Understanding Mind and Brain

Random Article

Latest Posts

Capturing Daydreams





Bottom Line

Abstract The investigation of daydreaming provides a unique set of challenges to the would be researcher, but also is an invaluable source of novel evidence on existing psychological problems. This article will provide a summary of the problems that face an experimenter when attempting to sample inner experience and the different methods that people have […]

Posted September 13, 2003 by thomasr

Full Article


The investigation of daydreaming provides a unique set of challenges to the would be researcher, but also is an invaluable source of novel evidence on existing psychological problems. This article will provide a summary of the problems that face an experimenter when attempting to sample inner experience and the different methods that people have used to overcome them. In the second half, this article will describe how in the future, the investigation of daydreams may help address important issues, both practical and theoretical, in the attempt to understand how the brain generates conscious awareness.

Capturing Daydreams

Daydreaming can be considered a special form of conscious awareness because it reflects a state in which “conscious awareness is to some extent decoupled from the current situation” (Smallwood et al in 2003 B). Over the last thirty years this line of research has been pursued by a number of psychologists including Jerome Singer, John Antrobus, Eric Klinger, John Teasdale and Leonard Giambra. To investigate daydreams the most common experimental method employed is thought sampling which is an attempt to sample the inner experience of an individual either (i) as they complete a task under laboratory conditions or (ii) as they go about their lives on a day to day basis. A variety of different approaches have been used to measure daydreaming using a thought sampling methodology and they broadly fall into two categories: (i) the thought probe technique, in which Daydreaming can be considered a special form of conscious awareness at various points in a task the thinking of the individual is probed and the sample is recorded either by verbal report (Smallwood et al 2003 A & B) or via a button push (Giambra, 1995; Schooler et al in press) and (ii) the self-report method in which the individual is primed with the purpose of the experiment and responds with a button push whenever they experience a daydream (Antrobus, 1968). It is important that research should continue to employ these different approaches because we do not at present know which of the two is most appropriate as an experimental tool, and both approaches have their own unique advantages.

The technique that I have employed in my investigations is the thought probe technique, in which a participant is asked to report what is passing through their mind as they complete a series of experimental tasks. These verbally reported thought are recorded and coded in terms of the extent they reflect a phenomena which is described as task unrelated thought, using the published criteria (Smallwood et al in press B, see below). Before beginning the task the participant is informed that the thought probe will take place, but are not informed about the categories investigated. The participants are informed that if they do not feel happy about reporting all the specific details about their thoughts, it is okay to just give the general direction, such as “I was thinking about something that happened when I went for a walk in the woods”.

As the information generated is subjective it is important to make an attempt to validate the classification system we employ. In my research it is possible to make two attempts to validate this process. First, you can ensure that several independent judges can agree on whether a verbal report qualifies as an example of a daydream. Using trained judges agreement is usually high (often greater than 85%). The second method for ensuring the accuracy of Using trained judges, agreement is usually high the judgments is to give out a questionnaire which measures the frequency of daydreams during the task to the participants at the end of the task, and seeing if their assessment of the experience of daydreaming matches those of the independent judges. Again the questionnaire responses tend to correspond to the classifications made by the judges. Table One gives an idea of the sort of thoughts which would be classified as a daydream.

Task Unrelated Thought (TUT)

The participant reported thought bears no relationship to the task in hand. The thought is not directly related to any obvious task relevant stimuli, stimuli in their current environment or towards their performance on the task. Examples of TUT include thinking about the future, for example participants may report “I was wondering about what I was going to do for dinner” or “I was thinking about what I was going to do tonight”. In addition the individual may describe thoughts about memories such as “I was remembering a meal I had in a restaurant” or “I remembered going for a walk in the woods”. Finally, the participant may explicitly suggest that their attention had wandered from the task, i.e. “My mind was wandering, I can’t remember what I was thinking about”. Criteria developed from Smallwood, Obonsawin and Reid, in press

Table One: Criteria for the rating of Task Unrelated Thought (a daydream) and some examples of what these thoughts may be.

The problems associated with daydream research

Other than the method of thought sampling the researcher is faced by three key issues when examining daydreaming. The first is concerned with the degree of experimental control that one has over the phenomena. In the investigation of visual experience, for example, with the appropriate experimental control an experimenter can ensure each participant experiences the relevant stimulus for the same duration and with a similar intensity. By contrast when researching internally generated phenomena, such as a daydream, one cannot exert the same degree of experimental control. As scientists this challenges our ability to identify causality in the phenomenon of inner experience, because the data is fundamentally based on a series of correlations.

Whilst this problem is brought into stark contrast when investigating daydreaming, it is not unique to this area. In fact, many well respected psychological phenomena which are routinely investigated, such as learning or memory, require the individual to report some aspect of their internal state through verbal report. Moreover, in memory research, for example, all participants do not recall all items presented in a standard test, and it is only possible to generate broad and replicable patterns on the conditions in which memory for a stimulus is enhanced. Research into daydreaming is the same, and at present a lot of research has been devoted to understanding the conditions in which inner experience change. For example, converging evidence suggests that tasks with longer absolute durations (Teasdale et al. 1995; Smallwood 2003) and longer block lengths (Giambra, 1995; Smallwood et al in press) yield more subjective reports which reflect daydreams. Thus, whilst we cannot predict whether an individual’s verbal report will reflect a day dream or not, we can say with some certainty that daydream frequency increases with time on task.

A second problem concerned with investigating inner experience, is the frequency with which the phenomena occurs. Daydreams occur 10 – 20 % of the time depending on the nature of the task (Smallwood et al 2003 A & B; in press; see also Klinger and Cox, 1987). Given this low base rate, it is often difficult to ensure that we measure enough examples of the phenomenon to ensure we can conduct the robust analysis necessary for a scientific endeavour. One solution to this problem is to measure daydreaming in tasks that are very easy, such as a sustained vigilance task in which the participant detects a target tone over a prolonged period (Giambra, 1995). In these circumstances we can identify the reliability of the measurements by repeating them over a period of time (Grodsky and Giambra, 1991).

Whilst this attempt ensures that we can observe a high number of daydreams, we have no way to validate our judgements because the task is so easy that daydreaming does not interfere with task performance. A second solution to the base rate issue, therefore, is to measure daydreams in the context of tasks which one cannot perform accurately without attending to the task. Using this approach, experimental evidence suggests that when verbal reports indicate that the attention of the individual has strayed off task, random number generation is impaired (Teasdale et al. 1995), encoding accuracy decreases (Smallwood et al 2003 B; in press) and comprehension of written text is reduced (Schooler et al in press). This approach is likely to be important in the future because it provides a source of validation for the phenomena in question, and moreover, may provide a more sensitive measure of daydreaming than is provided by frequency alone. For example, it is possible that in the future we can classify the quality of a daydream in terms of the extent to which it interferes with attention to the external environment: providing a quantitative index of the stimulus-independence of the phenomenon. It is plausible that these more objective measures may prove to be important in the investigation of state of psychopathology which are associated with self-focus, such as depression (see Smallwood et al in press C).

Finally, the most obvious challenge facing research into daydreaming is that the data is based on introspection and therefore subject to the limitations that are reflected in the investigation of this type of data. In particular, the verbal reports may be (i) skewed because of desire of the participant to maintain face, or, (ii) alternatively that the participant does not have access to the necessary information to make the report successfully (Nisbett and Wilson, 1977). When researching daydreaming, one can overcome this issue, to some extent, by asking the participant to report direction of thought, without needing to report the specific details. Because the individual is not asked to produce any personal information, it is likely that the individual will be more likely to describe their subjective experience in an honest fashion. Either way, if the participants were unwilling or unable to report their inner experience, it is not clear that these reports would correlate with task performance. As was noted above, however, these verbal reports do correlate with task performance in the predicted manner and this relationship holds whether the participants are blind to the phenomena under investigation (Smallwood et al 2003 B) or are trained in its detection (Schooler et al in press). Within reasonable bounds, therefore, the fact that the measure is based on introspective data does not influence our ability to identify and validate broad properties of daydreams in the laboratory.

Implications of daydream analysis for consciousness research

One area of research in which daydream research may prove to be important is in our ability to understand consciousness awareness. Changes in inner experience similar to a daydream occur over a period of minutes (Smallwood et al in press B; Schooler et al in press; Giambra, 1995) and make measurable contributions to task performance for at least thirty seconds preceding the thought probe (Teasdale et al 1995; Smallwood et al 2003B: Schooler et al in press). This means that changes in phenomenal experience associated with a daydream represent a dynamic state which is intermediate in duration relative to alternative methods of investigation of consciousness, such as the sleep wake cycle (hours) or perceptual research (seconds). From a theoretical perspective, therefore, the understanding of phenomena, such as daydreaming, might yield an important source of information on changes in consciousness over an intermediate time duration. If, for example, one could detect changes in coherence between two brain areas which preceded the verbal report of a daydream, this would prove a novel source of information for how awareness is expressed in the brain. At present the majority of research has focused on detecting differences in external task performance, because this method is important in determining the status of inner experience as a measurable psychological phenomenon. As we gather more detailed information regarding the experience of daydreams, it is possible that we may be able to detect the correlates of this phenomena and this may ultimately yield important insight into how our brains generate consciousness.

From a practical perspective, however, the investigation of thought sampling raises an important question for how we interpret information based on measurements gained from either Position Emission Tomography (PET) or Functional Magnetic Resonance Imaging (FMRI). Research on daydreaming makes explicit the assumption that we cannot maintain task relevant stimuli in awareness during any prolonged cognitive task, without occasional lapses (see Schooler et al in press). These lapses occur most frequently when task demands are low (Giambra, 1995; Teasdale et al. 1993; 1995; Smallwood et al in press A) and increases with the length of testing session (Smallwood et al in press B). In fact, recent work by Binder and colleagues (1999) demonstrated that activity in the pre-frontal cortex during an eyes closed resting condition was broadly similar to the activity in a semantic/conceptual task. Scanning studies often contrast a resting control condition, as a baseline, with an active task, as the experimental task. As we know that daydreaming will occur with greater frequency in the control condition than in the experimental condition it seems plausible that some of the activity in the resting control condition will reflect self-generated processing such as daydreaming. This raises important questions about what constitutes a control condition in a scanning study because the neural baseline commonly employed may in fact be a state of self-generated stimulation (Gusnard and Raichle, 2001).

As the use of PET and fMRI technology increases in the psychological domain it is likely our ability to isolate specific psychological phenomena will become more accurate. However, as the nature of our measurements become more detailed, phenomena such as daydreams are increasingly likely to induce error in our measurements. If this is the case, then a comprehensive account of how the brain generates higher aspects of cognition will require an understanding of what is going on during states of self-generated stimulation such as daydreaming, if only to control for the phenomenon.

In the future, therefore, it is possible that thought sampling studies will provides an important source of information in the search to understand how the brain generates consciousness. It is possible that a comprehensive account of the processes involved in daydreaming will directly contribute to our knowledge of consciousness or alternatively it may allow us to control for activity in the resting brain which otherwise will contaminate our ability to measure specific components of conscious awareness. Either way, it is likely that the use of thought sampling procedures over the next decade will increase to reflect the importance of an understanding of the processes of self-generated stimulation such as daydreaming as an important and yet under researched cognitive phenomenon.

© 2003 J. Smallwood


  1. Binder JR, Frost JA, Hammeke PS, Bellgowan SF, Rao SM & Cox RW (1999). Conceptual processing during the conscious resting state: A Functional MRI Study. Journal of Cognitive Neurosciences, 11:1, 80 –93.
  2. Gusnard DA & Raichle ME (2001). Searching for a baseline: functional imaging and the resting human brain. Nature Reviews: Neuroscience, 2, 685-694.
  3. Grodsky A & Giambra L (1989). Task unrelated images and thoughts whilst reading. In J.Shorr, P. Robin, J.A. Connek and M. Wolpin, Imagery: current perspectives, NY, Plenum Press.
  4. Giambra L (1995). A laboratory based method for investigating influences on switching attention to task unrelated imagery and thought. Consciousness and Cognition, 4, 1-21.
  5. Klinger EC (1999). Thought flow: properties and mechanisms underlying shifts in content. In Jerome A. Singer & Peter Salovey (Eds) At play in the fields of consciousness: Essays in the honour of Jerome L Singer. Lawrence Erlbaum: Mahwah : NJ.
  6. Nisbet RE & Wilson TD (1977). Telling more than we can know: Verbal Reports on Mental Processes. Psychological Review, 84, 3.
  7. Schooler JW, Reichle ED & Halpern DV (in press). Zoning-out during reading: Evidence for dissociations between experience and meta-consciousness. Daniel Levin (Eds) Visual Meta-Cognition: Thinking About Seeing. Praeger.
  8. Smallwood JM, Obonsawin MC, & Heim SD (2003 A). Task Unrelated Thought: the role of distributed processing. Consciousness and Cognition
  9. Smallwood JM, Baraciaia SF, Lowe M & Obonsawin MC (2003 B). Task unrelated thought whilst encoding information. Consciousness and Cognition.
  10. Smallwood JM, Obonsawin MC, Reid H & Heim, SD (in press A). An investigation into the role of personality and situation in the maintenance of subjective experience in a laboratory. Imagination, cognition and personality.
  11. Smallwood JM, Obonsawin MC & Reid H. (in press B). The effects of block duration & task demands on the experience of task unrelated thought. Imagination, cognition and personality
  12. Smallwood JM, Obonsawin MC, Baracaia SF, Reid H, O’Connor RC & Heim SD. (in press C). The Relationship Between Rumination, Dysphoria and Self-referent Thinking: Some Preliminary Findings. Imagination, cognition and personality.



One Comment


    While reading this article I felt that I am being taken to nowhere voyage in the the sea of daydreaming. This does not underestimate in anyway the seriousness in such experiments but actually builds no much hope to acquire any fruitful results.
    I need not be told what “problems associated with daydreams research” that is because the mere saying that an experiment is to be carried out in this particular area will definitely recall most of those problems including some of the “implications” of analysis declared above. For me, I could not expect much to come out of such experiments. For the writer, as we have read in the article he is optimistic in the sense of pinning hopes to a “decade” or more to come expecting more improved “measurments” and suggesting to others to work on new experiments in the same pursuit. What about man’s really detoriorating faith and ethics? I have taken into consideration right from the beginning that an experimenter should bear in mind the idea of dishonesty. People who are accustomed to “lying” may be, unknowingly, invited to labs, and those – who are not few in number – will help an experimenter in registering so strange set of frequencies that will be of no use to science: Just a waste of time, and good for nothing. I was a little relieved when I came across the term “describe their (participant) subjective experience in an HONEST fashion”.
    Using a term like “Within reasonable bounds .. introspective data does not influence..” would fairly/logically suggest: Within reasonable bounds .. introspective data DOES influence the 50% remaining percentage ..etc.,.

Leave a Response