{"blocks": [{"key": "be0d0615", "text": "Scenario", "type": "header-two", "depth": 0, "inlineStyleRanges": [], "entityRanges": [], "data": {}}, {"key": "50a8b77b", "text": "Technical phone screen for a data-science role focusing on marketing experiment design and causal inference.", "type": "unstyled", "depth": 0, "inlineStyleRanges": [], "entityRanges": [], "data": {}}, {"key": "419d62c0", "text": "Question", "type": "header-two", "depth": 0, "inlineStyleRanges": [], "entityRanges": [], "data": {}}, {"key": "f690846c", "text": "Which Python or R packages do you usually use for causal-inference or experiment analysis, and why? Describe a project where you applied causal-inference methods. What was the business problem, which approach did you choose, and what impact did it deliver? Explain the Difference-in-Differences (DID) technique. What assumptions does it rely on and when would you prefer it over other causal methods? You need to launch an email campaign for the 1point3acres community. How would you select the target users, define success metrics, design the screening/hold-out test, and analyze the results?", "type": "unstyled", "depth": 0, "inlineStyleRanges": [], "entityRanges": [], "data": {}}, {"key": "3f03f240", "text": "Hints", "type": "header-two", "depth": 0, "inlineStyleRanges": [], "entityRanges": [], "data": {}}, {"key": "0c884c3f", "text": "Mention packages like statsmodels, EconML; cover parallel-trends, treatment vs. control, randomization, power, lift and significance.", "type": "unstyled", "depth": 0, "inlineStyleRanges": [], "entityRanges": [], "data": {}}], "entityMap": {}}