{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "cb995b42",
   "metadata": {},
   "source": [
    "# Testy *t*: jedna próba, próby zależne, próby niezależne\n",
    "\n",
    "Na tym wykładzie nauczymy się, jak w praktyce stosować test *t* w trzech klasycznych sytuacjach:\n",
    "\n",
    "1) **jedna próba** (średnia z próby vs wartość odniesienia),\n",
    "2) **próby zależne** (te same osoby w dwóch warunkach),\n",
    "3) **próby niezależne** (dwie niezależne grupy).\n",
    "\n",
    "W przykładach używamy bardzo prostych danych typowych dla kognitywistyki (czasy reakcji). Celem nie jest samo policzenie p-wartości, tylko odpowiedź na pytanie: **jak duży jest efekt i jak niepewny jest jego pomiar** (przedział ufności)."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8c11d8e8",
   "metadata": {},
   "source": [
    "## Cele\n",
    "\n",
    "Po tym notatniku potrafisz:\n",
    "\n",
    "- dobrać wariant testu *t* do pytania badawczego,\n",
    "- rozpoznać wspólny schemat wszystkich testów *t*: **efekt / błąd standardowy / rozkład $t$ / 95% PU / p-wartość**,\n",
    "- przeprowadzić test *t* krok po kroku: historia → pytanie → $H_0$ i $H_A$ → poziom istotności → statystyka i jej rozkład pod $H_0$ → p-wartość → weryfikacja funkcją,\n",
    "- raportować wynik w sposób interpretowalny: liczebność ($n$), estymata efektu, 95% przedział ufności, $t$, stopnie swobody i p,\n",
    "- wyjaśnić, dlaczego test Welcha jest bezpiecznym wyborem, gdy nie chcemy zakładać równości wariancji.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7e27e54f",
   "metadata": {},
   "source": [
    "## Wymagania wstępne\n",
    "\n",
    "- Rozkład normalny i rozkład $t$ (intuicyjnie).\n",
    "- Co to jest błąd standardowy i przedział ufności.\n",
    "- Poprzedni notatnik o przedziałach ufności: szczególnie zależność między 95% PU a testem dwustronnym.\n",
    "- Podstawy `numpy` oraz proste wykresy w `matplotlib`.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "1b153aba",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from scipy import stats\n",
    "\n",
    "from IPython.display import display\n",
    "\n",
    "SEED = 20250116\n",
    "\n",
    "sns.set_theme(style=\"whitegrid\")\n",
    "\n",
    "try:\n",
    "    import ipywidgets as widgets\n",
    "\n",
    "    WIDGETY_DOSTEPNE = True\n",
    "except ImportError:\n",
    "    widgets = None\n",
    "    WIDGETY_DOSTEPNE = False"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "43ba8be0",
   "metadata": {},
   "source": [
    "## Wspólna idea testów *t*\n",
    "\n",
    "Niezależnie od wariantu zawsze pytamy o jakąś **średnią** albo **różnicę średnich** w populacji.\n",
    "\n",
    "- jedna próba: interesuje nas $\\mu - \\mu_0$,\n",
    "- próby zależne: interesuje nas średnia różnica $\\mu_D$,\n",
    "- próby niezależne: interesuje nas różnica $\\mu_1 - \\mu_2$.\n",
    "\n",
    "We wszystkich trzech sytuacjach robimy to samo:\n",
    "\n",
    "1. estymujemy efekt z próby,\n",
    "2. obliczamy jego błąd standardowy,\n",
    "3. standaryzujemy efekt do statystyki\n",
    "\n",
    "$$\n",
    "t = \\frac{\\text{estymata efektu} - \\text{wartość z } H_0}{SE},\n",
    "$$\n",
    "\n",
    "4. odczytujemy zgodność danych z $H_0$ z rozkładu $t$ o odpowiednich stopniach swobody,\n",
    "5. raportujemy efekt w naturalnej skali, 95% PU, $t$, $df$ i p.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "226af0e9",
   "metadata": {},
   "source": [
    "### Most z notatnika o przedziałach ufności\n",
    "\n",
    "To jest dokładnie ta sama logika, co w notatniku o PU, tylko tutaj ustawiamy nacisk na decyzję testową. W teście dwustronnym przy $\\alpha = 0.05$ zachodzi prosta zależność:\n",
    "\n",
    "- odrzucamy $H_0$ wtedy i tylko wtedy, gdy wartość z $H_0$ leży poza 95% PU dla odpowiedniego efektu,\n",
    "- nie odrzucamy $H_0$ wtedy i tylko wtedy, gdy wartość z $H_0$ mieści się w tym PU.\n",
    "\n",
    "W dalszej części będziemy stale wracać do tej równoważności.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "153a337c",
   "metadata": {},
   "source": [
    "## Test *t* dla jednej próby\n",
    "\n",
    "Zaczynamy od najprostszego wariantu, bo najłatwiej na nim zobaczyć wspólną logikę wszystkich testów *t*.\n",
    "\n",
    "Chcemy porównać średnią w próbie z wartością odniesienia $\\mu_0$ (np. normą z poprzednich badań). To jest dokładnie ten sam problem, który w poprzednim notatniku zapisywaliśmy jako: „czy wartość z $H_0$ leży w przedziale ufności dla średniej?”.\n",
    "\n",
    "Jeśli znamy odchylenie standardowe w populacji $\\sigma$, można użyć testu *z*:\n",
    "\n",
    "$$\n",
    "z = \\frac{\\bar{X} - \\mu_0}{\\sigma/\\sqrt{n}}.\n",
    "$$\n",
    "\n",
    "W praktyce $\\sigma$ prawie nigdy nie jest znane, więc zastępujemy je odchyleniem standardowym z próby $s$ i otrzymujemy statystykę:\n",
    "\n",
    "$$\n",
    "t = \\frac{\\bar{X} - \\mu_0}{s/\\sqrt{n}},\\quad T \\sim t(df=n-1) \\ \\text{(gdy } H_0 \\text{ jest prawdziwa)}.\n",
    "$$\n",
    "\n",
    "**Założenia (w wersji klasycznej):** obserwacje są niezależne, a rozkład w populacji jest w przybliżeniu normalny. W praktyce test *t* bywa dość odporny na umiarkowane odstępstwa, zwłaszcza gdy pracujemy na średnich liczonych po wielu próbach (np. średni wynik testu liczony na osobę).\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b12a498e",
   "metadata": {},
   "source": [
    "### Przykład z psychologii: lęk egzaminacyjny a norma (jedna próba)\n",
    "\n",
    "Wyobraźmy sobie początek semestru i krótką ankietę wśród studentów pierwszego roku. Każda osoba wypełnia **krótką skalę lęku egzaminacyjnego**. Wynik mieści się w przedziale od 0 do 20 punktów: wyższy wynik oznacza silniejszy lęk.\n",
    "\n",
    "Z wcześniejszych badań na podobnej populacji mamy punkt odniesienia: średni wynik na tej skali wynosił $\\mu_0 = 11$ punktów.\n",
    "\n",
    "W tym roku zebraliśmy dane od 12 studentów. Każda liczba poniżej to wynik jednej osoby.\n",
    "\n",
    "**Pytanie**\n",
    "\n",
    "Czy średni poziom lęku egzaminacyjnego w tej grupie różni się od normy 11 punktów? Jaki jest oszacowany efekt (w punktach skali) i jego 95% przedział ufności?\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "063e4871",
   "metadata": {},
   "source": [
    "### Przykład (jedna próba) — krok po kroku\n",
    "\n",
    "**Historia:** 12 studentów przed pierwszym ważnym kolokwium; każda wartość to wynik na krótkiej skali lęku egzaminacyjnego; punkt odniesienia: $\\mu_0 = 11$ punktów.\n",
    "\n",
    "**Pytanie:** czy dane są zgodne z populacją o średnim wyniku 11 punktów?\n",
    "\n",
    "**Hipotezy** (dwustronnie):\n",
    "\n",
    "- $H_0: \\mu = 11$\n",
    "- $H_A: \\mu \\neq 11$\n",
    "\n",
    "**Poziom istotności:** $\\alpha = 0.05$.\n",
    "\n",
    "**Statystyka i rozkład pod $H_0$:**\n",
    "\n",
    "$$\n",
    "t = \\frac{\\bar{x} - \\mu_0}{s/\\sqrt{n}},\\quad T \\sim t(df=n-1).\n",
    "$$\n",
    "\n",
    "**p-wartość:**\n",
    "\n",
    "$$\n",
    "p = 2\\,P\\bigl(T \\ge |t_{\\text{obs}}|\\mid H_0\\bigr).\n",
    "$$\n",
    "\n",
    "**Weryfikacja funkcją:** `scipy.stats.ttest_1samp`.\n",
    "\n",
    "**Raportowanie:** poza p-wartością podajemy estymatę efektu $\\bar{x} - \\mu_0$ (punkty skali) oraz 95% przedział ufności.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "a4d8a94e-d69e-446b-84bd-e42f7fc5875b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "n = 12, df = 11\n",
      "Średni wynik w próbie = 13.0 pkt\n",
      "Odchylenie standardowe w próbie = 1.86 pkt\n",
      "Błąd standardowy (SE) = 0.54 pkt\n",
      "\n",
      "Efekt (x̄ - μ0) = 2.0 pkt\n",
      "95% CI dla μ: (11.82, 14.18) pkt\n",
      "95% CI dla efektu (μ - μ0): (0.82, 3.18) pkt\n",
      "d Cohena (1 próba) = 1.08\n",
      "\n",
      "t = 3.728\n",
      "p (dwustronnie) = 0.0033\n",
      "p (SciPy, dwustronnie) = 0.0033\n"
     ]
    }
   ],
   "source": [
    "anxiety_scores = np.array(\n",
    "    [13, 11, 15, 12, 14, 16, 10, 13, 12, 15, 11, 14],\n",
    "    dtype=float,\n",
    ")\n",
    "mu0 = 11.0\n",
    "alpha = 0.05\n",
    "\n",
    "n = anxiety_scores.size\n",
    "sample_mean = float(anxiety_scores.mean())\n",
    "sample_sd = float(anxiety_scores.std(ddof=1))\n",
    "standard_error = sample_sd / np.sqrt(n)\n",
    "df = n - 1\n",
    "\n",
    "t_stat = (sample_mean - mu0) / standard_error\n",
    "p_value_two_sided = 2 * stats.t.sf(abs(t_stat), df=df)\n",
    "\n",
    "t_crit = stats.t.ppf(1 - alpha / 2, df=df)\n",
    "ci_low = sample_mean - t_crit * standard_error\n",
    "ci_high = sample_mean + t_crit * standard_error\n",
    "\n",
    "effect_points = sample_mean - mu0\n",
    "effect_ci_low = ci_low - mu0\n",
    "effect_ci_high = ci_high - mu0\n",
    "cohen_d = effect_points / sample_sd\n",
    "\n",
    "res = stats.ttest_1samp(anxiety_scores, popmean=mu0)\n",
    "\n",
    "print(f\"n = {n}, df = {df}\")\n",
    "print(f\"Średni wynik w próbie = {sample_mean:.1f} pkt\")\n",
    "print(f\"Odchylenie standardowe w próbie = {sample_sd:.2f} pkt\")\n",
    "print(f\"Błąd standardowy (SE) = {standard_error:.2f} pkt\")\n",
    "print()\n",
    "print(f\"Efekt (x̄ - μ0) = {effect_points:.1f} pkt\")\n",
    "print(f\"95% CI dla μ: ({ci_low:.2f}, {ci_high:.2f}) pkt\")\n",
    "print(f\"95% CI dla efektu (μ - μ0): ({effect_ci_low:.2f}, {effect_ci_high:.2f}) pkt\")\n",
    "print(f\"d Cohena (1 próba) = {cohen_d:.2f}\")\n",
    "print()\n",
    "print(f\"t = {t_stat:.3f}\")\n",
    "print(f\"p (dwustronnie) = {float(p_value_two_sided):.4f}\")\n",
    "print(f\"p (SciPy, dwustronnie) = {float(res.pvalue):.4f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "aafbd406",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "rng_plot = np.random.default_rng(SEED + 1)\n",
    "\n",
    "fig, axes = plt.subplots(1, 2, figsize=(12, 4))\n",
    "\n",
    "x_jitter = rng_plot.normal(loc=0.0, scale=0.035, size=n)\n",
    "axes[0].scatter(x_jitter, anxiety_scores, color=\"steelblue\", s=60, zorder=3)\n",
    "axes[0].axhline(mu0, color=\"black\", linestyle=\":\", linewidth=2, label=\"μ0 = 11 pkt\")\n",
    "axes[0].axhline(sample_mean, color=\"crimson\", linewidth=2, label=\"średnia z próby\")\n",
    "axes[0].set_xlim(-0.12, 0.12)\n",
    "axes[0].set_xticks([])\n",
    "axes[0].set_ylabel(\"wynik skali lęku egzaminacyjnego (pkt)\")\n",
    "axes[0].set_title(\"Jedna próba: wyniki osób i norma\")\n",
    "axes[0].grid(axis=\"y\", alpha=0.25)\n",
    "axes[0].legend(loc=\"upper left\")\n",
    "\n",
    "effect_margin = max(abs(effect_ci_low), abs(effect_ci_high), abs(effect_points)) + 1.0\n",
    "point_color = \"crimson\" if not (effect_ci_low <= 0 <= effect_ci_high) else \"steelblue\"\n",
    "\n",
    "axes[1].hlines(0, -effect_margin, effect_margin, color=\"0.85\", linewidth=2, zorder=0)\n",
    "axes[1].axvline(0, color=\"black\", linestyle=\":\", linewidth=2, label=\"H0: brak efektu\")\n",
    "axes[1].plot(\n",
    "    [effect_ci_low, effect_ci_high],\n",
    "    [0, 0],\n",
    "    color=\"steelblue\",\n",
    "    linewidth=8,\n",
    "    solid_capstyle=\"round\",\n",
    "    label=\"95% PU dla efektu\",\n",
    ")\n",
    "axes[1].scatter(\n",
    "    [effect_points],\n",
    "    [0],\n",
    "    color=point_color,\n",
    "    s=90,\n",
    "    zorder=3,\n",
    "    label=\"estymata efektu\",\n",
    ")\n",
    "axes[1].set_xlim(-effect_margin, effect_margin)\n",
    "axes[1].set_ylim(-0.3, 0.3)\n",
    "axes[1].set_yticks([])\n",
    "axes[1].set_xlabel(\"μ - μ0 (pkt)\")\n",
    "axes[1].set_title(\"Efekt i 95% PU\")\n",
    "axes[1].grid(axis=\"x\", alpha=0.25)\n",
    "axes[1].legend(loc=\"upper left\")\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f6a63774",
   "metadata": {},
   "source": [
    "**Interpretacja (ważne)**\n",
    "\n",
    "W tej próbie średni wynik wynosi `13.0` pkt, czyli o `2.0` pkt więcej niż wartość odniesienia `11` pkt. 95% PU dla efektu $(\\mu - \\mu_0)$ wynosi w przybliżeniu `[0.82, 3.18]` pkt, więc zero nie mieści się w tym przedziale.\n",
    "\n",
    "To daje dwa równoważne opisy tego samego wyniku:\n",
    "\n",
    "- testowo: statystyka $t = 3.73$ jest dość daleko od zera, a dwustronna p-wartość wynosi ok. `0.003`,\n",
    "- estymacyjnie: dane są zgodne z przeciętnie wyższym poziomem lęku egzaminacyjnego niż w normie, mniej więcej o 1-3 punkty skali."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "53e8661a",
   "metadata": {},
   "source": [
    "### Widżet: p-wartość jako pole pod krzywą\n",
    "\n",
    "Widżet poniżej pokazuje:\n",
    "\n",
    "- rozkład $t$ pod hipotezą zerową,\n",
    "- obszar odrzucenia przy $\\alpha = 0.05$,\n",
    "- p-wartość dla obserwowanego $t_{\\text{obs}}$.\n",
    "\n",
    "Na wykładzie warto pokazać kolejno trzy sytuacje: `t_obs` blisko 0, `t_obs` na granicy obszaru krytycznego i `t_obs` wyraźnie poza nim.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "ac022b66",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7571eaf5037c490a9a33136471467560",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "VBox(children=(IntSlider(value=11, continuous_update=False, description='df', max=200, min=1), FloatSlider(val…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ad44c6511e9045828575bec5d4406dea",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Output()"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "if not WIDGETY_DOSTEPNE:\n",
    "    print(\n",
    "        \"Widżety nie są dostępne. Jeśli pracujesz w Jupyterze, doinstaluj pakiet ipywidgets, \"\n",
    "        \"aby korzystać z części interaktywnych.\"\n",
    "    )\n",
    "else:\n",
    "    df_slider = widgets.IntSlider(\n",
    "        value=11,\n",
    "        min=1,\n",
    "        max=200,\n",
    "        step=1,\n",
    "        description=\"df\",\n",
    "        continuous_update=False,\n",
    "    )\n",
    "\n",
    "    t_slider = widgets.FloatSlider(\n",
    "        value=5.23,\n",
    "        min=-6.0,\n",
    "        max=6.0,\n",
    "        step=0.05,\n",
    "        description=\"t_obs\",\n",
    "        readout_format=\".2f\",\n",
    "        continuous_update=False,\n",
    "    )\n",
    "\n",
    "    alternative_dropdown = widgets.Dropdown(\n",
    "        options=[\n",
    "            (\"dwustronnie (H_A: μ ≠ μ0)\", \"two-sided\"),\n",
    "            (\"prawostronnie (H_A: μ > μ0)\", \"greater\"),\n",
    "            (\"lewostronnie (H_A: μ < μ0)\", \"less\"),\n",
    "        ],\n",
    "        value=\"two-sided\",\n",
    "        description=\"Test\",\n",
    "    )\n",
    "\n",
    "    def pokaz_p_wartosc(df, t_obs, alternative):\n",
    "        df = int(df)\n",
    "        t_obs = float(t_obs)\n",
    "        alpha = 0.05\n",
    "\n",
    "        x = np.linspace(-6, 6, 1000)\n",
    "        pdf = stats.t.pdf(x, df=df)\n",
    "\n",
    "        if alternative == \"two-sided\":\n",
    "            t_crit = float(stats.t.ppf(1 - alpha / 2, df=df))\n",
    "            p_value = float(2 * stats.t.sf(abs(t_obs), df=df))\n",
    "            crit_mask = (x <= -t_crit) | (x >= t_crit)\n",
    "            p_mask = (x <= -abs(t_obs)) | (x >= abs(t_obs))\n",
    "            opis = \"dwustronnie\"\n",
    "        elif alternative == \"greater\":\n",
    "            t_crit = float(stats.t.ppf(1 - alpha, df=df))\n",
    "            p_value = float(stats.t.sf(t_obs, df=df))\n",
    "            crit_mask = x >= t_crit\n",
    "            p_mask = x >= t_obs\n",
    "            opis = \"prawostronnie\"\n",
    "        else:\n",
    "            t_crit = float(stats.t.ppf(alpha, df=df))\n",
    "            p_value = float(stats.t.cdf(t_obs, df=df))\n",
    "            crit_mask = x <= t_crit\n",
    "            p_mask = x <= t_obs\n",
    "            opis = \"lewostronnie\"\n",
    "\n",
    "        decyzja = \"Odrzucamy H0\" if p_value < alpha else \"Nie odrzucamy H0\"\n",
    "\n",
    "        print(f\"df = {df}, α = {alpha:.2f}, test: {opis}\")\n",
    "        print(f\"t_obs = {t_obs:.2f}\")\n",
    "        print(f\"p-wartość = {p_value:.4f}\")\n",
    "        print(f\"Decyzja: {decyzja}\")\n",
    "\n",
    "        plt.figure(figsize=(10, 4))\n",
    "        plt.plot(x, pdf, color=\"black\", linewidth=2)\n",
    "\n",
    "        plt.fill_between(\n",
    "            x,\n",
    "            0,\n",
    "            pdf,\n",
    "            where=crit_mask,\n",
    "            color=\"crimson\",\n",
    "            alpha=0.20,\n",
    "            label=\"obszar odrzucenia (α = 0.05)\",\n",
    "        )\n",
    "        plt.fill_between(\n",
    "            x,\n",
    "            0,\n",
    "            pdf,\n",
    "            where=p_mask,\n",
    "            color=\"steelblue\",\n",
    "            alpha=0.20,\n",
    "            label=\"p-wartość\",\n",
    "        )\n",
    "\n",
    "        plt.axvline(t_obs, color=\"steelblue\", linestyle=\"--\", linewidth=2, label=\"t_obs\")\n",
    "\n",
    "        if alternative == \"two-sided\":\n",
    "            plt.axvline(t_crit, color=\"crimson\", linestyle=\":\", linewidth=2)\n",
    "            plt.axvline(-t_crit, color=\"crimson\", linestyle=\":\", linewidth=2)\n",
    "        else:\n",
    "            plt.axvline(t_crit, color=\"crimson\", linestyle=\":\", linewidth=2)\n",
    "\n",
    "        plt.xlim(-6, 6)\n",
    "        plt.title(\"Rozkład t, obszar odrzucenia i p-wartość\")\n",
    "        plt.xlabel(\"t\")\n",
    "        plt.ylabel(\"gęstość\")\n",
    "        plt.legend()\n",
    "        plt.tight_layout()\n",
    "        plt.show()\n",
    "\n",
    "    out = widgets.interactive_output(\n",
    "        pokaz_p_wartosc,\n",
    "        {\n",
    "            \"df\": df_slider,\n",
    "            \"t_obs\": t_slider,\n",
    "            \"alternative\": alternative_dropdown,\n",
    "        },\n",
    "    )\n",
    "\n",
    "    display(widgets.VBox([df_slider, t_slider, alternative_dropdown]), out)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2fab125d",
   "metadata": {},
   "source": [
    "## Test *t* dla prób zależnych (pomiary sparowane)\n",
    "\n",
    "Test sparowany jest w istocie testem jednopróbkowym wykonanym na **różnicach w parach**.\n",
    "\n",
    "Często porównujemy dwa warunki mierzone na **tych samych osobach** (np. dwa typy bodźców, dwa dni, „przed” i „po”). Wtedy obserwacje naturalnie łączą się w pary.\n",
    "\n",
    "W teście dla prób zależnych:\n",
    "\n",
    "- liczymy różnice w parach,\n",
    "- testujemy, czy **średnia różnica w populacji** wynosi 0.\n",
    "\n",
    "**Założenia (w wersji klasycznej):** różnice w parach są niezależne między osobami, a ich rozkład jest w przybliżeniu normalny.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7c9e2ab7",
   "metadata": {},
   "source": [
    "### Przykład z psychologii poznawczej: koncentracja w ciszy i przy powiadomieniach telefonu (próby zależne)\n",
    "\n",
    "Wyobraźmy sobie prosty eksperyment bliski codziennemu doświadczeniu studentów. Te same osoby rozwiązują krótki test koncentracji dwa razy:\n",
    "\n",
    "- raz w spokojnej ciszy,\n",
    "- raz wtedy, gdy obok co chwilę pojawiają się **powiadomienia telefonu**.\n",
    "\n",
    "Wynikiem jest liczba poprawnych odpowiedzi w krótkim teście uwagi. Ponieważ mierzymy **te same osoby w dwóch warunkach**, dane są sparowane.\n",
    "\n",
    "Poniżej mamy dane 15 osób. Każda liczba to wynik jednej osoby w danym warunku.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7807ecf5",
   "metadata": {},
   "source": [
    "### Przykład (próby zależne) — krok po kroku\n",
    "\n",
    "**Historia:** 15 studentów; każda osoba rozwiązuje ten sam krótki test koncentracji w dwóch warunkach: w ciszy i przy powiadomieniach telefonu.\n",
    "\n",
    "**Pytanie:** czy przeciętnie liczba poprawnych odpowiedzi jest inna w tych dwóch warunkach?\n",
    "\n",
    "Definiujemy różnicę jako:\n",
    "\n",
    "$$\n",
    "D_i = \\text{wynik}_{\\text{cisza}, i} - \\text{wynik}_{\\text{powiadomienia}, i}.\n",
    "$$\n",
    "\n",
    "Od tego miejsca wszystko wygląda tak samo jak w teście jednopróbkowym: zamiast surowych wyników analizujemy jedną nową zmienną, czyli różnice $D_i$.\n",
    "\n",
    "**Hipotezy** (dwustronnie):\n",
    "\n",
    "- $H_0: \\mu_D = 0$\n",
    "- $H_A: \\mu_D \\neq 0$\n",
    "\n",
    "**Statystyka i rozkład pod $H_0$:**\n",
    "\n",
    "$$\n",
    "t = \\frac{\\bar{D} - 0}{s_D/\\sqrt{n}},\\quad T \\sim t(df=n-1).\n",
    "$$\n",
    "\n",
    "**Raportowanie:** średnia różnica $\\bar{D}$ (punkty testu), 95% przedział ufności dla $\\mu_D$, $t$, $df$ i p.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "eb2e5f17-ac4f-44a7-845a-822eb7c5e750",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "n = 15, df = 14\n",
      "Średnia różnica (cisza - powiadomienia) = 1.40 pkt\n",
      "Odchylenie standardowe różnic = 1.50 pkt\n",
      "Błąd standardowy (SE) = 0.39 pkt\n",
      "\n",
      "95% CI dla średniej różnicy: (0.57, 2.23) pkt\n",
      "d_z (dla różnic) = 0.93\n",
      "\n",
      "t = 3.609\n",
      "p (dwustronnie) = 0.0028\n",
      "p (SciPy) = 0.0028\n"
     ]
    }
   ],
   "source": [
    "wynik_cisza = np.array(\n",
    "    [31, 34, 29, 35, 33, 32, 28, 36, 34, 31, 35, 29, 33, 32, 34],\n",
    "    dtype=float,\n",
    ")\n",
    "\n",
    "wynik_powiadomienia = np.array(\n",
    "    [27, 34, 29, 33, 33, 28, 27, 36, 31, 30, 32, 27, 33, 31, 34],\n",
    "    dtype=float,\n",
    ")\n",
    "\n",
    "d_pkt = wynik_cisza - wynik_powiadomienia\n",
    "\n",
    "n = d_pkt.size\n",
    "df = n - 1\n",
    "alpha = 0.05\n",
    "\n",
    "mean_diff = float(d_pkt.mean())\n",
    "sd_diff = float(d_pkt.std(ddof=1))\n",
    "se = sd_diff / np.sqrt(n)\n",
    "\n",
    "t_stat = mean_diff / se\n",
    "p_value_two_sided = 2 * stats.t.sf(abs(t_stat), df=df)\n",
    "\n",
    "t_crit = stats.t.ppf(1 - alpha / 2, df=df)\n",
    "ci_low = mean_diff - t_crit * se\n",
    "ci_high = mean_diff + t_crit * se\n",
    "\n",
    "d_z = mean_diff / sd_diff\n",
    "\n",
    "res = stats.ttest_rel(wynik_cisza, wynik_powiadomienia)\n",
    "\n",
    "print(f\"n = {n}, df = {df}\")\n",
    "print(f\"Średnia różnica (cisza - powiadomienia) = {mean_diff:.2f} pkt\")\n",
    "print(f\"Odchylenie standardowe różnic = {sd_diff:.2f} pkt\")\n",
    "print(f\"Błąd standardowy (SE) = {se:.2f} pkt\")\n",
    "print()\n",
    "print(f\"95% CI dla średniej różnicy: ({ci_low:.2f}, {ci_high:.2f}) pkt\")\n",
    "print(f\"d_z (dla różnic) = {d_z:.2f}\")\n",
    "print()\n",
    "print(f\"t = {t_stat:.3f}\")\n",
    "print(f\"p (dwustronnie) = {float(p_value_two_sided):.4f}\")\n",
    "print(f\"p (SciPy) = {float(res.pvalue):.4f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "39f75ebc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axes = plt.subplots(1, 2, figsize=(12, 4))\n",
    "\n",
    "for i in range(n):\n",
    "    axes[0].plot(\n",
    "        [0, 1],\n",
    "        [wynik_cisza[i], wynik_powiadomienia[i]],\n",
    "        color=\"lightgrey\",\n",
    "        linewidth=1,\n",
    "    )\n",
    "\n",
    "axes[0].scatter(np.zeros(n), wynik_cisza, color=\"steelblue\", label=\"cisza\", zorder=3)\n",
    "axes[0].scatter(\n",
    "    np.ones(n),\n",
    "    wynik_powiadomienia,\n",
    "    color=\"crimson\",\n",
    "    label=\"powiadomienia\",\n",
    "    zorder=3,\n",
    ")\n",
    "axes[0].set_xticks([0, 1], [\"cisza\", \"powiadomienia\"])\n",
    "axes[0].set_ylabel(\"wynik testu koncentracji (pkt)\")\n",
    "axes[0].set_title(\"Te same osoby w dwóch warunkach\")\n",
    "axes[0].grid(axis=\"y\", alpha=0.25)\n",
    "axes[0].legend()\n",
    "\n",
    "effect_margin = max(abs(ci_low), abs(ci_high), abs(mean_diff)) + 0.8\n",
    "point_color = \"crimson\" if not (ci_low <= 0 <= ci_high) else \"steelblue\"\n",
    "\n",
    "axes[1].hlines(0, -effect_margin, effect_margin, color=\"0.85\", linewidth=2, zorder=0)\n",
    "axes[1].axvline(0, color=\"black\", linestyle=\":\", linewidth=2, label=\"H0: brak efektu\")\n",
    "axes[1].plot(\n",
    "    [ci_low, ci_high],\n",
    "    [0, 0],\n",
    "    color=\"steelblue\",\n",
    "    linewidth=8,\n",
    "    solid_capstyle=\"round\",\n",
    "    label=\"95% PU dla μ_D\",\n",
    ")\n",
    "axes[1].scatter([mean_diff], [0], color=point_color, s=90, zorder=3, label=\"estymata efektu\")\n",
    "axes[1].set_xlim(-effect_margin, effect_margin)\n",
    "axes[1].set_ylim(-0.3, 0.3)\n",
    "axes[1].set_yticks([])\n",
    "axes[1].set_xlabel(\"μ_D = średnia różnica (pkt)\")\n",
    "axes[1].set_title(\"Średnia różnica i 95% PU\")\n",
    "axes[1].grid(axis=\"x\", alpha=0.25)\n",
    "axes[1].legend(loc=\"upper left\")\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3616dcdc",
   "metadata": {},
   "source": [
    "### Wniosek z przykładu\n",
    "\n",
    "W tej próbie średnia różnica `(cisza - powiadomienia)` wynosi `1.40` pkt, a 95% PU dla średniej różnicy to ok. `[0.57, 2.23]` pkt. Ponieważ zero nie mieści się w tym przedziale, dwustronny test `t` daje wynik istotny (`p ≈ 0.0028`)."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "26aa9f99",
   "metadata": {},
   "source": [
    "### Widżet: dlaczego parowanie pomaga\n",
    "\n",
    "W pomiarach sparowanych często jest tak, że osoby, które ogólnie osiągają wysokie wyniki, wypadają lepiej **w obu warunkach**. To powoduje dodatnią **korelację** między pomiarami.\n",
    "\n",
    "Test dla prób zależnych korzysta z tej informacji, ponieważ pracuje na różnicach w parach. Gdy korelacja jest dodatnia, rozrzut różnic jest zwykle mniejszy, a błąd standardowy średniej różnicy maleje.\n",
    "\n",
    "Na wykładzie najlepiej najpierw trzymać `różnicę` i `odchylenie` stałe, a zmieniać tylko korelację.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "47b5a835",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "cea4f4c9ba36425d8012a8c4dd5caef6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "VBox(children=(IntSlider(value=15, continuous_update=False, description='n', max=80, min=5), FloatSlider(value…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "173cb19c73e94a34bb785331a8422202",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Output(outputs=({'name': 'stdout', 'text': 'n = 15\\nZadana korelacja: 0.60, empiryczna korelacja: 0.63\\nŚredni…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "if not WIDGETY_DOSTEPNE:\n",
    "    print(\n",
    "        \"Widżety nie są dostępne. Jeśli pracujesz w Jupyterze, doinstaluj pakiet ipywidgets, \"\n",
    "        \"aby korzystać z części interaktywnych.\"\n",
    "    )\n",
    "else:\n",
    "    n_slider = widgets.IntSlider(\n",
    "        value=15,\n",
    "        min=5,\n",
    "        max=80,\n",
    "        step=1,\n",
    "        description=\"n\",\n",
    "        continuous_update=False,\n",
    "    )\n",
    "\n",
    "    delta_slider = widgets.FloatSlider(\n",
    "        value=1.5,\n",
    "        min=-6.0,\n",
    "        max=6.0,\n",
    "        step=0.1,\n",
    "        description=\"różnica\",\n",
    "        readout_format=\".1f\",\n",
    "        continuous_update=False,\n",
    "    )\n",
    "\n",
    "    sd_slider = widgets.FloatSlider(\n",
    "        value=4.0,\n",
    "        min=1.0,\n",
    "        max=10.0,\n",
    "        step=0.5,\n",
    "        description=\"odch.\",\n",
    "        readout_format=\".1f\",\n",
    "        continuous_update=False,\n",
    "    )\n",
    "\n",
    "    corr_slider = widgets.FloatSlider(\n",
    "        value=0.60,\n",
    "        min=0.00,\n",
    "        max=0.95,\n",
    "        step=0.05,\n",
    "        description=\"korelacja\",\n",
    "        readout_format=\".2f\",\n",
    "        continuous_update=False,\n",
    "    )\n",
    "\n",
    "    def pokaz_parowanie(n, delta_points, sd_points, corr):\n",
    "        n = int(n)\n",
    "        delta_points = float(delta_points)\n",
    "        sd_points = float(sd_points)\n",
    "        corr = float(corr)\n",
    "        alpha = 0.05\n",
    "\n",
    "        seed_seq = np.random.SeedSequence(\n",
    "            [SEED, n, int((delta_points + 20) * 100), int(sd_points * 100), int(corr * 1000)]\n",
    "        )\n",
    "        rng_local = np.random.default_rng(seed_seq)\n",
    "\n",
    "        mu_base = 30.0\n",
    "        cov = np.array(\n",
    "            [\n",
    "                [sd_points**2, corr * sd_points**2],\n",
    "                [corr * sd_points**2, sd_points**2],\n",
    "            ],\n",
    "            dtype=float,\n",
    "        )\n",
    "\n",
    "        data = rng_local.multivariate_normal(\n",
    "            mean=[mu_base, mu_base - delta_points],\n",
    "            cov=cov,\n",
    "            size=n,\n",
    "        )\n",
    "\n",
    "        wynik_cisza = data[:, 0]\n",
    "        wynik_powiadomienia = data[:, 1]\n",
    "        d = wynik_cisza - wynik_powiadomienia\n",
    "\n",
    "        corr_emp = float(np.corrcoef(wynik_cisza, wynik_powiadomienia)[0, 1])\n",
    "\n",
    "        res_paired = stats.ttest_rel(wynik_cisza, wynik_powiadomienia)\n",
    "        res_ind = stats.ttest_ind(wynik_cisza, wynik_powiadomienia, equal_var=False)\n",
    "\n",
    "        se_paired = float(d.std(ddof=1) / np.sqrt(n))\n",
    "        se_ind = float(\n",
    "            np.sqrt(wynik_cisza.var(ddof=1) / n + wynik_powiadomienia.var(ddof=1) / n)\n",
    "        )\n",
    "\n",
    "        tcrit = float(stats.t.ppf(1 - alpha / 2, df=n - 1))\n",
    "        ci_low = float(d.mean() - tcrit * se_paired)\n",
    "        ci_high = float(d.mean() + tcrit * se_paired)\n",
    "\n",
    "        decyzja = \"odrzucamy H0\" if float(res_paired.pvalue) < alpha else \"nie odrzucamy H0\"\n",
    "\n",
    "        print(f\"n = {n}\")\n",
    "        print(f\"Zadana korelacja: {corr:.2f}, empiryczna korelacja: {corr_emp:.2f}\")\n",
    "        print(f\"Średnia różnica (cisza - powiadomienia) = {float(d.mean()):.2f} pkt\")\n",
    "        print(f\"95% CI dla średniej różnicy = ({ci_low:.2f}, {ci_high:.2f}) pkt\")\n",
    "        print(f\"Błąd standardowy (sparowane) = {se_paired:.2f} pkt\")\n",
    "        print(f\"Błąd standardowy (ignorując pary) = {se_ind:.2f} pkt\")\n",
    "        print()\n",
    "        print(\"Test t dla prób zależnych (sparowanych):\")\n",
    "        print(\n",
    "            f\"  t = {float(res_paired.statistic):.3f}, df = {int(res_paired.df)}, p = {float(res_paired.pvalue):.4f}\"\n",
    "        )\n",
    "        print(\"Test t dla prób niezależnych (ignoruje pary):\")\n",
    "        print(\n",
    "            f\"  t = {float(res_ind.statistic):.3f}, df = {float(res_ind.df):.2f}, p = {float(res_ind.pvalue):.4f}\"\n",
    "        )\n",
    "        print(f\"Decyzja przy α={alpha:.2f} (sparowane): {decyzja}.\")\n",
    "\n",
    "        fig, axes = plt.subplots(1, 2, figsize=(12, 4))\n",
    "\n",
    "        axes[0].scatter(wynik_cisza, wynik_powiadomienia, color=\"steelblue\", alpha=0.7)\n",
    "        min_v = float(min(wynik_cisza.min(), wynik_powiadomienia.min()))\n",
    "        max_v = float(max(wynik_cisza.max(), wynik_powiadomienia.max()))\n",
    "        axes[0].plot(\n",
    "            [min_v, max_v],\n",
    "            [min_v, max_v],\n",
    "            color=\"grey\",\n",
    "            linestyle=\"--\",\n",
    "            linewidth=1,\n",
    "        )\n",
    "        axes[0].set_title(\"Te same osoby w dwóch warunkach\")\n",
    "        axes[0].set_xlabel(\"cisza (pkt)\")\n",
    "        axes[0].set_ylabel(\"powiadomienia (pkt)\")\n",
    "\n",
    "        axes[1].hist(d, bins=12, color=\"lightgrey\", edgecolor=\"white\")\n",
    "        axes[1].axvline(0, color=\"black\", linestyle=\":\", linewidth=2, label=\"H0: brak efektu\")\n",
    "        axes[1].axvline(float(d.mean()), color=\"crimson\", linewidth=2, label=\"średnia różnica\")\n",
    "        axes[1].set_title(\"Rozkład różnic w parach\")\n",
    "        axes[1].set_xlabel(\"cisza - powiadomienia (pkt)\")\n",
    "        axes[1].set_ylabel(\"liczność\")\n",
    "        axes[1].legend()\n",
    "\n",
    "        plt.tight_layout()\n",
    "        plt.show()\n",
    "\n",
    "    out = widgets.interactive_output(\n",
    "        pokaz_parowanie,\n",
    "        {\n",
    "            \"n\": n_slider,\n",
    "            \"delta_points\": delta_slider,\n",
    "            \"sd_points\": sd_slider,\n",
    "            \"corr\": corr_slider,\n",
    "        },\n",
    "    )\n",
    "\n",
    "    display(widgets.VBox([n_slider, delta_slider, sd_slider, corr_slider]), out)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fefe1621",
   "metadata": {},
   "source": [
    "### Skąd bierze się błąd standardowy różnicy średnich?\n",
    "\n",
    "To znowu ta sama logika, co wcześniej: jeśli umiemy opisać rozkład jednej średniej z próby, to umiemy też opisać rozkład **różnicy dwóch niezależnych średnich**.\n",
    "\n",
    "W dużym uproszczeniu:\n",
    "\n",
    "- średnia z próby ma wariancję $\\sigma^2/n$,\n",
    "- a **różnica dwóch niezależnych średnich** ma (w przybliżeniu) wariancję będącą sumą wariancji.\n",
    "\n",
    "Poniższa symulacja pokazuje tę własność na sztucznej populacji wyników testu.\n",
    "\n",
    "Skoro umiemy opisać błąd standardowy różnicy średnich, możemy zbudować test dla dwóch niezależnych prób.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "464f1113",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sprawdzenie na symulacji:\n",
      "var(średnia1) ≈ 0.199\n",
      "teoria: sigma1^2/n = 0.200\n",
      "\n",
      "var(średnia2) ≈ 0.450\n",
      "teoria: sigma2^2/n = 0.450\n",
      "\n",
      "var(różnica)  ≈ 0.650\n",
      "teoria: sigma1^2/n + sigma2^2/n = 0.650\n"
     ]
    }
   ],
   "source": [
    "rng_sim = np.random.default_rng(SEED + 10)\n",
    "\n",
    "n = 20\n",
    "mu_points = 10.0\n",
    "sigma1 = 2.0\n",
    "sigma2 = 3.0\n",
    "repeats = 80_000\n",
    "\n",
    "samples1 = rng_sim.normal(loc=mu_points, scale=sigma1, size=(repeats, n))\n",
    "samples2 = rng_sim.normal(loc=mu_points, scale=sigma2, size=(repeats, n))\n",
    "\n",
    "means1 = samples1.mean(axis=1)\n",
    "means2 = samples2.mean(axis=1)\n",
    "\n",
    "diff = means1 - means2\n",
    "\n",
    "print(\"Sprawdzenie na symulacji:\")\n",
    "print(f\"var(średnia1) ≈ {float(means1.var(ddof=0)):.3f}\")\n",
    "print(f\"teoria: sigma1^2/n = {float(sigma1**2 / n):.3f}\")\n",
    "print()\n",
    "print(f\"var(średnia2) ≈ {float(means2.var(ddof=0)):.3f}\")\n",
    "print(f\"teoria: sigma2^2/n = {float(sigma2**2 / n):.3f}\")\n",
    "print()\n",
    "print(f\"var(różnica)  ≈ {float(diff.var(ddof=0)):.3f}\")\n",
    "print(\n",
    "    f\"teoria: sigma1^2/n + sigma2^2/n = {float(sigma1**2 / n + sigma2**2 / n):.3f}\"\n",
    ")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "d3553419",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sd_diff_theory = float(np.sqrt(sigma1**2 / n + sigma2**2 / n))\n",
    "x_limit = float(np.ceil(np.max(np.abs(diff))))\n",
    "\n",
    "plt.figure(figsize=(10, 4))\n",
    "plt.hist(diff, bins=60, density=True, color=\"lightgrey\", edgecolor=\"white\")\n",
    "\n",
    "x = np.linspace(-x_limit, x_limit, 400)\n",
    "plt.plot(\n",
    "    x,\n",
    "    stats.norm.pdf(x, loc=0.0, scale=sd_diff_theory),\n",
    "    color=\"crimson\",\n",
    "    linewidth=2,\n",
    "    label=\"normalny (teoria)\",\n",
    ")\n",
    "plt.axvline(0, color=\"black\", linestyle=\":\", linewidth=2)\n",
    "\n",
    "plt.xlim(-x_limit, x_limit)\n",
    "plt.title(\"Rozkład różnicy średnich z prób: symulacja vs teoria\")\n",
    "plt.xlabel(r\"$\\bar{X}_1 - \\bar{X}_2$ (pkt)\")\n",
    "plt.ylabel(\"gęstość\")\n",
    "plt.grid(axis=\"x\", alpha=0.25)\n",
    "plt.legend()\n",
    "plt.tight_layout()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0856e0f5",
   "metadata": {},
   "source": [
    "## Test *t* dla prób niezależnych\n",
    "\n",
    "Gdy porównujemy **dwie niezależne grupy**, interesuje nas nie pojedyncza średnia, tylko **różnica średnich**.\n",
    "\n",
    "Najczęstsze pytanie ma postać:\n",
    "\n",
    "- czy $\\mu_1 - \\mu_2 = 0$?\n",
    "\n",
    "W praktyce mamy dwie popularne wersje testu:\n",
    "\n",
    "- **test Studenta** (zakłada równość wariancji w populacjach),\n",
    "- **test Welcha** (nie zakłada równości wariancji).\n",
    "\n",
    "Obie wersje opierają się na tej samej idei: bierzemy **oszacowaną różnicę średnich** i dzielimy ją przez **oszacowany błąd standardowy tej różnicy**. Różnią się tym, jak liczymy ten błąd standardowy i ile stopni swobody przypisujemy statystyce testowej.\n",
    "\n",
    "**Założenia (w wersji klasycznej):**\n",
    "\n",
    "- obserwacje w każdej grupie są niezależne,\n",
    "- grupy są niezależne od siebie,\n",
    "- rozkłady w populacjach są w przybliżeniu normalne albo próby są na tyle duże, że rozkład średnich jest bliski normalnemu,\n",
    "- tylko test Studenta dodatkowo zakłada równość **wariancji w populacjach**.\n",
    "\n",
    "W praktyce test Welcha jest dobrym domyślnym wyborem, jeśli nie mamy mocnych podstaw, by zakładać równość wariancji.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "81fc23c0",
   "metadata": {},
   "source": [
    "### Przykład z psychologii poznawczej: trening pamięci a wynik testu (dwie grupy)\n",
    "\n",
    "Wyobraźmy sobie krótkie badanie na zajęciach. Jedna grupa dostaje prostą strategię pamięciową przed testem (np. grupowanie słów w sensowne kategorie), a druga dostaje tylko zwykłą instrukcję.\n",
    "\n",
    "Następnie wszyscy próbują odtworzyć listę 15 słów. Wynikiem jest liczba poprawnie przypomnianych słów.\n",
    "\n",
    "Każda liczba poniżej to wynik jednej osoby. Grupy są niezależne: student z grupy treningu nie ma „pary” w grupie kontrolnej.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3d974aac",
   "metadata": {},
   "source": [
    "### Przykład (próby niezależne) — krok po kroku\n",
    "\n",
    "**Historia:** dwie niezależne grupy: grupa kontrolna bez treningu pamięci i grupa po krótkim treningu strategii zapamiętywania.\n",
    "\n",
    "**Pytanie:** czy średnie wyniki w populacjach różnią się?\n",
    "\n",
    "**Hipotezy** (dwustronnie):\n",
    "\n",
    "- $H_0: \\mu_{\\text{trening}} - \\mu_{\\text{kontrola}} = 0$\n",
    "- $H_A: \\mu_{\\text{trening}} - \\mu_{\\text{kontrola}} \\neq 0$\n",
    "\n",
    "W wariancie Welcha statystyka ma postać:\n",
    "\n",
    "$$\n",
    "t_W = \\frac{\\bar{x}_1 - \\bar{x}_2}{\\sqrt{s_1^2/n_1 + s_2^2/n_2}},\n",
    "$$\n",
    "\n",
    "a stopnie swobody są przybliżane formułą Welcha-Satterthwaite'a:\n",
    "\n",
    "$$\n",
    "df_W \\approx \\frac{\\left(s_1^2/n_1 + s_2^2/n_2\\right)^2}{\\dfrac{\\left(s_1^2/n_1\\right)^2}{n_1-1} + \\dfrac{\\left(s_2^2/n_2\\right)^2}{n_2-1}}.\n",
    "$$\n",
    "\n",
    "W klasycznym wariancie Studenta zakładamy równość wariancji i najpierw liczymy wspólne oszacowanie wariancji:\n",
    "\n",
    "$$\n",
    "s_p^2 = \\frac{(n_1 - 1)s_1^2 + (n_2 - 1)s_2^2}{n_1 + n_2 - 2}.\n",
    "$$\n",
    "\n",
    "Wtedy statystyka ma postać:\n",
    "\n",
    "$$\n",
    "t_S = \\frac{\\bar{x}_1 - \\bar{x}_2}{s_p\\sqrt{1/n_1 + 1/n_2}},\n",
    "$$\n",
    "\n",
    "a liczba stopni swobody wynosi po prostu:\n",
    "\n",
    "$$\n",
    "df_S = n_1 + n_2 - 2.\n",
    "$$\n",
    "\n",
    "W teście Welcha liczba stopni swobody jest zwykle **niższa** niż w klasycznym teście Studenta. To oznacza \"grubsze\" ogony rozkładu $t$, trochę większe wartości krytyczne i w konsekwencji test bardziej **konserwatywny**. Gdy wariancje są bardzo podobne, oba warianty mogą jednak dawać niemal ten sam wynik.\n",
    "\n",
    "**Raportowanie:** różnica średnich (liczba słów), 95% przedział ufności, $t$, $df$ i p.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "91dc262e-03b1-4eb9-81ad-9d30aa946974",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Podstawowe statystyki:\n",
      "  kontrola: n = 15, średnia = 9.60 słowa, sd = 1.35\n",
      "  trening:  n = 15, średnia = 11.67 słowa, sd = 2.06\n",
      "\n",
      "Różnica średnich (trening - kontrola) = 2.07 słowa\n",
      "d Cohena (2 grupy) = 1.19\n",
      "\n",
      "Welch:\n",
      "  t = 3.250, df = 24.18, p = 0.0034\n",
      "  95% CI dla różnicy: (0.75, 3.38) słowa\n",
      "\n",
      "Student (równość wariancji):\n",
      "  t = 3.250, df = 28, p = 0.0030\n",
      "  95% CI dla różnicy: (0.76, 3.37) słowa\n",
      "\n",
      "Weryfikacja SciPy:\n",
      "  Welch:   t = 3.250, df = 24.18, p = 0.0034\n",
      "  Student: t = 3.250, df = 28.00, p = 0.0030\n"
     ]
    }
   ],
   "source": [
    "wynik_kontrola = np.array(\n",
    "    [9, 10, 11, 8, 12, 9, 10, 11, 7, 10, 9, 8, 11, 10, 9],\n",
    "    dtype=float,\n",
    ")\n",
    "\n",
    "wynik_trening = np.array(\n",
    "    [10, 14, 12, 9, 15, 11, 12, 14, 8, 13, 11, 10, 14, 12, 10],\n",
    "    dtype=float,\n",
    ")\n",
    "\n",
    "alpha = 0.05\n",
    "\n",
    "n1 = wynik_trening.size\n",
    "n2 = wynik_kontrola.size\n",
    "\n",
    "mean1 = float(wynik_trening.mean())\n",
    "mean2 = float(wynik_kontrola.mean())\n",
    "diff_means = mean1 - mean2\n",
    "\n",
    "var1 = float(wynik_trening.var(ddof=1))\n",
    "var2 = float(wynik_kontrola.var(ddof=1))\n",
    "\n",
    "se_welch = float(np.sqrt(var1 / n1 + var2 / n2))\n",
    "df_welch = float(\n",
    "    (var1 / n1 + var2 / n2) ** 2\n",
    "    / ((var1**2) / (n1**2 * (n1 - 1)) + (var2**2) / (n2**2 * (n2 - 1)))\n",
    ")\n",
    "t_welch = diff_means / se_welch\n",
    "p_welch = 2 * stats.t.sf(abs(t_welch), df=df_welch)\n",
    "tcrit_welch = float(stats.t.ppf(1 - alpha / 2, df=df_welch))\n",
    "ci_welch_low = diff_means - tcrit_welch * se_welch\n",
    "ci_welch_high = diff_means + tcrit_welch * se_welch\n",
    "\n",
    "sp2 = float(((n1 - 1) * var1 + (n2 - 1) * var2) / (n1 + n2 - 2))\n",
    "se_student = float(np.sqrt(sp2 * (1 / n1 + 1 / n2)))\n",
    "df_student = n1 + n2 - 2\n",
    "t_student = diff_means / se_student\n",
    "p_student = 2 * stats.t.sf(abs(t_student), df=df_student)\n",
    "tcrit_student = float(stats.t.ppf(1 - alpha / 2, df=df_student))\n",
    "ci_student_low = diff_means - tcrit_student * se_student\n",
    "ci_student_high = diff_means + tcrit_student * se_student\n",
    "\n",
    "cohen_d = diff_means / np.sqrt(sp2)\n",
    "\n",
    "res_welch = stats.ttest_ind(wynik_trening, wynik_kontrola, equal_var=False)\n",
    "res_student = stats.ttest_ind(wynik_trening, wynik_kontrola, equal_var=True)\n",
    "\n",
    "print(\"Podstawowe statystyki:\")\n",
    "print(f\"  kontrola: n = {n2}, średnia = {mean2:.2f} słowa, sd = {np.sqrt(var2):.2f}\")\n",
    "print(f\"  trening:  n = {n1}, średnia = {mean1:.2f} słowa, sd = {np.sqrt(var1):.2f}\")\n",
    "print()\n",
    "print(f\"Różnica średnich (trening - kontrola) = {diff_means:.2f} słowa\")\n",
    "print(f\"d Cohena (2 grupy) = {cohen_d:.2f}\")\n",
    "print()\n",
    "print(\"Welch:\")\n",
    "print(f\"  t = {t_welch:.3f}, df = {df_welch:.2f}, p = {float(p_welch):.4f}\")\n",
    "print(f\"  95% CI dla różnicy: ({ci_welch_low:.2f}, {ci_welch_high:.2f}) słowa\")\n",
    "print()\n",
    "print(\"Student (równość wariancji):\")\n",
    "print(f\"  t = {t_student:.3f}, df = {df_student}, p = {float(p_student):.4f}\")\n",
    "print(f\"  95% CI dla różnicy: ({ci_student_low:.2f}, {ci_student_high:.2f}) słowa\")\n",
    "print()\n",
    "print(\"Weryfikacja SciPy:\")\n",
    "print(\n",
    "    f\"  Welch:   t = {float(res_welch.statistic):.3f}, df = {float(res_welch.df):.2f}, p = {float(res_welch.pvalue):.4f}\"\n",
    ")\n",
    "print(\n",
    "    f\"  Student: t = {float(res_student.statistic):.3f}, df = {float(res_student.df):.2f}, p = {float(res_student.pvalue):.4f}\"\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "8d12008e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "rng_plot = np.random.default_rng(SEED + 20)\n",
    "\n",
    "fig, axes = plt.subplots(1, 2, figsize=(12, 4))\n",
    "\n",
    "x_kontrola = rng_plot.normal(loc=0, scale=0.04, size=n2)\n",
    "x_trening = rng_plot.normal(loc=1, scale=0.04, size=n1)\n",
    "\n",
    "axes[0].scatter(x_kontrola, wynik_kontrola, color=\"grey\", alpha=0.8, label=\"kontrola\")\n",
    "axes[0].scatter(x_trening, wynik_trening, color=\"steelblue\", alpha=0.8, label=\"trening\")\n",
    "axes[0].set_xticks([0, 1], [\"kontrola\", \"trening\"])\n",
    "axes[0].set_ylabel(\"liczba poprawnie przypomnianych słów\")\n",
    "axes[0].set_title(\"Trening pamięci: dwie niezależne grupy\")\n",
    "axes[0].grid(axis=\"y\", alpha=0.25)\n",
    "axes[0].legend()\n",
    "\n",
    "effect_margin = max(\n",
    "    abs(ci_welch_low),\n",
    "    abs(ci_welch_high),\n",
    "    abs(ci_student_low),\n",
    "    abs(ci_student_high),\n",
    "    abs(diff_means),\n",
    ") + 0.8\n",
    "\n",
    "axes[1].axvline(0, color=\"black\", linestyle=\":\", linewidth=2, label=\"H0: brak różnicy\")\n",
    "\n",
    "for y, ci, color in [\n",
    "    (1, (ci_welch_low, ci_welch_high), \"steelblue\"),\n",
    "    (0, (ci_student_low, ci_student_high), \"darkorange\"),\n",
    "]:\n",
    "    axes[1].plot([ci[0], ci[1]], [y, y], color=color, linewidth=6, solid_capstyle=\"round\")\n",
    "    axes[1].scatter([diff_means], [y], color=color, s=80, zorder=3)\n",
    "\n",
    "axes[1].set_xlim(-effect_margin, effect_margin)\n",
    "axes[1].set_ylim(-0.7, 1.7)\n",
    "axes[1].set_yticks([1, 0], [\"Welch\", \"Student\"])\n",
    "axes[1].set_xlabel(\"μ_trening - μ_kontrola (słowa)\")\n",
    "axes[1].set_title(\"Różnica średnich i 95% PU\")\n",
    "axes[1].grid(axis=\"x\", alpha=0.25)\n",
    "axes[1].legend(loc=\"upper left\")\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e7b6ccab",
   "metadata": {},
   "source": [
    "### Wniosek z przykładu\n",
    "\n",
    "W obu wersjach testu oszacowana różnica średnich `(trening - kontrola)` jest dodatnia i wynosi ok. `2.07` słowa. Zarówno PU Welcha `[0.75, 3.38]`, jak i PU Studenta `[0.76, 3.37]` leżą powyżej zera.\n",
    "\n",
    "Na tych danych widać też różnicę w liczbie stopni swobody: dla Welcha mamy `df \\approx 24.18`, a dla klasycznego Studenta `df = 28`. Welch używa więc nieco mniejszej liczby stopni swobody, czyli bardziej konserwatywnego rozkładu odniesienia. Tutaj praktyczny wniosek pozostaje ten sam, ale przy trudniejszych danych ta różnica może już mieć znaczenie.\n",
    "\n",
    "To jest zgodne z lepszym wynikiem w grupie po krótkim treningu pamięci. Na tych danych Welch i Student dają bardzo podobny wniosek, ale dydaktycznie i praktycznie bezpieczniej traktować Welcha jako wybór domyślny.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "13263c7a",
   "metadata": {},
   "source": [
    "### Widżet: Student vs Welch\n",
    "\n",
    "Widżet poniżej pozwala zobaczyć, jak zmieniają się wyniki testu, gdy:\n",
    "\n",
    "- zmieniamy liczebności $n_1$ i $n_2$,\n",
    "- zmieniamy wariancje w grupach,\n",
    "- ustawiamy różnicę średnich w populacji.\n",
    "\n",
    "Dane w widżecie są symulowane, więc traktuj go jako ilustrację intuicji. Na wykładzie warto najpierw pokazać przypadek podobnych rozrzutów wyników, a potem wyraźnie nierównych.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "64918e95",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3c52538b658945998f4f348666a10d40",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "VBox(children=(IntSlider(value=15, continuous_update=False, description='n K', max=80, min=5), IntSlider(value…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a8e93287d462474aa62f10634ef03cf1",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Output()"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "if not WIDGETY_DOSTEPNE:\n",
    "    print(\n",
    "        \"Widżety nie są dostępne. Jeśli pracujesz w Jupyterze, doinstaluj pakiet ipywidgets, \"\n",
    "        \"aby korzystać z części interaktywnych.\"\n",
    "    )\n",
    "else:\n",
    "    n1_slider = widgets.IntSlider(\n",
    "        value=15,\n",
    "        min=5,\n",
    "        max=80,\n",
    "        step=1,\n",
    "        description=\"n K\",\n",
    "        continuous_update=False,\n",
    "    )\n",
    "\n",
    "    n2_slider = widgets.IntSlider(\n",
    "        value=15,\n",
    "        min=5,\n",
    "        max=80,\n",
    "        step=1,\n",
    "        description=\"n T\",\n",
    "        continuous_update=False,\n",
    "    )\n",
    "\n",
    "    sd1_slider = widgets.FloatSlider(\n",
    "        value=2.0,\n",
    "        min=0.5,\n",
    "        max=6.0,\n",
    "        step=0.1,\n",
    "        description=\"odch. K\",\n",
    "        readout_format=\".1f\",\n",
    "        continuous_update=False,\n",
    "    )\n",
    "\n",
    "    sd2_slider = widgets.FloatSlider(\n",
    "        value=3.0,\n",
    "        min=0.5,\n",
    "        max=6.0,\n",
    "        step=0.1,\n",
    "        description=\"odch. T\",\n",
    "        readout_format=\".1f\",\n",
    "        continuous_update=False,\n",
    "    )\n",
    "\n",
    "    diff_slider = widgets.FloatSlider(\n",
    "        value=2.0,\n",
    "        min=-5.0,\n",
    "        max=5.0,\n",
    "        step=0.1,\n",
    "        description=\"różnica\",\n",
    "        readout_format=\".1f\",\n",
    "        continuous_update=False,\n",
    "    )\n",
    "\n",
    "    max_sd = max(sd1_slider.max, sd2_slider.max)\n",
    "    min_n = min(n1_slider.min, n2_slider.min)\n",
    "    max_tcrit = float(stats.t.ppf(0.975, df=min_n - 1))\n",
    "    max_half_width = max_tcrit * float(np.sqrt(max_sd**2 / min_n + max_sd**2 / min_n))\n",
    "    effect_margin = float(\n",
    "        np.ceil(max(abs(diff_slider.min), abs(diff_slider.max)) + max_half_width)\n",
    "    )\n",
    "\n",
    "    def pokaz_student_welch(n1, n2, sd1, sd2, diff_points):\n",
    "        n1 = int(n1)\n",
    "        n2 = int(n2)\n",
    "        sd1 = float(sd1)\n",
    "        sd2 = float(sd2)\n",
    "        diff_points = float(diff_points)\n",
    "        alpha = 0.05\n",
    "\n",
    "        seed_seq = np.random.SeedSequence(\n",
    "            [SEED, n1, n2, int(sd1 * 100), int(sd2 * 100), int((diff_points + 20) * 100)]\n",
    "        )\n",
    "        rng_local = np.random.default_rng(seed_seq)\n",
    "\n",
    "        mu = 10.0\n",
    "\n",
    "        grupa_kontrola = rng_local.normal(loc=mu, scale=sd1, size=n1)\n",
    "        grupa_trening = rng_local.normal(loc=mu + diff_points, scale=sd2, size=n2)\n",
    "\n",
    "        mean_kontrola = float(grupa_kontrola.mean())\n",
    "        mean_trening = float(grupa_trening.mean())\n",
    "        diff = mean_trening - mean_kontrola\n",
    "\n",
    "        res_student = stats.ttest_ind(grupa_trening, grupa_kontrola, equal_var=True)\n",
    "        res_welch = stats.ttest_ind(grupa_trening, grupa_kontrola, equal_var=False)\n",
    "\n",
    "        var_kontrola = float(grupa_kontrola.var(ddof=1))\n",
    "        var_trening = float(grupa_trening.var(ddof=1))\n",
    "\n",
    "        se_welch = float(np.sqrt(var_trening / n2 + var_kontrola / n1))\n",
    "        df_welch = float(\n",
    "            (var_trening / n2 + var_kontrola / n1) ** 2\n",
    "            / (\n",
    "                (var_trening**2) / (n2**2 * (n2 - 1))\n",
    "                + (var_kontrola**2) / (n1**2 * (n1 - 1))\n",
    "            )\n",
    "        )\n",
    "        tcrit_welch = float(stats.t.ppf(1 - alpha / 2, df=df_welch))\n",
    "        ci_welch = (diff - tcrit_welch * se_welch, diff + tcrit_welch * se_welch)\n",
    "\n",
    "        sp2 = float(((n1 - 1) * var_kontrola + (n2 - 1) * var_trening) / (n1 + n2 - 2))\n",
    "        se_student = float(np.sqrt(sp2 * (1 / n1 + 1 / n2)))\n",
    "        df_student = n1 + n2 - 2\n",
    "        tcrit_student = float(stats.t.ppf(1 - alpha / 2, df=df_student))\n",
    "        ci_student = (\n",
    "            diff - tcrit_student * se_student,\n",
    "            diff + tcrit_student * se_student,\n",
    "        )\n",
    "\n",
    "        print(f\"n kontrola = {n1}, n trening = {n2}\")\n",
    "        print(\n",
    "            f\"Średnie w próbie: kontrola = {mean_kontrola:.2f} pkt, \"\n",
    "            f\"trening = {mean_trening:.2f} pkt\"\n",
    "        )\n",
    "        print(\n",
    "            f\"Odchylenia w próbie: kontrola = {float(np.sqrt(var_kontrola)):.2f} pkt, \"\n",
    "            f\"trening = {float(np.sqrt(var_trening)):.2f} pkt\"\n",
    "        )\n",
    "        print(f\"Różnica średnich (trening - kontrola) = {diff:.2f} pkt\")\n",
    "        print()\n",
    "        print(\"Welch:\")\n",
    "        print(\n",
    "            f\"  t = {float(res_welch.statistic):.3f}, df = {float(res_welch.df):.2f}, p = {float(res_welch.pvalue):.4f}\"\n",
    "        )\n",
    "        print(f\"  95% CI: ({ci_welch[0]:.2f}, {ci_welch[1]:.2f}) pkt\")\n",
    "        print()\n",
    "        print(\"Student:\")\n",
    "        print(\n",
    "            f\"  t = {float(res_student.statistic):.3f}, df = {float(res_student.df):.2f}, p = {float(res_student.pvalue):.4f}\"\n",
    "        )\n",
    "        print(f\"  95% CI: ({ci_student[0]:.2f}, {ci_student[1]:.2f}) pkt\")\n",
    "\n",
    "        fig, axes = plt.subplots(1, 2, figsize=(12, 4.5))\n",
    "\n",
    "        x_kontrola = rng_local.normal(loc=0.0, scale=0.04, size=n1)\n",
    "        x_trening = rng_local.normal(loc=1.0, scale=0.04, size=n2)\n",
    "\n",
    "        axes[0].scatter(x_kontrola, grupa_kontrola, color=\"grey\", alpha=0.8, label=\"kontrola\")\n",
    "        axes[0].scatter(x_trening, grupa_trening, color=\"steelblue\", alpha=0.8, label=\"trening\")\n",
    "        axes[0].set_xticks([0, 1], [\"kontrola\", \"trening\"])\n",
    "        axes[0].set_ylabel(\"wynik testu (pkt)\")\n",
    "        axes[0].set_title(\"Dane symulowane: dwie niezależne grupy\")\n",
    "        axes[0].grid(axis=\"y\", alpha=0.25)\n",
    "        axes[0].legend()\n",
    "\n",
    "        axes[1].axvline(0, color=\"black\", linestyle=\":\", linewidth=2, label=\"H0: brak różnicy\")\n",
    "\n",
    "        for y, ci, color in [\n",
    "            (1, ci_welch, \"steelblue\"),\n",
    "            (0, ci_student, \"darkorange\"),\n",
    "        ]:\n",
    "            axes[1].plot([ci[0], ci[1]], [y, y], color=color, linewidth=6, solid_capstyle=\"round\")\n",
    "            axes[1].scatter([diff], [y], color=color, s=80, zorder=3)\n",
    "\n",
    "        axes[1].set_xlim(-effect_margin, effect_margin)\n",
    "        axes[1].set_ylim(-0.7, 1.7)\n",
    "        axes[1].set_yticks([1, 0], [\"Welch\", \"Student\"])\n",
    "        axes[1].set_xlabel(\"μ_trening - μ_kontrola (pkt)\")\n",
    "        axes[1].set_title(\"Różnica średnich i 95% PU\")\n",
    "        axes[1].grid(axis=\"x\", alpha=0.25)\n",
    "        axes[1].legend(loc=\"upper left\")\n",
    "\n",
    "        plt.tight_layout()\n",
    "        plt.show()\n",
    "\n",
    "    out = widgets.interactive_output(\n",
    "        pokaz_student_welch,\n",
    "        {\n",
    "            \"n1\": n1_slider,\n",
    "            \"n2\": n2_slider,\n",
    "            \"sd1\": sd1_slider,\n",
    "            \"sd2\": sd2_slider,\n",
    "            \"diff_points\": diff_slider,\n",
    "        },\n",
    "    )\n",
    "\n",
    "    display(\n",
    "        widgets.VBox(\n",
    "            [\n",
    "                n1_slider,\n",
    "                n2_slider,\n",
    "                sd1_slider,\n",
    "                sd2_slider,\n",
    "                diff_slider,\n",
    "            ]\n",
    "        ),\n",
    "        out,\n",
    "    )\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c0872ad1",
   "metadata": {},
   "source": [
    "### Widżet: empiryczna alfa testu Studenta i Welcha\n",
    "\n",
    "Ten widżet bada **kalibrację testu**, więc wszystkie symulacje prowadzimy przy prawdziwej hipotezie zerowej:\n",
    "\n",
    "$$\n",
    "\\mu_{\\text{kontrola}} = \\mu_{\\text{trening}}.\n",
    "$$\n",
    "\n",
    "Manipulujemy więc tylko liczebnościami i odchyleniami standardowymi w grupach. Po kliknięciu `Symuluj` generujemy wiele badań i sprawdzamy, jak często test Studenta oraz test Welcha odrzucają $H_0$ przy zadanym nominalnym poziomie $\\alpha$.\n",
    "\n",
    "Jeśli empiryczna alfa jest wyraźnie wyższa od nominalnej, test jest **zbyt liberalny**. Jeśli jest wyraźnie niższa, test jest **zbyt konserwatywny**. To jest widżet o **alfie**, a nie o mocy.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "88955062",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b856a4150d1548ff8587a24cf50f7202",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "VBox(children=(HTML(value='<b>Symulacja przy prawdziwej H0:</b> &mu;<sub>kontrola</sub> = &mu;<sub>trening</su…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4883359d4300484eb0c2e497f4196978",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Output()"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "if not WIDGETY_DOSTEPNE:\n",
    "    print(\n",
    "        \"Widżety nie są dostępne. Jeśli pracujesz w Jupyterze, doinstaluj pakiet ipywidgets, \"\n",
    "        \"aby korzystać z części interaktywnych.\"\n",
    "    )\n",
    "else:\n",
    "    n1_slider_alpha = widgets.IntSlider(\n",
    "        value=15,\n",
    "        min=5,\n",
    "        max=80,\n",
    "        step=1,\n",
    "        description=\"n K\",\n",
    "        continuous_update=False,\n",
    "    )\n",
    "\n",
    "    n2_slider_alpha = widgets.IntSlider(\n",
    "        value=15,\n",
    "        min=5,\n",
    "        max=80,\n",
    "        step=1,\n",
    "        description=\"n T\",\n",
    "        continuous_update=False,\n",
    "    )\n",
    "\n",
    "    sd1_slider_alpha = widgets.FloatSlider(\n",
    "        value=2.0,\n",
    "        min=0.5,\n",
    "        max=6.0,\n",
    "        step=0.1,\n",
    "        description=\"odch. K\",\n",
    "        readout_format=\".1f\",\n",
    "        continuous_update=False,\n",
    "    )\n",
    "\n",
    "    sd2_slider_alpha = widgets.FloatSlider(\n",
    "        value=3.0,\n",
    "        min=0.5,\n",
    "        max=6.0,\n",
    "        step=0.1,\n",
    "        description=\"odch. T\",\n",
    "        readout_format=\".1f\",\n",
    "        continuous_update=False,\n",
    "    )\n",
    "\n",
    "    alpha_dropdown = widgets.Dropdown(\n",
    "        options=[(\"0.01\", 0.01), (\"0.05\", 0.05), (\"0.10\", 0.10)],\n",
    "        value=0.05,\n",
    "        description=\"alpha\",\n",
    "    )\n",
    "\n",
    "    repeats_dropdown = widgets.Dropdown(\n",
    "        options=[(\"500\", 500), (\"2000\", 2_000), (\"5000\", 5_000)],\n",
    "        value=2_000,\n",
    "        description=\"symulacje\",\n",
    "    )\n",
    "\n",
    "    simulate_button = widgets.Button(\n",
    "        description=\"Symuluj\",\n",
    "        button_style=\"primary\",\n",
    "    )\n",
    "\n",
    "    info_box = widgets.HTML(\n",
    "        value=(\n",
    "            \"<b>Symulacja przy prawdziwej H0:</b> \"\n",
    "            \"&mu;<sub>kontrola</sub> = &mu;<sub>trening</sub>\"\n",
    "        )\n",
    "    )\n",
    "\n",
    "    alpha_output = widgets.Output()\n",
    "\n",
    "    def klasyfikuj_alfe(emp_alpha, nominal_alpha):\n",
    "        tolerance = max(0.005, 0.15 * nominal_alpha)\n",
    "        if emp_alpha > nominal_alpha + tolerance:\n",
    "            return \"zbyt liberalny\"\n",
    "        if emp_alpha < nominal_alpha - tolerance:\n",
    "            return \"zbyt konserwatywny\"\n",
    "        return \"blisko poziomu nominalnego\"\n",
    "\n",
    "    def uruchom_symulacje_alfa(_):\n",
    "        n1 = int(n1_slider_alpha.value)\n",
    "        n2 = int(n2_slider_alpha.value)\n",
    "        sd1 = float(sd1_slider_alpha.value)\n",
    "        sd2 = float(sd2_slider_alpha.value)\n",
    "        nominal_alpha = float(alpha_dropdown.value)\n",
    "        repeats = int(repeats_dropdown.value)\n",
    "\n",
    "        seed_seq = np.random.SeedSequence(\n",
    "            [SEED, n1, n2, int(sd1 * 100), int(sd2 * 100), int(nominal_alpha * 10_000), repeats]\n",
    "        )\n",
    "        rng_local = np.random.default_rng(seed_seq)\n",
    "\n",
    "        mu0 = 10.0\n",
    "        grupa_kontrola = rng_local.normal(loc=mu0, scale=sd1, size=(repeats, n1))\n",
    "        grupa_trening = rng_local.normal(loc=mu0, scale=sd2, size=(repeats, n2))\n",
    "\n",
    "        mean_kontrola = grupa_kontrola.mean(axis=1)\n",
    "        mean_trening = grupa_trening.mean(axis=1)\n",
    "        diff = mean_trening - mean_kontrola\n",
    "\n",
    "        var_kontrola = grupa_kontrola.var(axis=1, ddof=1)\n",
    "        var_trening = grupa_trening.var(axis=1, ddof=1)\n",
    "\n",
    "        sp2 = ((n1 - 1) * var_kontrola + (n2 - 1) * var_trening) / (n1 + n2 - 2)\n",
    "        se_student = np.sqrt(sp2 * (1 / n1 + 1 / n2))\n",
    "        t_student = diff / se_student\n",
    "        df_student = n1 + n2 - 2\n",
    "        p_student = 2 * stats.t.sf(np.abs(t_student), df=df_student)\n",
    "\n",
    "        se_welch = np.sqrt(var_trening / n2 + var_kontrola / n1)\n",
    "        t_welch = diff / se_welch\n",
    "        df_welch = (var_trening / n2 + var_kontrola / n1) ** 2 / (\n",
    "            (var_trening**2) / (n2**2 * (n2 - 1))\n",
    "            + (var_kontrola**2) / (n1**2 * (n1 - 1))\n",
    "        )\n",
    "        p_welch = 2 * stats.t.sf(np.abs(t_welch), df=df_welch)\n",
    "\n",
    "        emp_alpha_student = float(np.mean(p_student < nominal_alpha))\n",
    "        emp_alpha_welch = float(np.mean(p_welch < nominal_alpha))\n",
    "\n",
    "        reject_student = int(np.sum(p_student < nominal_alpha))\n",
    "        reject_welch = int(np.sum(p_welch < nominal_alpha))\n",
    "\n",
    "        mean_df_welch = float(np.mean(df_welch))\n",
    "        mc_margin = 1.96 * np.sqrt(nominal_alpha * (1 - nominal_alpha) / repeats)\n",
    "\n",
    "        ocena_student = klasyfikuj_alfe(emp_alpha_student, nominal_alpha)\n",
    "        ocena_welch = klasyfikuj_alfe(emp_alpha_welch, nominal_alpha)\n",
    "\n",
    "        alpha_output.clear_output(wait=True)\n",
    "        with alpha_output:\n",
    "            print(\"Symulacja przy prawdziwej H0: μ_kontrola = μ_trening\")\n",
    "            print(\n",
    "                f\"n kontrola = {n1}, n trening = {n2}, \"\n",
    "                f\"odch. kontrola = {sd1:.1f}, odch. trening = {sd2:.1f}\"\n",
    "            )\n",
    "            print(f\"Nominalne alpha = {nominal_alpha:.2f}, liczba symulacji = {repeats}\")\n",
    "            print()\n",
    "            print(\n",
    "                f\"Student: empiryczna alpha = {emp_alpha_student:.3f} \"\n",
    "                f\"({reject_student}/{repeats}), df = {df_student}, ocena: {ocena_student}\"\n",
    "            )\n",
    "            print(\n",
    "                f\"Welch:   empiryczna alpha = {emp_alpha_welch:.3f} \"\n",
    "                f\"({reject_welch}/{repeats}), średnie df ≈ {mean_df_welch:.2f}, \"\n",
    "                f\"ocena: {ocena_welch}\"\n",
    "            )\n",
    "            print(f\"Orientacyjny margines błędu Monte Carlo: ±{mc_margin:.3f}\")\n",
    "\n",
    "            fig, axes = plt.subplots(1, 2, figsize=(12, 4.5))\n",
    "\n",
    "            labels = [\"Student\", \"Welch\"]\n",
    "            alpha_values = [emp_alpha_student, emp_alpha_welch]\n",
    "            colors = [\"darkorange\", \"steelblue\"]\n",
    "\n",
    "            upper_y = max(0.12, nominal_alpha * 2.4, max(alpha_values) + 0.03)\n",
    "            lower_band = max(0.0, nominal_alpha - mc_margin)\n",
    "            upper_band = min(upper_y, nominal_alpha + mc_margin)\n",
    "\n",
    "            axes[0].axhspan(\n",
    "                lower_band,\n",
    "                upper_band,\n",
    "                color=\"0.92\",\n",
    "                label=\"margines Monte Carlo\",\n",
    "            )\n",
    "            bars = axes[0].bar(labels, alpha_values, color=colors, width=0.6)\n",
    "            axes[0].axhline(\n",
    "                nominal_alpha,\n",
    "                color=\"black\",\n",
    "                linestyle=\":\",\n",
    "                linewidth=2,\n",
    "                label=f\"nominalne alpha = {nominal_alpha:.2f}\",\n",
    "            )\n",
    "            axes[0].set_ylim(0, upper_y)\n",
    "            axes[0].set_ylabel(\"empiryczna alfa\")\n",
    "            axes[0].set_title(\"Częstość odrzuceń przy prawdziwej H0\")\n",
    "            axes[0].grid(axis=\"y\", alpha=0.25)\n",
    "            axes[0].legend(loc=\"upper left\")\n",
    "\n",
    "            for bar, value in zip(bars, alpha_values):\n",
    "                axes[0].text(\n",
    "                    bar.get_x() + bar.get_width() / 2,\n",
    "                    value + 0.005,\n",
    "                    f\"{value:.3f}\",\n",
    "                    ha=\"center\",\n",
    "                    va=\"bottom\",\n",
    "                )\n",
    "\n",
    "            df_values = [df_student, mean_df_welch]\n",
    "            df_bars = axes[1].bar(labels, df_values, color=colors, width=0.6)\n",
    "            axes[1].set_ylabel(\"stopnie swobody\")\n",
    "            axes[1].set_title(\"Rozkład odniesienia dla obu testów\")\n",
    "            axes[1].grid(axis=\"y\", alpha=0.25)\n",
    "\n",
    "            for bar, value in zip(df_bars, df_values):\n",
    "                axes[1].text(\n",
    "                    bar.get_x() + bar.get_width() / 2,\n",
    "                    value + 0.2,\n",
    "                    f\"{value:.2f}\",\n",
    "                    ha=\"center\",\n",
    "                    va=\"bottom\",\n",
    "                )\n",
    "\n",
    "            plt.tight_layout()\n",
    "            plt.show()\n",
    "\n",
    "    simulate_button.on_click(uruchom_symulacje_alfa)\n",
    "\n",
    "    controls_box = widgets.VBox(\n",
    "        [\n",
    "            info_box,\n",
    "            widgets.HBox([n1_slider_alpha, n2_slider_alpha]),\n",
    "            widgets.HBox([sd1_slider_alpha, sd2_slider_alpha]),\n",
    "            widgets.HBox([alpha_dropdown, repeats_dropdown]),\n",
    "            simulate_button,\n",
    "        ]\n",
    "    )\n",
    "\n",
    "    display(controls_box, alpha_output)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cdb6a061",
   "metadata": {},
   "source": [
    "## Jak wybrać wariant testu *t*?\n",
    "\n",
    "Najpierw spójrz na jednostkę analizy:\n",
    "\n",
    "- **jedna próba**: masz jedną grupę i jedną wartość odniesienia $\\mu_0$,\n",
    "- **próby zależne**: masz te same osoby lub obiekty w dwóch warunkach; testujesz średnią różnicę,\n",
    "- **próby niezależne**: masz dwie różne grupy; testujesz różnicę średnich.\n",
    "\n",
    "Dodatkowa decyzja pojawia się tylko w ostatnim wariancie:\n",
    "\n",
    "- jeśli nie chcesz mocno zakładać równości wariancji, wybieraj **Welcha**,\n",
    "- test Studenta traktuj jako przypadek szczególny, a nie jako automatyczny domyślny wybór.\n",
    "\n",
    "Niezależnie od wariantu raportuj efekt w jednostkach problemu i 95% PU, a p-wartość traktuj jako jedną z informacji.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "06944512",
   "metadata": {},
   "source": [
    "## Podsumowanie\n",
    "\n",
    "- Każdy test *t* opiera się na tym samym szkielecie: **estymata efektu / błąd standardowy / rozkład $t$**.\n",
    "- W teście dwustronnym z $\\alpha = 0.05$ odrzucenie $H_0$ jest równoważne temu, że wartość z $H_0$ leży poza 95% PU.\n",
    "- Test sparowany to w praktyce test jednopróbkowy wykonany na różnicach w parach.\n",
    "- W dwóch niezależnych grupach bezpiecznym wyborem domyślnym jest test Welcha.\n",
    "- Raportuj zawsze efekt w naturalnej skali problemu i 95% PU; p-wartość traktuj jako jedną z informacji, a nie jedyny wynik analizy.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cf048a6f",
   "metadata": {},
   "source": [
    "## Ćwiczenia\n",
    "\n",
    "1. **Rozpoznanie wariantu.** Dla każdej historii nazwij właściwy test:  \n",
    "   a) jedna grupa studentów vs norma 11 punktów na skali lęku egzaminacyjnego,  \n",
    "   b) te same osoby w ciszy i przy powiadomieniach telefonu,  \n",
    "   c) grupa kontrolna vs grupa po krótkim treningu pamięci.\n",
    "2. **Jedna próba (lęk egzaminacyjny).** Zmień wartość odniesienia $\\mu_0$ (np. 10, 11, 12 pkt). Jak zmieniają się: efekt, 95% PU i p-wartość?\n",
    "3. **Próby zależne (koncentracja).** Dla danych `wynik_cisza` i `wynik_powiadomienia` policz średnią różnicę i 95% PU. Napisz 3-4 zdania interpretacji bez słów „udowodniono”.\n",
    "4. **Próby niezależne (trening pamięci).** Porównaj wyniki testu Studenta i Welcha na tych samych danych. Co jest identyczne, a co może się różnić?\n",
    "5. **(Opcjonalnie) Student vs Welch.** W widżecie ustaw takie parametry, żeby oba testy dały bardzo podobny wynik, a potem takie, żeby różnica była wyraźniejsza. Spróbuj opisać, skąd bierze się ta rozbieżność.\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.14.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}