Abstract

Measures of sexual interest have been fraught with issues, such as being susceptible to
manipulation and not easily portable. Indirect measures aim to overcome this by being less
transparent, and thus supposedly harder to manipulate (Schmidt, Banse, & Imhoff, 2015).
However, as many exist, it is hard to know which is the most effective. Due to the variance in
the reliability and validity of current measures, a more comprehensive and easy-to�use method of assessing sexual interest is required. The present research attempts to solve
this problem by using an emerging indirect measure - namely, an approach-avoidance task
(AAT) - to measure sexual interest, implemented via a tablet PC. This thesis is original in
both applying approach-avoidance procedures to the measurement of sexual interest, as well
as implementing AATs on a tablet PC.
Across six studies, the tablet AAT was rigorously tested, varying the exact methodology,
including the instructions and stimuli, in order to determine the most optimal version. The
first two studies used explicit instructions whereby participants had to evaluate the stimulus
content (i.e., sexual attraction). The first study demonstrated that the AAT was somewhat
effective at assessing age-specific preferences, but not sex-specific preferences, which was
unexpected. The second study (which did not include child images) mainly demonstrated
results that were contrary to what was expected (i.e., more avoidance for preferred images).
Subsequent studies used implicit instructions whereby a directional cue indicated swipe
direction. The third study's results were not in line with predictions, but this was most likely
an artefact of the design. The fourth study also had mixed results, however, it appeared that
the 'response time' was more in line with predictions, compared with the 'movement time'
variable. The unexpected results from the fourth study (which largely replicated the
methodology from another approach-avoidance that produced expected results by Piqueras-
ii
Fiszman, Kraus, & Spence, 2014) may have been due to an issue with the stimuli. Therefore,
a survey was conducted for the fifth study to assess the attractiveness of the stimuli. These
ratings were then used to reanalyse the data from the prior studies to determine whether the
sexual attractiveness (or lack thereof) of the stimuli had an impact on the approach-avoidance
reactions. However, it was found that sexual attractiveness did not have an effect. The sixth
and final study aimed to test whether the equipment (tablet PC) was inappropriate for
conducting an AAT, or whether an AAT was inappropriate method for assessing sexual
interest. This was done by comparing the tablet PC AAT to a traditional joystick AAT. This
sixth study also produced results that were not in line with predictions.
The discussion chapter of this thesis thoroughly explores potential explanations as to why the
application of approach-avoidance procedures to sexual interest was unsuccessful. This
covers methodological considerations, such as the equipment, stimuli, instructions, and study
sample, as well as wider considerations, such as the ‘file drawer’ issue and application of
open science. The thesis concludes with recommendations for future research, including a
discussion surrounding the utility (or futility) of continuing to research the AAT as a measure
of sexual interest.