'Association implies Causation? - May 2022' (AsPredicted #98,012)
Author(s) This pre-registration is currently anonymous to enable blind peer-review. It has 2 authors.
Pre-registered on 2022/05/24 07:33 (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 statement "X is associated with Y" imply that "X causes Y"? While the statement "X is associated with Y" is intentionally used in scholarly papers to _not_ create a statement of causality, nevertheless it may _imply_ that the intention on the part of the speaker/writer is that the first item in the sentence causes the second item.
Participants will be given sentences of the general form "[X] is associated with [Y]", and a forced-choice binary response of whether this means that "[X] causes [Y]" or "[Y] causes [X]".
We will use several variations on this:
Study 1A [Basic]: We will use the sentence construction 'X is associated with Y' and give the binary choice 'X causes Y' / 'Y causes X'. X and Y will be nonsense words such as Denoglin and Fembik.
Study 1B [Basic]: Same as Study 1A, but with pure symbols [e.g. X, Y]
Study 2A [Directionality]: We will use the sentence construction 'X is associated with [higher/lower/--] [risk/probability/--] of Y' and the binary choice 'X causes higher/lower/-- Y' vs. 'Y causes higher/lower/-- X'. X and Y will be nonsense words such as Denoglin and Fembik.
Study 2B: Same as 2A but with pure symbols [e.g. X, Y]
3) Describe the key dependent variable(s) specifying how they will be measured. If the presented sentence had [X] before [Y], we code the binary causality judgement as '1' if it has [X] before [Y].
For example, if the sentence was "Denoglin is associated with Bomvik", and the participant chose "Denoglin causes Bomvik" over "Bomvik causes Denoglin", this is coded as a '1', otherwise as '0'.
Our dependent variable within each study, and within question type (increase, decrease, neutral) is the proportion of people who choose responses coded as '1'.
4) How many and which conditions will participants be assigned to? There will be 2 studies, each with 2 sub-conditions.
Study 1A will use the sentence construction 'X is associated with Y' and give the binary choice 'X causes Y' / 'Y causes X'. X and Y will be nonsense words such as Denoglin and Fembik. There will be 3 questions of this sort.
Each participant will see all questions. Question order, answer order, and specific word order will be randomized.
Study 2A is an expansion of Study 1A, and adds directionality. We will use the sentence construction 'X is associated with [higher/lower/increased/--] [risk/probability/--] of Y' and the binary choice 'X causes higher/lower/-- Y' vs. 'Y causes higher/lower/-- X'.
There will be 7 sentences total: "X is associated with Y", "X is associated with an increased probability of Y", "X is associated with a decreased probability of Y", "X is associated with an increased risk of Y", "X is associated with a decreased risk of Y", "X is associated with a higher risk of Y", "X is associated with a lower risk of Y".
Each participant will see all questions. Question order, answer order, and specific word order will be randomized.
Studies 1B and 2B are the same as 1A and 2A, but using symbols [X,Y, ...] instead of nonsense words [Flembin, Epiprex, ...]
5) Specify exactly which analyses you will conduct to examine the main question/hypothesis. Within each study, we compare the proportion of participants with responses coded as "1" (first item causes second item) to 0.5 (random), using a one-sided test of proportions. We correct for multiple hypothesis testing using the Holm-Bonferroni method.
6) Describe exactly how outliers will be defined and handled, and your precise rule(s) for excluding observations. We will exclude participants who fail to answer simple catch questions (e.g. answering 'describe what this study was about' with 'fine, thanks').
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. We will recruit 100 participants per study (=400 participants total). Pilot studies suggest 'X is associated with an increase in Y' implies "X causes Y" with an effect size detectable with 30-40 participants, but we want to also be able to detect effects that are smaller than that in the neutral conditions.
8) Anything else you would like to pre-register? (e.g., secondary analyses, variables collected for exploratory purposes, unusual analyses planned?) Nothing else to pre-register.
