diagnostic) or on the order in which causes and effects were to appear (Cause-Effect vs. Effect-Cause). It is well possible that word order in and by itself induces predictive vs. diagnostic reasoning which in turn affects the perceived strength of the link between cause and effect.
Within the normative framework of Bayes nets in which a causal structure is used to define a probability (Spirtes, Glymour, & Scheines, 1993), Meder, Mayrhofer, and Waldmann (2014) showed that in case of uncertainty regarding the underlying causal link, the structure of test questions helps people to distinguish between the two processes. Questions used to explore predictive and diagnostic causal inferences were created in order to mirror both the Cause - Effect and the Effect – Cause direction: in the former case the word related to the cause was placed before the one related to the effect. Again, the order of causes and effect within the questions (Cause- Effect vs. Effect-Cause) and type of reasoning (predictive vs. diagnostic frame) were confounded and it is not clear which of the two variables (word order or framing) is responsible for the type of
adopted reasoning. The authors suggested that the process of diagnostic reasoning is affected by both the belief about the existence of a causal relation between a given cause and an effect, and the probability that the effect has been produced by alternative causal by alternative causes. The authors asked participants to make two judgments about the relation between a disease and a substance present in the blood of a patient. They were presented with two different questions: one invited them to make a predictive inference from cause (the disease) to effect (the substance) (e.g., “How certain are you that a novel patient who has been infected with Midosis has the substance Rothan in his blood?”), while the other prompted the causal inference in the opposite direction, from effect to cause (diagnostic) (e.g. “How certain are you that a novel patient who has the substance Rothan in his blood has been infected with Midosis?”, Meder, et al., 2014). In addition to the experimental questions, participants were given a learning data set in which levels of diagnostic and predictive probabilities were combined. Results showed that diagnostic judgments were affected by higher levels of diagnostic probability in the learning data set. Of particular interest, individuals were also more likely to judge the presence of a substance (effect) as being causally related to the disease (cause) when the learning matrix showed higher predictive probabilities. Therefore, a stronger causal link between the two components led participants to judge the effect as being produced by the given cause (Meder, et al., 2014).
The previous literature supports the importance of both processes during the elaboration of a causal event. However, in previous studies the order of cause-effect vs. effect-cause was confounded with the inference frame, with cause-effect order being systematically associated with the induction of predictive reasoning (such as in Meder et al., 2014), and effect-cause ordering being associated with a diagnostic frame. The main purpose of my studies is to investigate whether the order in which Causes and Effects are disposed within sentences affects the strength of causal
Hegarty et al., 2011), in this research project I propose a novel approach to Word Order according to which the focus on the first term influences the strength perception of causal relations between elements. By switching the order in which causes and effects appear (Cause-Effect vs. Effect- Cause) individuals will be induced to either focus on the cause (in Cause-Effect order) or on the effect (in Effect-Cause order). I predict that cause-effect order will lead people to perceive a stronger link between cause and effect and that this may, in turn, influence subsequent intentions of changing current habits.
Despite a substantial literature focusing on causal reasoning and a growing literature focusing on the role of language in cognitive processes, the role of word order in the perception of causal relations has not yet received attention. Given the consistent confounding of framing and word order in prior research (see Meder et al., 2014; Fernbach et al., 2010), the separate effects of framing and word order on causal reasoning remain to be investigated. The studies reported below aim at exploring the impact of order on the perceived strength of the relation between two variables (cause and effect) Specifically, I predict that by manipulating only the order of presentation of cause and effect in the health domain, will induce participants to focus either on the cause (in cause- effect order) or on the effect (in effect-cause). This, in turn, will affect the perceived causal link, the perceived health risk and the perception of personal relevance as will be explained in greater detail below.
The potential relation between word order and causal reasoning will be tested in three distinct studies (Studies 3a, 3b and 3c). In all three studies I will present a series of health related binomials that are causally related (Cause-Effect, e.g., smoking-cancer; sodium-blood pressure). Sentences will be presented in the same form across all the three studies and the main independent variable will be the order in which the two terms of the binomials are arranged. In one case, the cause will precede the effect (e.g., Smoking has a link with lung cancer) and in the other, the effect will precede the cause (e.g., Lung cancer has a link with smoking). The linguistic frame of the causal relation as “a link” allows to investigate Cause-Effect order while keeping the remaining
information constant, and without imposing a specific causal reasoning (diagnostic vs. predictive) which may confound the effect of ordering.
The three studies differed with respect to the dependent variables. In Study 3a, participants reported the perceived correlation between the two health-related elements, in Study 3b they evaluated the degree of risk for themselves and for others, whereas in Study 3c they rated the personal relevance. Also, the intention to change risk behaviors was assessed in both Study 3a and Study 3c. . I predicted that the perceived relation (investigated in Study 3a) and the risk for self and others (investigated in Study 3b) will be stronger when participants are faced with cause-effect ordering of binomials because this order reflects the natural order of causes preceding the effects. Specifically, by focusing first on the cause (Study 2b), participants will think about that cause (smoking) as the most likely reason behind a certain effect (lung cancer). By contrast, by focusing on the effect (lung cancer), participants may think of different causes (smoking, pollution, genetics) than may have generated the effect. In study 3c, rating the personal relevance, I predicted that information in which effects precede the cause will be evaluated as more relevant than information where causes are presented first. The underlying mechanism responsible for this could lay in the fact that by placing in first position the effect, consequentially the focus of attention is primarily on the effect. My hypothesis consists in demonstrating that, in focusing on an issue of personal relevance such as health, the potential presence of a great number of causes behind a given effect may intensify the perception of the issue’s personal relevance.