{"blocks": [{"key": "f9853dd2", "text": "Scenario", "type": "header-two", "depth": 0, "inlineStyleRanges": [], "entityRanges": [], "data": {}}, {"key": "1755ffa4", "text": "On-site statistics round", "type": "unstyled", "depth": 0, "inlineStyleRanges": [], "entityRanges": [], "data": {}}, {"key": "94c5c5cf", "text": "Question", "type": "header-two", "depth": 0, "inlineStyleRanges": [], "entityRanges": [], "data": {}}, {"key": "c82295c3", "text": "Population contains one-, two- and three-child families. Estimate the proportion of each family type from a sample of 100 children. Construct a 95 % confidence interval. Explain why the OLS coefficient of Y~X differs from X~Y and relate to causal direction. A regression model shows all coefficients statistically insignificant yet high predictive performance. Provide a statistical explanation and propose a fix.", "type": "unstyled", "depth": 0, "inlineStyleRanges": [], "entityRanges": [], "data": {}}, {"key": "5d3b7130", "text": "Hints", "type": "header-two", "depth": 0, "inlineStyleRanges": [], "entityRanges": [], "data": {}}, {"key": "f33b98d4", "text": "Multinomial proportions, bootstrap CI; reverse causality; multicollinearity and LASSO/ridge.", "type": "unstyled", "depth": 0, "inlineStyleRanges": [], "entityRanges": [], "data": {}}], "entityMap": {}}