Measuring the Emotional Experience of Visitors to Dark Sites

While there is clearly a need for more data on visitor experience to museums, including dark sites, there is a need for more effective and efficient data collection methods. Data collection should be done in such a way that it does not hamper the visitor experience while being as informative as possible. Additionally, researchers strive for a combination of both subjective and objective measures to obtain a more rich and complete data set. In this post I will describe a proposed method of collecting both objective and subjective data regarding visitors' emotional experience developed by Doreen and I.

First I will describe what type of emotion data we will be able to collect. As I mentioned earlier we will want to collect both subjective and objective measures. Subjective measures are explicit measures, commonly in the form of self report measures. Self report emotion measures simply ask the individual to indicate how they are presently feeling (state emotion) or how they generally feel (trait emotion). There are several emotion self report measures which can easily be completed by the visitor. We can also create a new self report questionnaire for our purposes using items from existing self report measures. That way we can make sure we collect data for specific items of interest while limiting the length of the questionnaire, increasing the likelihood visitors complete the questionnaire.

Perhaps more interesting is how we can collect objective emotion data. This is where my research and content from my previous blog posts come in. If visiting dark sites have an emotional effect on visitors, it may very well temporarily bias their cognitive functions, such as their memory, interpretations, and perceptions. I am particularly interested in perceptual biases, where my research investigates biases associated with the perception of neutral or ambiguous faces. If dark sites illicit powerful emotional experiences, we may be able to assess this by determining whether visitors demonstrate any negative cognitive biases following their visit. To assess this we would use versions of the emotion recognition training task. Visitors would only complete what is traditionally the first part of the task: visitors would be presented a range of images of an individual portraying emotional expressions ranging from happy to a negative emotion (anger, disgust, fearful, sadness) including images portraying neutral or ambiguous expressions. Based on how many of these images visitors rate as a negative emotion rather than happy, we could determine whether they are biased towards more negative perceptions, and identify the specific negative emotion associated with the bias. This is an objective measure given participants are not explicitly being told what the task is assessing.

A negative perceptual bias where more faces are perceived as sad rather than happy.

Another way we can collect objective emotion data would be using the emotion recognition software described in my last post. If it was possible to set up a live video feed by placing video cameras at various locations throughout the sites, we would be able to collect emotional data using software such as FaceReader or CERT would allow us to collect emotion data in real time. In fact, the software could do this while maintaining the anonymity of the visitor, by simply recording the response (or emotional data) without recording or storing the image of the visitor. That means that using the live video in conjunction with the software, we could collect information about emotional expressions portrayed by visitors, giving us insight into their emotional experience at that very moment. Again this is especially effective given it would give us data on their objective emotional experience rather than subjective, self report data. If a live video feed is not an option, video can be recorded and then sent to Emotient. However doing so would require additional consent to be obtained from the visitors given their image would be stored. As such, a live video feed would be more effective.

Finally, it is important to consider the delivery of these measures. Data collection is most effective when it is done efficiently and with little time commitment to the visitor. Additionally we must remember that for emotion data to be most accurate it should be collected before, during, or immediately after a visit, preferably before the visitor leaves the site. In order to accomplish this in order to collect the emotional data described above, we turn to the audio guide.

Audio guides are available at most museums and many dark sites and offer us with a very efficient and effective way of collecting data. While traditional audio guides were generally limited to simple number input and the presentation of audio recordings, audio guides have seen important technological advances over the last couple of decades.

Newer audio guides are now able to show full colour images and text, and allow visitors to input more varied information. The physical hardware for audioguides has substantially changed as well, with some museums now using Apple iPhones to deliver the audio guide program to their visitors. We believe using an audio guide with features similar to a smart phone, such as full colour images, extensive input options, as well as front facing cameras, would be the most effective way to collect emotional data. Visitors could complete self report measures and questionnaires presented by the device's screen and quickly input their response using the touchscreen functionality. Given that many popular smart phones now have front facing cameras, these could be used to capture the live video necessary for objective emotion data collection mentioned earlier. Attention can be directed to the screen when displaying colour images. At those moments when visitors are reacting to what they are viewing on the screen, the font-facing camera can be used to deliver live video to emotion recognition software and their expression data can be recorded. Finally, the emotion recognition training task has already been adapted for smartphone use and can run on the same smartphone as the audio guide before and after the start and end of the visit. In addition to these emotion measures, we will collect sociodemographic data using a visitor survey developed by Doreen as part of her PhD.

Therefore, using a smart phone audio guide as the data collection method, we would be able to obtain rich data regarding the emotional experience of visitors to museums and dark sites. Given the data collection would be integrated into their audio guides, data would be collected easily and efficiently, not requiring the visitor to use an external data collection device and would be a practical way for them to complete questionnaires and surveys. Given that many museums and sites are already using such devices as audio guides, developing the software to combine the existing audio guide content to include these measures would not require too much further effort and would be extremely effective for data collection. This offers museums and dark sites a novel and effective way of collecting emotional and sociodemographic data about their visitors in an unprecedented way and allow us to better understand the emotional experiences related to visiting dark sites in the future.