import pandas as pd

class Experiment:
    def __init__(self, csv_path):
        self.raw = pd.read_csv(csv_path)
        self.clean = None

    def preprocess(self, cutoff=3):
        df = self.raw.dropna(subset=["score"]).copy()
        z = (df["score"] - df["score"].mean()) / df["score"].std()
        self.clean = df[z.abs() < cutoff]
        return self

    def summary(self):
        if self.clean is None:
            raise RuntimeError("call preprocess() before summary()")
        return self.clean.groupby("condition")["score"].agg(["mean", "std", "count"])
>>> exp = Experiment("data/survey_responses.csv").preprocess(cutoff=3)
>>> exp.summary()
