Jones, C., Hollenhorst, S., Perna, F.  (2003)  'An Empirical Comparison of the Four  Channel Flow Model and Adventure Experience Paradigm', in Leisure Sciences, 25: 17 - 31.

A classic quantitative analysis, attempting to see how well different approaches and models will predict the findings from various questionnaires administered to white water kayakers. Technical conclusions aside, the paper does contain some information about how to operationalize both 4 channel flow models and adventure experience paradigms.

A number of empirical and conceptual approaches have tried to explain the experience is delivered by adventure activities. Csikszentmihalyi began by considering what Maslow's  'peak experiences' would actually look like, or later, what intrinsic rewards felt like. This led to the work on flow as:

(1) Merging of action and awareness in which complete attention is given to the activity at the present moment and actions become automatic.
(2) Clear goals and feedback in which the goals of the activity are clearly defined and feedback is immediate, allowing the individual to assess the potential of meeting their goals and, thus, become completely involved in the activity.
(3) Concentration on the task at hand in which an individual specifically focuses on the activity with total concentration  [!].
(4) Sense of control over actions and the environment where there is no conscious awareness of control but rather a lack of worry about loss or lack of control.
(5) Loss of self consciousness in which an individual becomes one with the activity to the degree that concern for the self disappears  (i.e. lack of awareness of physical pain or appearance).
(6) Transformation of time where time has altered by the rhythm of the activity rather than the referents of time of day  (ie time can be perceived to speed up, slowdown, or stand still).  (18).

This work has been operationalised to yield  'flow indicators of positive emotion, cognition, affect, activation, cognitive efficiency, and intrinsic motivation' or, in another study,  'positive mood, tension, freedom of choice, intrinsic motivation, concentration, competence, physical awareness, and potency' (18).

For Csikszentmihalyi, flow depends on matching challenge to skill, and this led to the four-channel flow model:  '(1) flow occurs when challenge and skill are above one's personal mean,  (2) anxiety occurs when challenge is above the personal mean and skill is below,  (3) boredom occurs when skill is above the personal mean and challenge is below, and  (4) apathy occurs when both challenge and skill are below the personal mean' (19). This model has been tested in daily settings, involving self rating of conscious experience. However, in the daily setting, really absorbing adventurous activities are less common. Better results should be obtained by focusing on adventure settings.

In whitewater river paddling, we need to be careful to remember that 'challenge' may not just relate to physical events. Skill too can have 'a physical, cognitive, or emotional meaning' (19). This makes it difficult to classify actual kayaking activities [so that the most alarming looking can actually be generating apathy or boredom for an experienced kayaker]. Csikszentmihalyi has also pointed out that it is a matter of balancing perceived risks and with the ability to minimise risk that can generate flow. Nevertheless, this is the first study that actually tries to see how real life challenges, such as difficult rapids, can influence perceptions.

There is another approach to understanding the experience -- the Adventure Experience Paradigm  (AEP) (Martin and Priest 1986). This is based on the idea of flow, but focuses on 'the balance of perceived risk and competence' in adventure activities specifically, and provides for five conditions --  'peak adventure' where the balance is sufficient to deliver euphoria,  'adventure' and  'exploration and experimentation' where competence exceeds risk, while  'misadventure'and  'devastation and disaster' represent the excess of risk over competence. This model too has been operationalised by getting paddlers to complete questionnaires before and after they coped with various up rapids. Different sorts of comparison between the scores before and after are supposed to indicate the different stages of the model -- for example,  'pre-trip and pre-first rapid evaluations revealed the perceived risks exceeded competence and these experiences were characterised as misadventure' (21). Data taken post trip and after the second major rapid reveal the experience has been an adventure where competence exceeded risk. This study shows apparently that perception is affected by the actual experience of overcoming difficulties. The discussion is further extended by comparing  'absolute, perceived, and real forms of risk and competence' (22)  [ 'absolute' seems to mean worst case].

Perceptions of risk and competence depend on how well the adventurers can demonstrate their  'ability to influence or control the environment' (22). This is why  inexperienced adventurers perceived there to be greater danger and risk. Experienced adventurers require  '"edgework skills" which refer to a unique ability much more complex than basic hard skills... the ability to exert cognitive control over a potentially disastrous situation which most individuals would perceive as uncontrollable... a type of mental stamina that allow an individual to avoid being overcome by fear and to focus attention and actions on critical moments of survival'  [citing Schuett 1991] (22).

Flow models and are similar conceptually and the categories can be related -- peak adventure with flow, for example. Actual procedures for using the models vary, however  [flow models operationalize flow in terms of average scores for challenge and skill compared to perceived challenge and skill, while AEP uses combinations of scores to produce a particular sector on a graph].

[For some strange reason, it is these methodological issues that are pursued in the rest of the paper. This particular focus involves a test of 4 particular hypotheses -- such as that the explanatory power of the two models will be similar. There is an interesting discussion of different types of validity.  'Convergent validity' refers to  'the extent to which the measure correlate highly with other measures designed to evaluate the same construct', while  'ecological validity' refers to 'the association between the experimental design and  "real life" contexts' (23). It is the latter that refers to the novelty of this study, which argues that real difficulties, rapids classified as extremely challenging for example, will produce higher perceptions of challenge and risk, while paddlers' perceptions of their skill and competence will be higher when paddling less challenging rapids.]

Research consisted of asking kayakers a number of questions in exchange for a small bribe,  'at the put-in, take-out and immediately below six rapids of varying difficulty level' (24). Prompts were supplied to guide recollected experience. Special 10 point scales were devised to measure perceived challenge,skill, risk and competence. Scores were standardised  'a control for individual response bias' (24). Data were then manipulated to conform to both flow model and AEP procedures as above. Previously used operational items were also offered for response  (as above), especially  'perception of the transformation of time... intrinsic motivation... involvement... merging of action and the winners... concentration on the task at hand... paradox of control  ("Were you in control of the situation?"), and  (lack of) physical awareness' (25). Respondents ticked positions on seven point scales. There was a background questionnaire to gather details of experience of paddling difficult rapids.

Much statistical processing then ensued. Convergent validity proved to be moderately good in most cases. The sample chosen was experienced and displayed little internal variation. Both models explained the data to about the same extent  [we already know that this is not actually very great for either, in terms of prediction at least -- 'both models explained only a new small proportion of the variance in the factor score of flow indicators, statistically neither of these models performed strongly in explaining the optimal experience construct' (29)]. Ecological validity was also supported, so that there was 'a statistically significant difference in... scores... between low and high difficulty levels on the river' (29). Perceptions of risk and competence also varied according to real difficulties, as predicted, at least when extremes of difficulty were included [less so with the closer rapids].

It seems that 'flow is still an elusive concept that is difficult to empirically capture' (29). Further work to operationalize the models is required. Perhaps flow should be further stratified according to intensity, and this more intense kind of flow might require  'a much more complex set of parameters than above-average challenge-skill or the literal balance of risk-competence' (30). The proposed solution here is to develop more valid scales to measure the different dimensions, and then try to capture intensity as an effect of interaction of these dimensions.

Martin, P. and Priest, S. (1986) 'Understanding the adventure experience' in Journal of Adventure Education, 3, 1: 18--21
Schuett, M. (1991) (unpublished PhD)