Political Humor and Anxiety: Measure of Anxiety (Update #3)

Since my previous posts tackled the link between political humor and anxiety, the next question in my literature review is straightforward: how do we measure anxiety?

As a non-psych major who typically studies government and economics, I had to do some digging. After reviewing the literature, anxiety can be split into two categories (state and trait) and is typically measured by either survey methods or biological measures.

  1. State Anxiety: State anxiety measures how anxious the participant is in a particular moment
  2. Trait Anxiety: Trait anxiety measures how anxious the participant is in general as part of their personality

Survey data

  • Anxiousness is associated with neuroticism, one of the Big Five personality characteristics in psychology. As such, it can be measured by rating neuroticism-associated statements like “I get stressed out easily” or “I am relaxed most of the time” on a scale from Almost Never to Almost Always (numerically represented as a 4 point scale), then tallying up the responses to get a number to represent trait anxiety. Likewise, state anxiety statements like “I feel tense” or “I feel calm” can be rated on the same scale to measure how anxious someone is in that moment.
  • Pros and Cons of survey data:
    • Survey data is much easier to collect, since it would be via a survey form and wouldn’t even require that participants are in the lab.
    • Survey data can be easily quantified by the numerical scale
    • However, there are some validity concerns with this measure. For example, participants could give the answer they think the experimenter wants, be poor judges of their own anxiety, or give arbitrary or incomplete responses.

Physiological measure for state anxiety:

Anxiety typically creates higher levels of physiological arousal, which is associated with a more attentive and excited state:

  1. Skin conductance: The electrical response of nerves gives quantitative measure of how intensely a participant is experiencing an emotion. Greater response means that the participant is more excited, less response indicates the opposite.
  2. Heart rate: Heart rate initially drops upon exposure to a stimulus, but then accelerates faster if the participant is excited by the stimulus.
  • Pros and Cons of physiological measures:
    • While physiological data is harder to collect, since it requires that the participant is in the lab and equipment must be attached to them, it also circumvents some of the issues with traditional survey data since participants cannot lie about their own heart rate or give misleading information.
    • There are other hazards, such as equipment malfunction or participants jostling equipment, but this can be corrected manually or omitted.
    • Heart rate also accelerates quickly after the presence of humor, so it may be hard to tell whether anxiety or humor is causing a spike in heart rate

Implications for experimental design:

Since I need data about the participant’s partisan affiliation and political ideology, I’ll be administering a survey anyway to gather this data and I could add in questions about trait anxiety to control for the participants that are more anxious in general. My real question is whether or not to measure state anxiety with survey methods in addition to the BIOPAC sensors for heart rate. Currently, I’m leaning towards including a state anxiety  pre- and post-test to disentangle the effect of humor and anxiety. One alternative is adding an apolitical humor treatment to compare to the political humor treatments.

Programming update:

  • PRAW continues to return a 403 error message, which I’m still working on. The authentication process is not very evident, so I’ve been looking back through the documentation to figure out my issue. It might be an issue with my laptop, so I’ll check it on another machine.
  • My application for a Twitter developer account was approved and made into a team account. This was a surprisingly thorough process, which asked about my intent, my specific use of the Twitter API, and how it would affect users. In addition to an online application, I had to reply to two emails with detailed responses. This makes sense, since Twitter has been receiving a lot of criticism for bot accounts abusing the platform and is likely wary of people asking for developer accounts.
  • After the developer account was approved, I wrote code in R using the twitteR package to pull the top 20 Tweets from the last week. It works fairly well, but needs some tweaks to pull images directly.