Stability of height fear acquisition pathways over one year.
Abstract: Compared to associative pathways to fear, non-associative pathways are more frequently endorsed by people with evolutionary-relevant fears such as heights and water. However, the ways in which onset pathways are often defined and used can generate biases, and pathway ascriptions have been found to be highly unstable over time. The present study examined the stability of pathways to the onset of height fear. To examine the stability of retrospective recall of height fear onset over time, 22 participants completed the same questionnaire at Time 1, and then three months (Time 2) and 12 months (Time 3) later. The questionnaire comprised the Origins Questionnaire-II as well as measures of fear severity and avoidance behaviour. There were no differences in the frequency of ascription to associative and non-associative pathways. Instability in pathway ascriptions were observed in 18% of cases over three months, and 27% of cases over nine months. This study supports prior research that demonstrates the limitations of a retrospective approach in ascertaining fear onset pathways.
Article Type: Report
Subject: Phobias (Psychological aspects)
Phobias (Analysis)
Authors: Chin, Edwin C.H.
Taylor, Joanne E.
Pub Date: 03/01/2011
Publication: Name: New Zealand Journal of Psychology Publisher: New Zealand Psychological Society Audience: Academic Format: Magazine/Journal Subject: Psychology and mental health Copyright: COPYRIGHT 2011 New Zealand Psychological Society ISSN: 0112-109X
Issue: Date: March, 2011 Source Volume: 40 Source Issue: 1
Geographic: Geographic Scope: New Zealand Geographic Code: 8NEWZ New Zealand
Accession Number: 257859219
Full Text: Contemporary understandings of fears and phobias have been expanded from the basic Pavlovian conditioning framework to encompass a variety of ways in which associative learning of fears can occur (Field, 2006; Mineka & Oehlberg, 2008). Rachman (1977) proposed that, in addition to the classical conditioning pathway, persistent fear responses may also be acquired through vicarious and informational transmission, although the influence of these pathways appears to vary with different fear types. For example, while most children who fear dogs and other animals have experienced associative learning experiences prior to the development of their fear (Doogan & Thomas, 1992; King, Clowes-Hollins, & Ollendick, 1997; King, Gullone, & Ollendick, 1998), the experience of associative events is much less prevalent for people with fears regarding water or heights (Graham & Gaffan, 1997; Menzies & Clarke, 1993b). There is also accumulating evidence that associative conditioning experiences are not always sufficient or necessary in establishing persistent fear to a wide range of objects and situations (e.g., Davey, 1992, 1997; King et al., 1997; Menzies & Clarke, 1993a; Merckelbach, Muris, & Schouten, 1996).

Given these limitations, the associative framework has been expanded to incorporate neo-conditioning perspectives such as UCS revaluation (where initial encounters with a stimulus that do not elicit fear are revaluated through subsequent experiences; Davey, 1989; de Jong, Muris, & Merckelbach, 1996) and interoceptive conditioning (where the physiological arousal that occurs during a CS encounter is considered to constitute UCSs that produce fear; Mineka & Ohman, 2002). However, these approaches have been seen as an attempt to fit all cases of fear acquisition into the conditioning framework, and to make the associative perspective all-encompassing and non-falsifiable (Poulton & Menzies, 2002a). In light of this, non-associative accounts have been proposed to complement the associative perspective and offer a theory for conceptualising fears that do not arise through conditioning pathways. It is argued that most people will express fear towards a range of evolutionary-relevant stimuli that have historically been threatening to human life (e.g., heights, water, early separation) on their first encounter, without having previously experienced aversive conditioning events (Menzies & Clarke, 1995b; Menzies & Harris, 1997; Poulton, Davies, Menzies, Langley, & Silva, 1998). These stimuli have the propensity to elicit fear in the absence of any external UCS, or in other words, a non-conditioning traumatic event. In a study with 148 adults who met the DSM-III-R criteria for height phobia, 56% of the sample either reported having always been fearful of heights, or described fear-onset events that were consistent with the non-conditioning traumatic event pathway, while associative accounts were attributed to only 31% of the cases (Menzies & Clarke, 1995a). These results have been replicated with a sample of 54 students (Menzies & Parker, 2001).

Comparisons of these findings to research on evolutionary-neutral fears reveal some similarities. For example, 40% of those with driving-related fears described non-associative onset pathways (Taylor & Deane, 1999). This is potentially problematic for the non-associative framework to explain, as it appears that the evolutionary relevance of a feared stimulus does not fully account for the fear to be acquired through non-associative pathways. One possible explanation is that the non-conditioning traumatic event and always been fearful non-associative pathways also encapsulate those who have simply forgotten about an associative encounter. Such limitations inherent to human memory impact on the accuracy of retrospective recall and possibly increase the endorsement of non-associative pathways to the development of height fear.

A few studies have demonstrated various types of limitations of retrospective recall of fear onset, raising questions over the accuracy and reliability of the retrospective research on which much of the fear acquisition literature is based. People may be unaware of conditioning encounters that occurred before the age at which autobiographical memory can be encoded and recalled, resulting in childhood amnesia (Mineka & Sutton, 2006). Menzies and Harris (1997) suggest that such capacities develop after the age of two, but even when the capacity to register autobiographical events has developed, the detail and complexity of these early memories is limited (Mineka & Sutton, 2006). Being unable to recall ambiguous associative events may add to some people's belief that they have always been fearful, a belief that has provided much of the support for the non-associative view of fear acquisition (Poulton & Menzies, 2002b). It has also been found that students' reports of past classical conditioning experiences were described with higher levels of certainty than reports of other pathways (Withers & Deane, 1995). The implication of traumatic associative experiences being more readily remembered is that mild, non-aversive encounters are comparatively less memorable (White & Davey, 1989). As a result, such biases may lead to the under-detection of fears established through UCS revaluation, where the initial encounter with the UCS may only be mildly aversive (Davey, 1989; Rescorla, 1968).

Even when information can be accurately recalled, there is also concern about the consistency of the recalled information, and only one study has explicitly investigated the stability of retrospective accounts of fear onset. Taylor, Deane, and Podd (1999) used the Origins Questionnaire (OQ; Menzies & Clarke, 1993b) with a community sample of 85 participants to examine the stability of retrospective accounts of driving fear onset across a 12-month period. After one year, 46% had ascribed to a different fear-onset pathway than what was originally described. The highest proportion of change occurred for the 11 participants who initially could not remember how they became fearful of driving, from which 9 participants ascribed to either a conditioning or non-conditioning fear-onset pathway after 12 months. Similarly, 11 of the 25 participants who initially ascribed to an associative pathway had changed their pathway ascriptions to either a different associative category or to a non-associative pathway after one year. These changes did not appear to be due to intervening driving-related events, fluctuations in fear severity, or trait anxiety (Taylor et al., 1999). Thus it appears that, over a period of 12 months, people have the potential to either remember past experiences or forget previously remembered events, highlighting the issue that the accuracy of retrospective data collected at any one point in time can be affected by the presence of memory limitations and errors.

However, a major limitation of the study by Taylor et al. (1999) was that the questionnaire used to determine the fear onset pathway at follow-up changed due to practical considerations. This may have contributed to the overall instability that was observed, and the findings need to be replicated in a study that exerts greater measurement control. The present study constitutes such a replication, using a stable method of ascertaining fear onset over time, and with height fear as the target fear. Based on the driving fear research conducted by Taylor et al. (1999), it was hypothesised that changes in participants' ascriptions of height-fear onset pathways over a 12-month period would be observed, although to a lesser extent because of the consistency in measurement over time in the present study.


Undergraduate psychology and media studies students were approached in class and in web-based communication forums and invited to take part in the study by taking or being mailed a pack containing study information as well as the Fear Survey Schedule Second Edition (FSS-II; Geer, 1965) and returning the completed FSS-II. Selection of participants into the main study was based on their responses to the height item on the Fear Survey Schedule-Second Edition (FSS-II; Geer, 1965). The selected participants subsequently completed a questionnaire about height fear at three points in time over 12 months: immediately after selection (Time 1), and three months (Time 2) and 12 months (Time 3) after Time 1. At all phases, the nature of the study and its procedures were explained prior to volunteers consenting to take part. Clinical levels of fear or anxiety were not assessed. Participants who completed the Time 1 questionnaire were sent a $20 book voucher, and an additional $15 book voucher was given on completion of the two subsequent questionnaires, given the length of the questionnaire (21 pages) and effort required to participate over the course of one year.


Time 1. From a sample of 185 university students who completed the FSS-II, 33 reported having very much or extreme levels of fear towards heights. Thirty of these students completed and returned the questionnaire, resulting in a response rate of 90%. A non-height-fearful comparison group of 47 people was also recruited from the 54 participants who reported no fear or very little fear towards heights. There was a majority of women in the fear (n = 28, 93%) and non-fear groups (n = 37, 79%), and an independent samples t-test showed that the fearful participants were older (M = 33.87, SD = 12.61) than the non-fearful participants (M = 28.32, SD = 10.28), t(75) = 2.10, p < .05, d = .50. As expected, the fearful group scored higher than the non-fearful group for the severity of height fear in terms of levels of anxiety (fear group: M = 56.60, SD = 18.58; non-fear group: M = 13.57, SD = 11.77), t(75) = 12.45, p < .05, d = 2.95, and avoidance (fear group: M = 13.23, SD = 6.20; non-fear group: M = 2.38, SD = 2.97), t(75) = 10.32,p < .05, d = 2.44. The fear severity scores for the height fear group were higher than those obtained by previous research with students with height fear (e.g., Menzies & Clarke, 1993b; Menzies & Parker, 2001), but comparatively lower than those obtained with clinical acrophobic samples (e.g., Cohen, 1972, cited in Antony, 2001; Menzies and Clarke, 1995a). However, unlike previous studies by Menzies and Parker (2001) and Taylor et al. (1999), the fearful group had higher levels of trait anxiety (M = 47.17, SD = 9.44) compared to the non-fearful group (M = 39.36, SD = 9.32), t(74) = 3.53, p < .01, d = .84.

Times 2 & 3. At Time 2, there were 24 responses from the fear group (80% response rate) and 37 from the non-fear group (79% response rate). At Time 3, there were 26 responses from the fear group (87% response rate) and 40 from the non-fear group (85% response rate). Overall, 22 participants in the fear group and 33 in the non-fear group completed all three questionnaires. There were no significant differences in fear severity or trait anxiety between those who did and did not complete all three questionnaires. All 30 fearful participants were included in analyses regarding pathway frequency, but only those who completed all three questionnaires (i.e., n = 22) were included in assessing the stability of fear onset pathway descriptions over time.


The questionnaire used in the present study was a compilation of instruments. Participants in the height fear group completed a questionnaire that enquired about past experiences with heights and possible pathways to height fear. The non-fearful group's questionnaire enquired generally about past experiences with heights. Both groups completed measures of height fear severity and trait anxiety.

Origins Questionnaire-Second Edition (OQ-II). The OQ-II (Menzies & Parker, 2001) identifies both associative and non-associative experiences related to a feared stimulus. In terms of associative pathways, respondents are asked to describe any direct, vicarious, or informational events experienced prior to the development of their fear, and each such event is explored in terms of how much fear and pain was experienced and whether excessive fear of heights developed after the event. The OQ-II also asks about the level of fear prior to the events, whether similar non-noxious events had occurred before, and whether the perceived dangerousness of an event was subsequently revaluated. These questions have the potential to capture many of the fear acquisition theories such as latent inhibition, UCS inflation, and the role of non-specific stress in the dishabituation of fears. Finally, the OQ-II gives participants the option of describing themselves as having always been fearful of heights.

Based on the OQ-II, participants were grouped into one of nine fear onset pathways: classical conditioning, vicarious conditioning, informational conditioning, multiple associative pathways, non-conditioning traumatic event, always been this way, mixed pathways, cannot remember, and cannot classify. With the exception of the multiple associative pathways and always been this way categories, the classification criteria used were those adopted in previous studies (Menzies & Clarke, 1995a; Menzies & Parker, 2001). The multiple associative pathways category encapsulated onset events that involved more than one type of conditioning being experienced simultaneously, such as having a combination of classical and vicarious conditioning in the same event. The mixed pathway category applied to reports where both associative and non-associative events contributed to fear onset. The authors independently reviewed all of the 54 questionnaire responses received from Times 1 and 2, with 87% agreement in classifications. Cases of disagreement were discussed between the authors and decisions made. Menzies and Parker's (2001) modified version of the OQ-II was used for the non-fearful group, which asked about previous anxiety-provoking exposure to heights as well as other related UCS revaluation experiences, without making reference to height fear.

At Times 2 and 3, participants were asked to complete an additional page at the start of the OQ-II that asked whether they had an incident involving heights during the two retest intervals, to identify any intervening events that could account for changes over time. Participants were asked to rate how much their fear towards heights had improved or worsened since they last completed the questionnaire on a seven-point scale, using the anchors A (much worse), B (moderately worse), C (mildly worse), D (the same), E (mildly better), F (moderately better), and G (much better).

Acrophobia Questionnaire (AQ). The AQ (Cohen, 1977) consists of two scales that assess the severity of anxiety and avoidance associated with 20 situations involving heights. Anxiety is rated on a seven-point scale ranging from 0 (Not at all anxious) to 6 (Extremely anxious), and avoidance on a three-point scale of 0 (Would not avoid doing it), 1 (Would try to avoid doing it), and 2 (Would not do it under any circumstances). Total scores range from 0-120 for the anxiety scale and 0-40 for the avoidance scale. Test-retest reliabilities were r = .86 and r = .82 for anxiety and avoidance, respectively, over a three-month period (Baker, Cohen, & Saunders, 1973), and were .94 and .90 for anxiety and avoidance, respectively, over three months in the present study.

Trait Scale of the State-Trait Anxiety Inventory (STAI-T Form Y-2). The STAI-T (Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983) is a 20-item self-report measure that assesses trait anxiety. Items are rated on a four-point Likert scale from 1 (Not at all) to 4 (Very Much). The psychometric qualities of the STAI have been well established through its extensive use in clinical and non-clinical research. In the present study, test-retest reliabilities over 3 and 12 months were high, with r = .82 and .79, respectively. Internal consistency was high at [alpha] = .92.


Height Fear Acquisition

Table 1 shows the frequency of pathway classifications for the height fear group across the three time points. At Time 1, 8 (26.67%) respondents ascribed to one of the associative pathways, compared to 17 (56.67%) to non-associative pathways. Despite having more than twice as many participants ascribing to a non-associative pathway as to an associative pathway, a chi-square test for goodness of fit showed that this difference in proportions was not significant, [chi square](1) = 3.24, p = 0.07, 9 = 0.33.

There were no differences between participants who ascribed to associative and non-associative pathways in terms of AQ-Anxiety [t(23) = 1.04, p = 0.31, d = 0.46], AQ-Avoidance [t(22.94) = 0.59, p = 0.57, d = 0.26], and STAI-T scores [t(21.67) = 1.29, p = 0.21, d = 0.59]. Of the non-fear group, 83% had experienced a height-related associative event that could have led to the conditioning of height fear. This was significantly greater than the 60% in the fearful group who experienced such events, [chi square](1) = 3.91, p < 0.05, [phi]9 = 0.23. However, when examining the frequency for each type of conditioning event separately, the difference was only significant for the experience of vicarious events (27% vs. 77% of the fearful and non-fearful groups, respectively), [chi square](1) = 11.65, p < 0.01, [phi] = 0.39. The aversiveness of various associative experiences was also examined. Table 2 provides data for the subjective fear experienced during the various types of associative events. Independent samples t-tests revealed that the classical and vicarious learning events experienced by height-fearful participants were rated as more frightening than those experienced by non-fearful participants [t(39.78) = 4.41, p < 0.01, d = 1.40; t(42) = 2.64, p < 0.05, d = 0.98, respectively]. There was no difference in the aversiveness of informational learning events across the two groups, t(24) = 1.64, p = 0.11, d = 0.73.

In terms of other non-associative components of fear development, the potential role of non-specific stressors in triggering the dishabituation of height fear was investigated (e.g., physical and mental illness, financial hardship). A chi-square test of independence found no differences between the fearful and non-fearful groups in the proportion of participants whose fear-learning event had occurred during periods of stress, [chi square](1) = 0.35, p = 0.56, [phi] = 0.07. Examples of UCS inflation were identified in order to evaluate the relevance of this process to the development of height fear in the sample. Eight participants reported having revaluated the dangerous of heights or a related UCS, although all were revaluations of an event that was already the most etiologically significant event, so the reports of revaluation of previous events mainly involved an increase in fear severity rather than the emergence of persistent height fear.

Stability of Retrospective Recall of Fear Onset

The main purpose of the present study was to examine the stability of fear onset pathways reported at three points in time over 12 months. Data from the 22 fearful respondents who completed all three questionnaires were included in these analyses. Table 3 provides the data for the pathways ascribed to at Time 1, as well as the percentage of participants who reported the same pathway at Time 2. Over this three-month period, 82% of pathway ascriptions remained the same, while 4 out of 22 (18%) respondents provided inconsistent ascriptions of the pathway to their height fear. Table 4 provides the pathways ascribed to by participants at Time 2, as well as the percentage who reported the same pathway at Time 3. Over this nine-month period, 73% of ascriptions remained the same, while 6 (27%) had a change in pathway ascription. Three of these six participants also had changes in pathway ascriptions between Time 1 and Time 2.

Overall, out of a sample of 22, 7 participants contributed to 10 incidences of changes in pathway ascriptions at some point over 12 months. The cannot remember category was most frequently involved in changes of pathway ascriptions. Specifically, three participants who originally ascribed to a conditioning or non-conditioning traumatic pathway endorsed the cannot remember category at a subsequent retest. Additionally, two other participants who originally endorsed cannot remember about the onset of their fear subsequently ascribed to the always been this way category at retest. The multiple associative pathways category was involved in four incidences of pathway change. One participant who originally ascribed to the vicarious conditioning pathway at Time 1 reported an additional classical conditioning component to the same observational experience at Time 2, which led to a multiple associative pathways ascription. Another participant changed from an always been fearful ascription at Time 2 to the multiple associative pathways category at Time 3, as more etiologically-significant associative experiences were reported at retest.

The stability of participants' AQ and STAI-T scores over the three test periods were examined using three one-way repeated measures ANOVAs (see Table 5). For the fearful group, there were no significant differences between the three time periods on the AQ-Anxiety scale (Wilk's Lambda = .91, F(2,20) = 1.04, p = .37), AQ-Avoidance scale (Wilks' Lambda = .95, F(2,20) = .55, p = .59), and STAI-T (Wilks' Lambda = .96, F(2,19) = .40, p = .68). Similarly for the non-fearful group, no significant differences between the three time periods were found for AQ-Anxiety (Wilks' Lambda = 0.99, F(2,31) = 0.13, p = .88), AQ-Avoidance (Wilks' Lambda = 1.00, F(2,31) = 0.01, p =.99), and STAI-T (Wilks' Lambda = 0.85, F(2,31) = 2.74, p = .08).

The influence of any height-related intervening experiences during the 12-month retest period was also examined. The described events mostly resembled a non-conditioning traumatic exposure to heights or involved vicarious conditioning where the respondent personally witnessed another person being anxious or hurt because of heights. However, none of the five participants who reported an intervening event between Times 1 and 2 had a change in pathway ascription at Time 2. Similarly, none of the six participants who reported an intervening event between Times 2 and 3 had a change in pathway ascription at Time 3.

Discussion and Conclusions

Although the present study supported previous research in finding a larger number of non-associative than associative pathways to height fear, the proportion of respondents who ascribed to non-associative pathways (57%) was not significantly greater than those who ascribed to associative pathways (27%). This finding is consistent with previous research by Menzies and Parker (2001), which also found a higher but nonsignificant proportion of height-fearful participants ascribing to non-associative than associative pathways (54% vs. 36%, respectively). Such results are consistent with what is expected from the non-associative model, which argues that the non-associative pathway is the primary (but not only) pathway to fear for evolutionary-relevant fears such as heights (Poulton & Menzies, 2002b). The non-significant results may be due to insufficient statistical power, given the p value of 0.07 and medium to large effect size (phi-coefficient) of 0.33. This is comparable to the effect size of 0.38 obtained by Menzies and Clarke (1995a), who found that their sample of 129 clinically height-phobic participants had ascribed to non-associative pathways significantly more often than to associative pathways. Following Cohen's (1988) convention, detecting a medium effect through a chi-square analysis should ideally involve a sample size greater than 80. Clearly, with only 25 participants involved in the present analysis, the risk of Type II error is substantial. Indeed, having one additional participant ascribing to a non-associative pathway would be sufficient to achieve a significant finding at the .05 alpha level. A replication of this study using a larger sample would provide some clarification as to whether the null hypothesis is in fact true, or if it was due to insufficient statistical power as a result of low sample size.

The present study identified scenarios where each of Rachman's (1977) three pathways to fear were by themselves sufficient in conditioning height fear. In addition, several processes that are consistent with the associative framework were identified. Firstly, despite many of the non-fearful participants reporting having experienced associative learning events, these events were generally less aversive or frightening than those experienced by fearful participants. This is consistent with associative theories which propose a positive relationship between the aversiveness of a stimulus or punishment and the severity of the conditioned response (Rachman, 1977, 1991). Menzies and Parker (2001), on the other hand, found that non-fearful participants reported past height-related experiences that were on average more aversive than those reported by height-fearful participants. One possible explanation to reconcile these findings is that the present height-fearful group had higher trait anxiety levels than the non-fearful group. As a result, fearful participants may report more aversive events or perceive them as being more frightening than those with lower trait anxiety. In any case, differences in trait anxiety between the height-fearful and non-fearful groups may be a confounding factor for between-group comparisons.

A potential limitation of the study is that participants could endorse through answering a single question on the OQ-II that they had always been fearful of heights. The OQ-II instructs participants who endorse this option to not complete the rest of the questionnaire. This potentially leads to a premature exclusion of the etiological significance of any associative experiences. The relative ease of endorsing this option may also generate biases and inflate the frequency of always been fearful ascriptions. Future studies should instruct participants to complete the entire OQ-II when they report having always been fearful. This would enable a greater exploration of the influence of associative experiences for those who have already developed their fear through non-associative pathways, and may provide better insight into the process of habituation in terms of how fearful or non-fearful exposures may help to reduce dishabituated fears.

Reliability of Retrospective Reports

The present study identified various aspects of retrospective recall that are subject to change over time. Specifically, 18% (n = 4) of pathway ascriptions were unstable over three months, and 27% (n = 6) were unstable over nine months. This is comparatively less than the 46% (n = 39) change identified by Taylor et al. (1999) over 12 months for driving fear. As hypothesised, by adopting a more consistent method of retesting compared to Taylor et al.'s study, the level of change observed was notably reduced. When examining the proportions of pathway change that occurred throughout the study, the cannot remember classification accounted for half of the changes in pathway ascriptions observed. Specifically, two participants who originally endorsed cannot remember about the onset of their fear had subsequently ascribed to the always been this way category at retest. Additionally, three other participants who originally ascribed to a particular pathway had ascribed to the cannot remember category at a subsequent retest. The previous study by Taylor et al. (1999) could not definitively conclude that participants had either remembered or forgotten events during the retest period because such scenarios were only deduced from the changes in pathway ascriptions made on two different measures. The present study provides some clarification of this finding, and demonstrate the likelihood for people to either forget or remember past experiences within a one-year period.

In addition to the instability of pathway ascriptions, other aspects of retrospective recall in this study were observed to be unreliable over time, and in many cases, substantial variations in reports were observed even though the fear onset pathway remained consistent. For example, although one participant ascribed to the non-conditioning traumatic events pathway at both Times 1 and 2, the events described that led to these ascriptions were completely different. Specifically, the event reported at Time 2 occurred at a younger age and at a different location from that reported at Time 1. This suggests that the participant had remembered an earlier event during the retest interval, and thus shifted her belief as to what contributed to her fear. Despite a short three-month retest interval, participants appeared to have remembered much more of their past histories with heights. It is possible that participants had thought about their own scenario after the Time 1 test, or had asked family members about past height-related experiences, and had integrated new information which was subsequently described at Time 2.

Similarly, over a retest period of three months between Times 1 and 2, 32% (n = 7) of the fearful group had omitted events at Time 2 that were previously reported at Time 1, and 18% (n = 4) had reported additional events at Time 2 that were not reported at Time 1. For the non-fearful group, 30% (n = 10) of participants had omitted events on the Time 2 questionnaire that were previously reported at Time 1, and almost 70% (n = 23) had reported additional events at Time 2 that were not reported at Time 1. It is important to note that many of these changes were not detected through the analysis of fear onset pathways as they did not lead to changes in pathway ascriptions. On the other hand, pathway ascriptions also appeared to be quite sensitive to minor variations in participants' reports, which may inflate the level of retrospective unreliability. For example, a participant ascribed to the multiple associative pathways category at Time 1 as she reported that both herself (classical conditioning) and her brother (vicarious conditioning) were hurt from sliding down a cliff edge. At Time 2, she did not mention the vicarious component of the event, and was therefore classified into the classical conditioning category. In these cases, it was not instability of memory of past experiences that contributed to the change in pathway ascriptions, but simply the level of detail that was elicited by the OQ-II. Thus, the current procedure and classification criteria for ascertaining pathway ascriptions can be both under- and over-sensitive to the unreliability inherent in retrospective reports, which raises questions over the accuracy of pathway ascriptions in capturing fear onset.

The observed increase in the number of past experiences being reported in follow-up tests raises the possibility that the information obtained at the later retests reflected a more accurate picture of participants' past experiences. If so, future research could consider integrating similar follow-up tests in order to capture participants' experiences more fully. Replication of this study with a height phobic sample would have greater clinical relevance in relation to the role of onset events and subsequent experiences on the longterm development and maintenance of height fear. Further use of the OQ-II with a clinical sample would also provide broader information about the psychometric properties of this instrument. Assuming that those with clinical levels of fear and phobia report more extreme experiences of prior fear, panic, or stress, the reliability of their accounts may be less subject to the biases associated with the sensitivity of the classification method used in the OQ-II.

In summary, the present study provided greater insight into the origins of height fear and how past height-related events are reported. The findings demonstrated the limitations of associative theories insofar as aversive experiences were found to be both insufficient and unnecessary for the acquisition of height fear (Field, 2006; Poulton & Menzies, 2002a). Furthermore, the role of non-associative pathways in the acquisition of this evolutionary-relevant fear is comparable to that of traditional associative pathways. However, having identified substantial variations in participants' reports of past height-related experiences over time, the reliability of retrospective accounts for exploring fear onset was again questioned (Taylor et al., 1999). The present study provided further support for the limitations of retrospective recall, and identified several methodological shortcomings that future studies should aim to address in their enquiry into the onset of fears and phobias.


The research presented in this paper formed part of the first author's Master's thesis in the School of Psychology at Massey University. The authors acknowledge those who took part in the study. The study received financial support from the School of Psychology in addition to a Massey University Research Fund grant.


Antony, M. M. (2001). Specific phobia: A brief overview and guide to assessment. In M. M. Antony, S. M. Orsillo, & L. Roemer (Eds.), Practitioner's guide to empirically based measures of anxiety (pp. 127-158). New York: Kluwer Academic/Plenum Publishers.

Baker, B. C., Cohen, D. C., & Saunders, J. T. (1973). Self-directed desensitization for acrophobia. Behaviour Research and Therapy, 11, 79-89.

Cohen, D. C. (1977). Comparison of self-report and overt-behavioral procedures for assessing acrophobia. Behavior Therapy, 8, 17-23.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: L. Erlbaum Associates.

Davey, G. C. (1989). UCS revaluation and conditioning models of acquired fears. Behaviour Research and Therapy, 27, 521-528.

Davey, G. C. (1992). Characteristics of individuals with fear of spiders. Anxiety Research, 4, 299-314.

Davey, G. C. (1997). A conditioning model of phobias. In G. Davey (Ed.), Phobias: A handbook of theory, research and treatment (pp. 301-322). West Sussex, England: John Wiley & Sons Ltd.

de Jong, P. J., Muris, P., & Merckelbach, H. (1996). UCS inflation and human aversive autonomic conditioning. Biological Psychology, 43, 227-239.

Doogan, S., & Thomas, G. V. (1992). Origins of fear of dogs in adults and children: The role of conditioning processes and prior familiarity with dogs. Behaviour Research and Therapy, 30, 387-394.

Field, A. P. (2006). Is conditioning a useful framework for understanding the development and treatment of phobias? Clinical Psychology Review, 26, 857-875.

Geer, J. H. (1965). The development of a scale to measure fear. Behaviour Research and Therapy, 3, 45-53.

Graham, J., & Gaffan, E. A. (1997). Fear of water in children and adults: Etiology and familial effects. Behaviour Research & Therapy, 35, 91-108.

King, N. J., Clowes-Hollins, V., & Ollendick, T. H. (1997). The etiology of childhood dog phobia. Behaviour Research and Therapy, 35, 77.

King, N. J., Gullone, E., & Ollendick, T. (1998). Etiology of childhood phobias: Current status of Rachman's three pathways theory. Behaviour Research and Therapy, 36, 297-309.

Menzies, R. G., & Clarke, J. C. (1993a). The etiology of childhood water phobia. Behaviour Research and Therapy, 31, 499-501.

Menzies, R. G., & Clarke, J. C. (1993b). The etiology of fear of heights and its relationship to severity and individual response patterns. Behaviour Research and Therapy, 31, 355-365.

Menzies, R. G., & Clarke, J. C. (1995a). The etiology of acrophobia and its relationship to severity and individual response patterns. Behaviour Research and Therapy, 33, 795-803.

Menzies, R. G., & Clarke, J. C. (1995b). The etiology of phobias: A nonassociative account. Clinical Psychology Review, 15, 215-221.

Menzies, R. G., & Harris, L. M. (1997). Mode of onset in evolutionary-relevant and evolutionary-neutral phobias: Evidence from a clinical sample. Depression and Anxiety, 5, 134-136.

Menzies, R. G., & Parker, L. (2001). The origins of height fear: An evaluation of neoconditioning explanations. Behaviour Research and Therapy, 39, 185-199.

Merckelbach, H., Muris, P., & Schouten, E. (1996). Pathways to fear in spider phobic children. Behaviour Research and Therapy, 34, 935-938.

Mineka, S., & Oehlberg, K. (2008). The relevance of recent developments in classical conditioning to understanding the etiology and maintenance of anxiety disorders. Acta Psychologica, 127, 567-580.

Mineka, S., & Ohman, A. (2002). Born to fear: Non-associative vs associative factors in the etiology of phobias. Behaviour Research and Therapy, 40, 173-184.

Mineka, S., & Sutton, J. (2006). Contemporary learning theory perspectives on the etiology of fears and phobias. In M. G. Craske, D. Hermans, & D. Vansteenwegen (Eds.), Fear and learning: From basic processes to clinical implications (pp. 75-98). Washington, DC: American Psychological Association.

Poulton, R., Davies, S., Menzies, R. G., Langley, J. D., & Silva, P. A. (1998). Evidence for a non-associative model of the acquisition of a fear of heights. Behaviour Research and Therapy, 36, 537-544.

Poulton, R., & Menzies, R. G. (2002a). Fears born and bred: Toward a more inclusive theory of fear acquisition. Behaviour Research and Therapy, 40, 197-208.

Poulton, R., & Menzies, R. G. (2002b). Non-associative fear acquisition: A review of the evidence from retrospective and longitudinal research. Behaviour Research and Therapy, 40, 127-149.

Rachman, S. (1977). The conditioning theory of fear-acquisition: A critical examination. Behaviour Research and Therapy, 15, 375-387.

Rachman, S. (1991). Neo-conditioning and the classical theory of fear acquisition. Clinical Psychology Review, 11, 155-173.

Rescorla, R. (1968). Probability of shock in the presence and absence of CS in fear conditioning. Journal of Comparative and Physiological Psychology, 66, 1-5.

Spielberger, C. D., Gorsuch, R. L., Lushene, R., Vagg, P. R., & Jacobs, G. A. (1983). Manual for the State-Trait Anxiety Inventory (Form Y). Palo Alto, CA: Mind Garden, Inc.

Taylor, J. E., & Deane, F. P. (1999). Acquisition and severity of driving-related fears. Behaviour Research and Therapy, 37, 435-449.

Taylor, J. E., Deane, F. P., & Podd, J. V. (1999). Stability of driving fear acquisition pathways over one year. Behaviour Research and Therapy, 37, 927-939.

White, K., & Davey, G. C. (1989). Sensory preconditioning and UCS inflation in human 'fear' conditioning. Behaviour Research and Therapy, 27, 161-166.

Withers, R. D., & Deane, F. P. (1995). Origins of common fears: Effects on severity, anxiety responses and memories of onset. Behaviour Research and Therapy, 33, 903-915.

Edwin C. H. Chin, Massey University

Joanne E. Taylor, Massey University

Corresponding Author:

Joanne E. Taylor

School of Psychology,

Massey University

Private Bag 11-222

Palmerston North
Table 1. Frequency of Pathway Classifications for the Fear Group at
Initial Assessment (Time 1), Three Months Later (Time 2), and 12
Months Later (Time 3)

Pathway                                Time 1     Time 2     Time 3
                                      (n = 30)   (n = 24)   (n = 26)
Associative pathways
  Classical conditioning                  4          3          1
  Vicarious conditioning                  2          2          2
  Informational conditioning              1          0          0
  Multiple associative pathways           1          2          3

  Total associative pathways              8          7          6

Non-associative pathways
  Non-conditioning traumatic event        6          5          5
  Always been fearful                    11          9         13

  Total non-associative pathways         17         14         18

Other pathways
  Mixed pathway                           2          0          0
  Cannot remember                         2          3          2
  Cannot classify                         1          0          0

  Total other pathways                    5          3          2

Note. This data represents the total frequency of each pathway being
described at each time of testing, and does not reflect the
consistency of pathway ascriptions over time. For example, even
though there were two ascriptions to the vicarious conditioning
pathway at all three times of testing, it does not indicate that these
ascriptions were made by the same two participants.

Table 2. Mean Level of Fear Experienced During Associative Events

Event Type          Fear group            Non-fear group

                n    Mean (a)   SD     n    Mean (a)   SD

Classical       16   7.88       1.59   30   5.13       2.19
Vicarious       8    5.00       1.77   36   2.47       2.57
Informational   8    4.00       2.73   18   2.39       2.12

(a) Self reported level of fear rated on a scale from 0 (No fear)
to 10 (Worst fear I can imagine).

Table 3. Stability of Pathway Ascriptions Over Three Months Between
Times 1 and 2

                                                             % of
Pathway                           Pathway classification     cations
                                         (n = 22)            which
                                    Time 1       Time 1      stayed the
Associative                       n    % (a)    n    % (a)   same

  Classical conditioning          3    13.63    2    9.10    66.70
  Vicarious conditioning          2    9.10     1    4.54    50.00
  Informational conditioning      0    0        -    -       N/A
  Multiple associative pathways   0    0        -    -       N/A
  Non-conditioning traumatic
    event                         5    22.72    5    22.72   100.00
  Always been fearful             10   45.45    9    40.91   90.00
  Mixed pathway                   0    0        -    -       N/A
  Cannot remember                 1    4.54     1    4.54    100.00
  Cannot classify                 1    4.54     0    0       0
Total                             22   100.00   18   81.81

N/A = not applicable.

(a) Calculated as the percentage of the total sample of 22.

Table 4. Stability of Pathway Ascriptions Over Nine Months Between
Times 2 and 3

                                                             % of
Pathway                           Pathway classification     original
                                         (n = 22)            classifi-
                                    Time 2        Time 3     which
                                                             stayed the
Associative                       n    % (a)    n    % (a)   same

  Classical conditioning          2    9.09     1    4.55    50.00
  Vicarious conditioning          1    4.55     1    4.55    100.00
  Informational conditioning      0    0        -    -       N/A
  Multiple associative pathways   2    9.09     1    4.55    50.00
  Non-conditioning traumatic
    event                         5    22.72    4    18.18   80.00
  Always been fearful             9    40.91    8    36.37   88.89
  Mixed pathway                   0    0        -    -       N/A
  Cannot remember                 3    13.64    1    4.55    33.33
  Cannot classify                 0    0        0    0       N/A
Total                             22   100.00   16   72.73

N/A = not applicable.
(a) Calculated as the percentage of the total sample of 22.

Table 5. Mean (SD) Fear Severity and Trait Anxiety Over 12 Months

                             Time 1          Time 2          Time 3

Fear group (n = 22)
  AQ-Anxiety              55.68 (15.37)   57.27 (15.66)   58.82 (16.67)
  AQ-Avoidance (a)        12.27 (5.25)    13.14 (5.24)    12.68 (5.77)
  STAI-T (a)              47.67 (9.74)    46.37 (10.97)   47.67 (10.41)
Non-fear group (n = 33)
  AQ-Anxiety              11.97 (7.05)    11.64 (6.38)    11.24 (7.55)
  AQ-Avoidance             1.82 (1.89)     1.79 (1.65)     1.82 (2.05)
  STAI-T                  38.06 (9.79)    36.15 (9.23)    37.79 (11.98)

AQ = Acrophobia Questionnaire. SD = standard deviation.
STAI-T = Trait scale of the State-Trait Anxiety Inventory. AQ-Anxiety
scale: range 0-120. AQ-Avoidance scale: range 0-40. STAI-T: range 20-80.
(a) n = 21 due to one case of pairwise exclusion.
Gale Copyright: Copyright 2011 Gale, Cengage Learning. All rights reserved.