Statistics - Topics

Why Anscombe's quartet looks identical until you plot it Interests drift fast enough that interest-based categories go stale. What is the difference between mediation and moderation in quantitative psychology? Why the central limit theorem does not rescue a sample of twelve Continental groupings put Russia in Europe and Turkey nowhere How a small charity ends up running the only fatality database for an entire sector How do you detect and handle missing data in longitudinal psychological research? Why does sorting people by a single characteristic produce groups that share almost nothing else? Why do measurable characteristics crowd out the ones that matter most to the people described? What does the variability hypothesis claim and where does the evidence fail? How does Bayesian estimation change the way psychologists report results? The youth rate excludes full-time students which flatters the figure considerably How do you test whether a set of points is spatially random? One hour of work last week is enough to be counted as employed in the household survey Rotating panels: the same households answer unemployment questions for four months then drop out Confidence intervals vs p-values: what do they actually tell you? Base rate neglect: the test is 99 percent accurate and you likely do not have it How is statistical power calculated before a psychology study begins? Discouraged workers who stop looking after a year drop out of the unemployment count entirely Is statistical randomness the same thing as metaphysical contingency? Sector level and industry group level are cited interchangeably and they are not the same Directory entries copy each other so a single error spreads across sites How do you pick an evaluation metric when one class is 99 percent of the data? Inclusion criteria are stated once and then quietly ignored Founding-year lists are survivorship bias with a chart attached Directory editors never merge duplicate entries for the same company Weibull analysis on a fleet of twenty pumps is mostly a confident guess Why model predictions lose skill outside the range they were fitted on Continental lists need a rule for territories and none of them agree Actuaries spend more time setting assumptions than doing math Comparing company counts across countries runs into different legal definitions of a company How do meta analysts correct for publication bias in effect size estimates?