The researchers investigated one factor that could contribute to the spread of misinformation: whether people are sensitive to important cues to the degree of consensusgeneral agreement about a conclusionwhen assessing a set of arguments on a specific topic. In particular, a person , tested whether people are more convinced by a conclusion that has "true consensus" (the conclusion of each argument is based on independent primary sources) than when it has "false consensus" (the conclusion of each argument is actually based on only one primary source).

In the experiment, participants were presented with fictional news articles about the Japanese economy, and were asked to rate their confidence in the conclusion that "Japan's economy will continue to improve". Across three conditions, Yousif et al. (2019) manipulated the sources that were cited in the news articles. Participants were randomly allocated to each condition in a between-participants design.

The three conditions were:

True-consensus condition, in which there were four positive news articles (supporting the conclusion) and one negative article (rejecting the conclusion), all citing different primary sources. False-consensus condition, in which there were four positive news articles and one negative article, but the positive articles all citied the same primary source. The negative article cited a different source. No-consensus condition, in which there was only one positive and one negative news article, each citing a different primary source. Yousif et al. (2019) expected that confidence ratings would be higher for the true-consensus condition than for the no-consensus condition, but the key question was what would happen in the false-consensus condition. If people care only about the number of primary sources, then confidence in the false-consensus condition should match that of the no-consensus conditionthere are two primary sources in each condition (one positive and one negative). However, if instead, people focus on just the number of secondary sources, then confidence in the false-consensus condition should match that of the true-consensus conditionboth have four positive news articles and one negative article.

To check that participants read the news articles properly and attended to the citations in the first place, at the end of the experiment participants were shown a list of 10 sources and were asked to indicate which had been cited in the articles.

The dataset has been analysed and the following results were produced..

The results of ANOVA indicate sufficient evidence to conclude that there is a significant difference

between the three group means, F (2, 297) = 52.099, p < .001, 2= .260.

The results of the Levene's test for equality of variances indicates no violation of the assumption of

the equality of variances of the three groups, F (2, 279) = 1.283, p = .279 > .05.

The Q-Q plot of the residuals indicates no violation of the normality assumption, as all the plotted

points are along the straight line drawn at the center.

The Tukey's post hoc tests indicate that,

The pairs of groups "False" and "No" (p < .001), "No" and "True" (p < .001) differ significantly in group

means. The pair of groups "False" and "True" (p = .065 > .05) does not differ.

Explanation:

The F-test in ANOVA for homogeneity correction is the test for comparing the three group means.

The test for equality of variances and Q-Q plots indicate the adequacy of the assumption under

ANOVA. The Tukey's Post-Hoc tests indicate the significance of the difference between the possible

pairs of the three groups of the independent variable.

Now based on the above results

What conclusion do you draw and why? That is, do people seem to care about the number of primary sources that support a conclusion, or just the number of secondary sources? Explain your answer