1.THINK-ALOUD PROTOCOLS1.1History of verbal reports1.1.1Verbal reports in cognitive sciencesFor decades, humanbeings were interested about inner processes ongoing in their minds.
The first attempts to investigate the inner workings of human’s mindwere highly speculative, and were not overly concered about providingsufficient empirical support for their findings. In early attempts,individuals, predominantly philosophers, applied basic introspectivemethods in order to try to elicit data about mechanisms andstructures of their minds. Only in later centuries, bigger emphasison applying scientific methods to investigate human mind andbehaviour has been put. In cognitive psychology and in other sciencessignificant efforts have been made in order to find agreementregarding reliability of observations and disambiguate and what maybe deemed as „indisputable evidence”, and what may not (Ericsson& Simon, 1993: 48-49). Franciscus Donderscan be regarded as one of the pioneers of analysing cognitiveprocesses and played a significant role in the development of moderncognitive sciences. In his studies, by observing latencies, hemeasured reaction time of the subjects. Donders differentiated threetypes of reactions, the most basic one a-reaction, where thesubject immediately responds to the stimuli with a given response, inthe case of b-reaction there are different responses to eachstimulus, and in the c-reactionthe subject ought to reply only to certain stimulus, anddiscriminating others. (Ericsson & Simon: 52) Later, WilhelmWundt adopted this model and added his own, d-reaction,where the subject respond immediately after discrimation of thestimulus.
In the early 20thcentury introspective method was mainly used by structuralist, mostnotably Wundt and Titchener. According to the views ofstructuralists, self-observations can be regarded as reliable data;Wundt’s and Titchener’s hypothesis claimed that all mental states andexperiences in terms of their sensory and imaginal components(Ericsson & Simon: 1993: 51). Due to the fact, that introspectiveand retrospective reports are given post hoc, the experimenterhas to believe that information reported by the subject are true andcomplete (van Someren, Barnand & Sandberg 1994: 21).
Anotherissue with such reports is the fact, that reports given by thesubject are retrieved from long-term memory (LTM), it may happen thatdue to the limitations of human minds the subject will not be able toretrieve all the processes that occurred in short-term memory (STM,also known as working memory), or processes that actually did nottake place will be retrieved, thus unreliable data will be elicited.(van Someren, Barnand & Sandberg 1994). As a consequence ofmethodological issues, structuralists’ approach to and introspectionas a method itself, was severely criticised by Behaviourists andGestalt school of psychology. With one of the most influentialfigures being John Watson and his famous paper called „Psychologyas the Behaviorists Views it” (Ericsson & Simon 1993: 57).Watson argued that introspection can not be used for scientificpurposes, as the subjects report how they think that they perform acertain task, data regarding actual inner processes is not elicited,therefore it is unreliable. It is worth noting that Watson was notcritical of all forms of verbal reports, actually he believed thattypically covert thought processes can be elicited by usingthink-aloud verbalisations while carrying out an assigned task.(Ericsson & Simon 1993: 58-59).
With the rise ofBehaviourism amount of introspective studies significantly lessened,albeit Think-aloud protocols (TAPs) gained in popularity, withinvestigators such as Watson (1920), Duncker (1926), Bulbrook (1932),Smoke (1932) and Claparède (1934) asking their subjects forthink-aloud verbalisations. (Ericsson and Simon 1993: 61). In 60-70spopularity of the think-aloud method continued to grow. Newell andSimon (1972) in their work Human Problem Solving managed tobuild comprehensive problem solving models with by using Think-aloudprotocols and computer simulation.
(vanSomeren, Barnand & Sandberg 1994: 31,) (Holyoak & Morrison2005: 327). Think-aloud protocols, alsohave not escaped criticism, with the issue of comepleteness of thedata and doubts wheter concurrent verbalisations influence thesubjects’ performance amongst them (Ericsson & Simon 1993: 61). In 1984 a ground-breaking work by Ericsson and Simon „ProtocolAnalysis: Verbal Reports as Data” was published (revised editionpublished in 1993), which remainsa ‘bible’ regarding verbal reports, and sheds the light on allegedapparent issues. Today,Think-aloud protocols are used in many sciences and are regarded asvaluable and reliable tool with some limitations to acquire insightinto ongoing mental processes in the subject’s mind (van Someren,Barnand & Sandberg 1994: 32). And with rapid technologicalprogress, and aid of Artificial Intelligence or tools such aseye-tracking devices there is further scope for development ofthink-aloud methods in various fields.
