Table 2.

Summary of significant individual differences effects reported in Bayesian word problems presenting normalized information or natural frequencies.

	Numeracy/education	IQ-raven	CRT I	Thinking disposition
NORMALIZED VERSIONS*
Chapman and Liu, 2009	No
Siegrist and Keller, 2011	Yes/No a
Hill and Brase, 2012	No
Garcia-Retamero and Hoffrage, 2013	Yes
Johnson and Tubau, 2013	Yes/No a
Lesage et al., 2013			No
Sirota et al., 2014a		Yes	No	Yes/No b
Ayal and Beyth-Marom, 2014	Yes c
McNair and Feeney, 2015	Yes/No d	Yes		No e
NATURAL FREQUENCIES
Brase et al., 2006	Yes
Chapman and Liu, 2009	Yes
Sirota and Juanchich, 2011	Yes		Yes
Siegrist and Keller, 2011	Yes/No f
Hill and Brase, 2012	Yes
Garcia-Retamero and Hoffrage, 2013	Yes
Johnson and Tubau, 2013	Yes/No g
Lesage et al., 2013			Yes
Sirota et al., 2014a		Yes	Yes	Yes/No b

Note that variation exists between the specific context and numbers used across studies, as well as specific measures and criteria used to determine low vs. high performers (see text for additional details, and original articles for full problems and explanations).

^*It is important to note that YES with normalized versions does not imply “good” reasoning, with most higher ability participants typically below 30% correct response.

^ICRT, Cognitive Reflection Test (Frederick, 2005).

^aYES with simple versions; NO with complex versions (floor effect).

^bYES with REI (rational-experiential inventory; rational thinking); NO with CAOMTS (actively open-minded thinking).

^cInformation was normalized, but problems manipulated to require only simple single-step arithmetic.

^dHigher numerate benefited more from causal manipulation used in Krynski and Tenenbaum (2007).

^eNO with REI.

^fYES in study 1; NO in study 2 (though clear trend).

^gYES with complex text; NO with short, simple text.