Michał Krawczyk

The Search for Significance: A Few Peculiarities in the Distribution of P Values in Experimental Psychology Literature

In this project I investigate the use and possible misuse of p values in papers published in five (high-ranked) journals in experimental psychology. I use a data set of over 135’000 p values from more than five thousand papers. I inspect (1) the way in which the p values are reported and (2) their distribution. The main findings are following: first, it appears that some authors choose the mode of reporting their results in an arbitrary way. Moreover, they often end up doing it in such a way that makes their findings seem more statistically significant than they really are (which is well known to improve the chances for publication). Specifically, they frequently report p values “just above” significance thresholds directly, whereas other values are reported by means of inequalities (e.g. “p<.1”), they round the p values down more eagerly than up and appear to choose between the significance thresholds and between one- and two-sided tests only after seeing the data. Further, about 9.2% of reported p values are inconsistent with their underlying statistics (e.g. F or t) and it appears that there are “too many” “just significant” values. One interpretation of this is that researchers tend to choose the model or include/discard observations to bring the p value to the right side of the threshold.

Do gender and physical attractiveness affect college grades

Abstract In this study, data on grades awarded for bachelor and master theses at a large Polish university were used to identify possible discrimination on gender or physical attractiveness. The focus is on the gap between the grades awarded by the advisor (who knows the student personally) and the referee (who typically does not, so that gender is less salient and attractiveness is not observable). This provides an excellent control for actual quality of the work, which is often problematic in previous literature. Observations on nearly 15,000 students are available, of which some 2600 also have their physical attractiveness rated and are included in the analysis. It can thus be checked if the advisor–referee grade gap depends on students’ gender and attractiveness, also in interaction with their genders. Based on stereotypes discussed in the literature one can hypothesise that females are treated relatively favourably by advisors (and males by referees). Likewise, it can be expected that physically attractive individuals’ advisors award them relatively high grades. Overall, we observe some evidence that females indeed get relatively high grades from advisors and no evidence of influence of physical attractiveness.

Are all researchers male? Gender misattributions in citations

I screen academic literature for cases of misattribution of cited author’s gender. While such mistakes are overall not common, their frequency depends dramatically on the gender of the cited author. Female scholar are cited as if they were male more than ten times more often than the opposite happens, probably revealing that citers are influenced by the gender-science stereotype. The gender of the citing author and the field of study appear to have only limited effect.

Do pirates play fair? Testing copyright awareness of sports viewers

ABSTRACT Ethical norms are believed to be followed more loosely on the Internet than in the ‘real world’. This proposition is often evoked to explain the prevalence of so-called digital piracy. In this study, we provide evidence from a vignette experiment that contradicts this claim. Analysing the case of sports broadcast, we compare explicitly the ethical judgement of legal and illegal sharing in the offline and online contexts. We find that the norms concerning legality, availability of alternatives and deriving material benefits from sharing content do not differ substantially between the virtual and real worlds. We also test explicitly for the role of legal awareness and find that emphasising what is prohibited (copyright infringement) is less effective than focusing on what is permitted (fair use) in reducing the disparity between legal and ethical norms.

Pirates in the Lab: Using Incentivized Choice Experiments to Explore Preference for (Un)Authorized Content

We report a laboratory experiment aimed at investigating factors affecting choice between different versions of a full-length movie. In particular, we estimate the willingness to pay for a legal, rather than pirated copy and compare it to the impact of such characteristics as picture quality or delay in delivery. We find a modest but highly significant preference for the authorized version. By conducting otherwise identical choice experiments both with and without actual experiential and monetary consequences, we conclude that the method does not seem to suffer from hypothetical bias. We also find that when the proceeds from legal sale are transferred to a good cause, willingness to pay for the unauthorized copy is reduced.

File sharing as conditional cooperation: evidence from a framed field experiment

ABSTRACT In this project, we investigate downloading and sharing behaviour in a novel ‘piracy game’ modelled after standard public good games. We find that willingness to share correlates positively with the sharing by others. By contrast, actual behaviour in the ‘piracy game’ is not correlated with self-reported behaviour.

Catastrophic Risk: Social Influences on Insurance Decisions [Dataset]

We study behavioral patterns of insurance demand for low-probability large-loss events (catastrophic losses). Individual patterns of belief formation and risk attitude that were suggested in the behavioral decisions literature emerge robustly in the current set of insurance choices. However, social comparison effects are less robust. We do not find any evidence for peer effects (through social loss aversion or imitation) on insurance take-up. In contrast, we find support for the prediction that people underweight others’ relevant information in their own decision making.

We all do it, but are we willing to admit? Incentivizing digital pirates’ confessions

In this study, we try to assess the prevalence of illicit downloading in the market of audio books and the willingness to admit to such practices. We compare the Bayesian Truth Serum (Prelec, 2004) treatment in which truthful responses and precise estimates are rewarded to the control treatment with a flat participation fee. We find a sizable treatment effect – incentivized ‘pirates’ admit approximately 60% more often than the nonincentivized ones.