'Application of AI in the Medical Field - The Influence of Expertise and Causability on Risk- and Opportunity-Perception. A Survey Study.' (AsPredicted #87,412)
Author(s) This pre-registration is currently anonymous to enable blind peer-review. It has 3 authors.
Pre-registered on 2022/02/08 05:16 (PT)
1) Have any data been collected for this study already? No, no data have been collected for this study yet.
2) What's the main question being asked or hypothesis being tested in this study? Does the level of medical expertise influence people's risk and opportunity perception of the application of AI in the medical field? Does the level of the AI's causability determine the risk- and opportunity perception?
Hypothesis 1:
H1a: Opportunity-perception decreases with increasing medical knowledge.
H1b: Risk-perception increases with increasing medical knowledge.
Hypothesis 2:
H2a: Opportunity-perception increases with increasing Causability.
H2b: Risk-perception decreases with increasing Causability.
Hypothesis 3:
H3a: Opportunity-perception increases with increasing confidence in one's own AI knowledge.
H3b: Risk-perception decreases with increasing confidence in one's own AI knowledge.
3) Describe the key dependent variable(s) specifying how they will be measured. DV: Risk- and Opportunity-Perception
Participants will be presented with one of six scenarios, in which a practical application of AI in the medical field will be described from a first-person perspective. After reading the scenario, participants are asked to rate to which extend they perceive the application of AI in this context as a risk or opportunity. Risk/opportunity perception will be rated in six questions each on a five-point Likert-scale.
Other measurements:
AI-Knowledge
AI-Knowledge will be measured as a control variable. All participants will answer ten questions about AI (yes/no answers). The mean score of correctly answered questions will indicate their level of expertise regarding AI.
Confidence in AI-knowledge
For each of the AI-knowledge questions, participants will be asked to indicate how confident they are in their own answer on a six-point Likert-scale (I guessed – 50%, 60%, …, I'm sure – 100%).
4) How many and which conditions will participants be assigned to? Level of medical knowledge
Participants will either be assigned to the "Expert"-group or "No-Expert"-group based on their profession. Participants who work in the medical field will be assigned to the "Expert"-group.
Causability
Participants will randomly be assigned to either the high or low causability condition. The high causability group will be presented with a scenario in which a more elaborate explanation of how the AI came to its result is given (i.e. which symptoms were analyzed, the size of the database, the probability for the correctness of the suspected diagnosis).
The low causability group however will be presented with the same scenario, but instead of the explanation participants will be told that the AI finished its analyzes without describing them further. Participants in the low causability group will be given the suspected diagnosis as well.
5) Specify exactly which analyses you will conduct to examine the main question/hypothesis. Age, gender, education and the level of AI-Knowledge will be controlled in all analyzes. The models' homoscedasticity, assumption of normality and leverage will be tested using graphics.
H1a: Linear regression predicting the opportunity perception (OP) from medical knowledge (MK, dummy-coded).
OPi = beta0 + beta1*MKi + εI
H1b: Linear regression predicting the risk perception (RP) from MK
RPi = beta0 + beta1*MKi + εI
H2a: Linear regression predicting OP from Causability (CA, dummy-coded)
OPi = beta0 + beta1*CAi + εI
H2b: linear regression predicting RP from CA
RPi = beta0 + beta1*CAi + εI
H3a: linear regression predicting OP from Confidence in AI-Knowledge (CK)
OPi = beta0 + beta1*CKi + εI
H3b: linear regression predicting RP from CK
RPi = beta0 + beta1*CKi + εI
6) Describe exactly how outliers will be defined and handled, and your precise rule(s) for excluding observations. Participants are required to be at least 18 years old and to be fluent in German in order to participate in this study.
7) How many observations will be collected or what will determine sample size? No need to justify decision, but be precise about exactly how the number will be determined. A survey period of three weeks is set. A precise sample size is not specified, all participant data gathered within three weeks will be comprised in the data set.
8) Anything else you would like to pre-register? (e.g., secondary analyses, variables collected for exploratory purposes, unusual analyses planned?) Exploratory analysis: Is the causability effect on the general risk/opportunity perception mediated by the participant's level of expertise?
