I recently wrote a post on using video in the most effective treatment we have for social anxiety ... "Treating social anxiety disorder: video (and still) feedback (6th post)".  Typically with social anxiety there are several "layers" to a sufferer's fears about potential negative judgements from others. For example they might be anxious that 1.)  They will blush.  2.)  Other people will notice that they are blushing.  3.)  They will then be judged negatively for blushing.  CBT treatment aims to reduce this anxiety by showing sufferers that their fears are exaggerated and that the methods they have developed for managing their difficulties are mostly making the problem worse.  David Clark, the key figure behind the development of this CBT approach, comments that "You are unlikely to get people to buy into these ideas just by talking ... it seems that it is actively experiencing & feeling this stuff that helps." 

Early in the course of treatment, therapists should aim to set aside a full session for an important videoed role play experiment.  Typically this session would occur after assessment interviews and after an individualised model has been developed with the client tracing out on a "flow chart" how their social anxiety episodes usually develop.  The videoed safety behaviours experiment aims to show clients in a very practical, experiential way that the methods that they are using to manage their anxiety (self-focused attention, self-evaluation, and 'safety behaviours') do NOT result in a better 'presentation' to others.  In fact these 'coping techniques' tend to make the situation worse, both for the sufferers' internal experience and for how they come across socially.  This is a big claim, and backing it up with genuine videoed evidence can be game-changing.

The "Social anxiety safety behaviours video experiment form" (Word doc or PDF file) can be used to pre-plan what the client will do & what will be assessed in the two stages of the experiment.  If one is using an assistant/stooge then they are NOT told beforehand what the client's particular fears are and they are asked to treat the role play as a "normal interaction with a stranger".  After the first part of the exercise the assistant is asked to jot down how they experienced the meeting ... "What was your impression of this person (the client)?  How did you find the meeting with them?" At this point the assistant hasn't been informed what the client is specifically worried about.  They are then asked to make more specific ratings of the client's particular concerns e.g. how was the quality of the conversation and/or how intense did the anxiety signs appear?  

Meanwhile, in the first stage of the experiment, the client has been asked to really focus on themselves, how they're feeling, and how they sense they're coming across (high levels of self-focused attention & self-evaluation) and to maximise their use of key pre-agreed safety behaviors (this might, for example, involve them tracking/planning their conversation to constantly monitor if it is witty and interesting enough, or to do one's best to avoid eye contact and keep still, etc).  It is also agreed beforehand what it will be most personally relevant to assess.  This might include both how anxious the client feels and how anxious they estimate they are looking.  It might include objective predictions about how bad anxiety signs will be (e.g. expected blushing severity pinpointed on a paint colour chart, predicted hand trembling or extent of shaking voice physically demonstrated by the client, etc).  It might also include predictions about general appearance and about various aspects of social competence.  The "Negative self-portrayal scale" can often be helpful in highlighting what to assess, as too can information from the "Social anxiety flow chart" - both these measures should have already been completed before getting to the treatment session where we introduce a video experiment.  Similarly the flow chart and measures like the "Subtle avoidance frequency examination (SAFE)" can also highlight what safety behaviours the client is to maximise and minimise during the two phases of the experiment.  

Now one runs through the first phase of the videoed experiment with high emphasis on the client self-focusing, self-evaluating, and using safety behaviours. The assistant/stooge then answers the general questions about their experience of the interaction (before knowing anything about the client's particular fears) ... see the "What was your impression of this person?  How did you find your interaction with them?" side of the "Social anxiety safety behaviours video experiment form" ... and then answers the more specific 0 to 100 severity level questions on the other side of the form.  It may be helpful for the assistant to leave the room for a few minutes now.  The client is asked to answer the phase one 0 to 100 severity questions.  They are then reminded that now, in the coming second phase of the experiment, the request is to reverse their normal methods for coping with their anxiety.  For example, one can ask them to "Try to be much more as you would with someone you’re close to & feel comfortable with, body posture more open, focused on the conversation not oneself, both more interested in the other person & more self-disclosing."  See the recent paper by Voncken & colleagues - "Socially anxious individuals get a second chance after being disliked at first sight: The role of self-disclosure in the development of likeability in sequential social contact" - for more on this.  They are also asked to stop using their typical safety behaviours.  

Now the assistant/stooge is brought back in for the second phase of this videoed experiment.  They and the client have a further short five minute or so conversation.  This time however the client is doing their best to focus out and really get involved with the conversation, while dropping their typical safety behaviours as much as possible.  At the end of the conversation, the assistant is again asked to rate the various pre-identified targets (e.g. how anxious the client appeared, etc) on the 0 to 100 severity scale.  They are thanked and leave.  The client now also rates the various items using the 0 to 100 scale.  The final section of the experiment involves looking at the videos of the first and second conversations and repeating the 0 to 100 ratings but this time from the objective record of what the client genuinely looked like.  Note when the client rates how they look on the video, it's important to ask them to do this "As if the person on the video is a stranger ... ".  This is because it seems possible that when the client observes themselves on the video, they may start to feel uncomfortable & self-conscious.  What we need to guard against is the possibilty that they then rate, for example, how anxious they think that they're looking on the video by estimating how anxious they are now feeling looking at the video. The "Assess as if a stranger" instruction is to try to reduce this "cross-contamination".  This is probably the first time the client has ever seen how they actually appear in this kind of social situation with a stranger ... fascinating & potentially very powerful. 

This experience can be hugely important in strongly challenging the client's beliefs that they need to self-focus & use a cluster of safety behaviours in order to come across acceptably in this kind of social interaction.  Very often they will have mind-changing experiences ... for example that they actually appeared more natural, more at ease and socially more competent when they dropped their safety behaviours and focused on spontaneous, open conversation.  They may well, to their surprise, note that they in fact felt more relaxed in the "exposed" second phase of the experiment (with minimised safety behaviours) than in (for them) the more normal, guarded first phase of the experiment.  If they say something like "Surely the main reason I come across better in the second phase of the experiment is simply down to familiarity, a sort of practice effect?"  One can reply something like "I strongly suspect that isn't the main explanation for why the second phase was so improved.  If it was, then a third conversation should be even easier again.  How about we do a third videoed conversation now, but this time reinstate all the self-focus, self-assessment and maximised safety behaviours."  If needed, one then goes through with a third phase of the experiment.  

Note that social anxiety sufferers often carry a negative image/felt sense of themselves that they take into new social interactions and that makes the interactions harder.  It can be helpful to get them to really note how they actually look on the video record and substitute this more accurate, realistic image of themselves for their exaggeratedly negative one when they go into future challenging social situations.  One has probably got to the end of the therapy session by now.  It can be helpful to encourage the client to explore dropping safety behaviours and focusing on conversations (not themselves) over the next days.  Their experience of the videoed experiment can be reviewed again at the next appointment.  Therapy typically then continues with attention training, a series of between-session behavioural experiments and, probably at some stage, then v's now discrimination training & rescripting of early socially traumatic memories.