t_test_independent_two_tailes.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib as mlt\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import statsmodels.formula.api as smf\n",
    "import statsmodels.stats.api as sms\n",
    "import statsmodels.stats.power as smp\n",
    "import scipy.stats\n",
    "from scipy.stats import t, norm, ttest_ind, ttest_rel, ttest_1samp\n",
    "import decimal\n",
    "import random\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def conf_int_two_sample(x,y,alfa,alternative = 'two-tailed'):\n",
    "    '''\n",
    "    alternative == 'two-tailed','left','right'\n",
    "    '''\n",
    "    mean1 = np.mean(x)\n",
    "    mean2 = np.mean(y)\n",
    "    mean = mean1-mean2\n",
    "    n1 = np.size(x)\n",
    "    n2 = np.size(y)\n",
    "    var1 = np.var(x,ddof=1)\n",
    "    var2 = np.var(y,ddof=1)\n",
    "    se = np.sqrt(var1/n1+var2/n2)\n",
    "    if np.mean(x)-np.mean(y) >= 0:\n",
    "        ts = (np.mean(x)-np.mean(y)-diff)/se\n",
    "    else:\n",
    "        ts = (np.mean(y)-np.mean(x)-diff)/se\n",
    "    df = (var1/n1+var2/n2)**2/((var1/n1)**2/(n1-1)+(var2/n2)**2/(n2-1))\n",
    "    #se = np.std(x,ddof=1)/np.sqrt(len(x))\n",
    "    if alternative == 'two-tailed':\n",
    "        return mean - t.ppf(1-alfa/2,df=df)*se,mean + t.ppf(1-alfa/2,df=df)*se\n",
    "    elif alternative == 'left':\n",
    "        return mean - t.ppf(1-alfa/2,df=df)*se\n",
    "    elif alternative == 'right':\n",
    "        return mean + t.ppf(1-alfa/2,df=df)*se"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_normal_sample(n,mean,sd):\n",
    "    return np.random.normal(loc=mean,scale=sd,size=n)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def t_test_for_independent_groups(x,y,diff = 0, alternative = 'two-tailed'): #Welch\n",
    "    n1 = np.size(x)\n",
    "    n2 = np.size(y)\n",
    "    var1 = np.var(x,ddof=1)\n",
    "    var2 = np.var(y,ddof=1)\n",
    "    se = np.sqrt(var1/n1+var2/n2)\n",
    "    ts = (np.mean(x)-np.mean(y)-diff)/se\n",
    "    df = (var1/n1+var2/n2)**2/((var1/n1)**2/(n1-1)+(var2/n2)**2/(n2-1)) #Welch\n",
    "    #if alternative == 'two-tailed':\n",
    "    p_val = 2*t.sf(np.abs(ts),df)\n",
    "    if alternative == 'two-tailed':\n",
    "        p_val = 2*t.sf(np.abs(ts),df)\n",
    "    if alternative == 'greater':\n",
    "        p_val = t.sf(ts,df)\n",
    "    if alternative == 'less':\n",
    "        p_val = t.cdf(ts, df)\n",
    "    return p_val"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def independent_1(sample1,sample2,diff=0, eps = 0):\n",
    "    mean1 = np.mean(sample1)\n",
    "    mean2 = np.mean(sample2)\n",
    "    mean = mean1-mean2\n",
    "    \n",
    "    if mean > diff+eps:\n",
    "        return 'd'\n",
    "    if  mean <= diff-eps:\n",
    "        return 'm'\n",
    "    if mean < delta+eps and mean > delta-eps  : #problematyczne eps > delta\n",
    "        return 'nwd'\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "def independent_2(sample1, sample2, alfa, diff=0):\n",
    "    mean = np.mean(sample1) - np.mean(sample2)\n",
    "    #if mean >= 0:\n",
    "    result = t_test_for_independent_groups(sample1,sample2,diff = 0, alternative = 'greater')\n",
    "    #else:\n",
    "    #    result = t_test_for_independent_groups(sample1,sample2,diff = 0, alternative = 'less')\n",
    "    \n",
    "    if result < alfa and np.abs(mean) > np.abs(diff):\n",
    "        return 'd'\n",
    "    elif result > alfa and (np.abs(mean) <= np.abs(diff)):\n",
    "        return 'm'\n",
    "    else:\n",
    "        return 'nwd'    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "def independent_3(sample1, sample2, alfa, diff=0):\n",
    "    mean = np.mean(sample1) - np.mean(sample2)\n",
    "    #if mean >= 0:\n",
    "    result = t_test_for_independent_groups(sample1,sample2,diff = diff, alternative = 'greater')\n",
    "    result2 = t_test_for_independent_groups(sample1,sample2,diff = diff,alternative='less')\n",
    "    #else:\n",
    "    #    result = t_test_for_independent_groups(sample1,sample2,diff = diff, alternative = 'less')\n",
    "    #    result2 = t_test_for_independent_groups(sample1,sample2,diff = diff,alternative='greater')\n",
    "    \n",
    "    if result < alfa and np.abs(mean) > np.abs(diff):\n",
    "        return 'd'\n",
    "    elif (result2 < alfa) and (np.abs(mean) <= np.abs(diff)):\n",
    "        return 'm'\n",
    "    #if result >= alfa:\n",
    "    else:\n",
    "        return 'nwd'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "duże1 = []\n",
    "małe1 = []\n",
    "duże2 = []\n",
    "duże3 = []\n",
    "małe2 = []\n",
    "małe3 = []\n",
    "źle_duże1 = []\n",
    "niewiadomo_duże1 = []\n",
    "źle_duże2 = []\n",
    "niewiadomo_duże2 = []\n",
    "źle_duże3 = []\n",
    "niewiadomo_duże3 = []\n",
    "źle_małe1 = []\n",
    "niewiadomo_małe1 = []\n",
    "źle_małe2 = []\n",
    "niewiadomo_małe2 = []\n",
    "źle_małe3 = []\n",
    "niewiadomo_małe3 = []\n",
    "przedziały_n = []\n",
    "ile_dużych = []\n",
    "ile_małych = []\n",
    "means = []\n",
    "delta = 1\n",
    "mean = 0\n",
    "eps = 1\n",
    "sds = [i*0.1 for i in range(20,50)]\n",
    "for k in range(1,25): \n",
    "    p1,p2 = 15+k*5,20+k*5\n",
    "    for sd in sds:\n",
    "        przedziały_n.append(p1)\n",
    "        pod1 = []\n",
    "        pod2 = []\n",
    "        pod3 = []\n",
    "        for n in range(p1,p2):\n",
    "            for j in range(0,30):\n",
    "                pop_mean = j*0.1\n",
    "                alfa = 0.05\n",
    "                sample1 = get_normal_sample(n,mean+pop_mean,sd)\n",
    "                sample2 = get_normal_sample(n,mean,sd)\n",
    "                pod1.append([independent_1(sample1,sample2,delta,eps),pop_mean,eps])\n",
    "                pod2.append([independent_2(sample1,sample2,alfa,delta),pop_mean,alfa])\n",
    "                pod3.append([independent_3(sample1,sample2,alfa,delta),pop_mean,alfa])\n",
    "    \n",
    "        # --------------------------------------------------      wśród wyników które były duże, ile z nich wskazał test - (wyniki wskazane przez test)/(faktycznie duże testy\n",
    "        #---------------------------------------------------      ile razy test wskazał wynik mały, gdy ten w rzeczywistości był duży\n",
    "        #---------------------------------------------------      ile razy test wskazał \"nie wiadomo\", gdy efekt w rzeczywistości był duży\n",
    "        #podejście 1\n",
    "        duże_pod1 = [i for i in pod1 if np.abs(i[1]) > np.abs(delta)]\n",
    "\n",
    "        if len(duże_pod1) != 0:\n",
    "            pod1_dobrze_duże = len([i for i in duże_pod1 if i[0] == 'd'])/len(duże_pod1)\n",
    "            pod1_źle_duże = len([i for i in duże_pod1 if i[0] == 'm'])/len(duże_pod1)\n",
    "            pod1_niewiadomo_duże = len([i for i in duże_pod1 if i[0] == 'nwd'])/len(duże_pod1)\n",
    "        else:\n",
    "            pod1_dobrze_duże = 'brak'\n",
    "            pod1_źle_duże = 'brak'\n",
    "            pod1_niewiadomo_duże = 'brak'\n",
    "        #podejście 2\n",
    "        duże_pod2 = [i for i in pod2 if np.abs(i[1]) > np.abs(delta)]\n",
    "        if len(duże_pod2) != 0:\n",
    "            pod2_dobrze_duże = len([i for i in duże_pod2 if i[0] == 'd'])/len(duże_pod2)\n",
    "            pod2_źle_duże = len([i for i in duże_pod2 if i[0] == 'm'])/len(duże_pod2)\n",
    "            pod2_niewiadomo_duże = len([i for i in duże_pod2 if i[0] == 'nwd'])/len(duże_pod2)\n",
    "        else:\n",
    "            pod2_dobrze_duże = 'brak'\n",
    "            pod2_źle_duże = 'brak'\n",
    "            pod2_niewiadomo_duże = 'brak'\n",
    "        #podejście 3\n",
    "        duże_pod3 = [i for i in pod3 if np.abs(i[1]) > np.abs(delta)]\n",
    "        if len(duże_pod3) != 0:\n",
    "            pod3_dobrze_duże = len([i for i in duże_pod3 if i[0] == 'd'])/len(duże_pod3)\n",
    "            pod3_źle_duże = len([i for i in duże_pod3 if i[0] == 'm'])/len(duże_pod3)\n",
    "            pod3_niewiadomo_duże = len([i for i in duże_pod3 if i[0] == 'nwd'])/len(duże_pod3)\n",
    "        else:\n",
    "            pod3_dobrze_duże = 'brak'\n",
    "            pod3_źle_duże = 'brak'\n",
    "            pod3_niewiadomo_duże = 'brak'\n",
    "\n",
    "        # --------------------------------------------------      wśród wyników które były małe, ile z nich wskazał test - (wyniki małe wskazane przez test)/(faktycznie małe wyniki)\n",
    "        #---------------------------------------------------      ile razy test wskazał wynik duży, gdy ten w rzeczywistości był mały\n",
    "        #---------------------------------------------------      ile razy test wskazał \"nie wiadomo\", gdy efekt w rzeczywistości był duży\n",
    "        #podejście 1\n",
    "        małe_pod1 = [i for i in pod1 if (np.abs(i[1]) < np.abs(delta))]\n",
    "        if len(małe_pod1) != 0:\n",
    "            pod1_dobrze_małe = len([i for i in małe_pod1 if i[0] == 'm'])/len(małe_pod1)\n",
    "            pod1_źle_małe = len([i for i in małe_pod1 if i[0] == 'd'])/len(małe_pod1)\n",
    "            pod1_niewiadomo_małe = len([i for i in małe_pod1 if i[0] == 'nwd'])/len(małe_pod1)\n",
    "        else:\n",
    "            pod1_dobrze_małe = 'brak'\n",
    "            pod1_źle_małe = 'brak'\n",
    "            pod1_niewiadomo_małe = 'brak'\n",
    "\n",
    "        #podejście 2\n",
    "        małe_pod2 = [i for i in pod2 if np.abs(i[1]) < np.abs(delta)]\n",
    "        if len(małe_pod2) != 0:\n",
    "            pod2_dobrze_małe = len([i for i in małe_pod2 if i[0] == 'm'])/len(małe_pod2)\n",
    "            pod2_źle_małe = len([i for i in małe_pod2 if i[0] == 'd'])/len(małe_pod2)\n",
    "            pod2_niewiadomo_małe = len([i for i in małe_pod2 if i[0] == 'nwd'])/len(małe_pod2)\n",
    "        else:\n",
    "            pod2_dobrze_małe = 'brak'\n",
    "            pod2_źle_małe = 'brak'\n",
    "            pod2_niewiadomo_małe = 'brak'\n",
    "\n",
    "        #podejście 3\n",
    "        małe_pod3 = [i for i in pod3 if np.abs(i[1]) < np.abs(delta)]\n",
    "        if len(małe_pod3) != 0:\n",
    "            pod3_dobrze_małe = len([i for i in małe_pod3 if i[0] == 'm'])/len(małe_pod3)\n",
    "            pod3_źle_małe = len([i for i in małe_pod3 if i[0] == 'd'])/len(małe_pod3)\n",
    "            pod3_niewiadomo_małe = len([i for i in małe_pod3 if i[0] == 'nwd'])/len(małe_pod3)\n",
    "        else:\n",
    "            pod3_dobrze_małe = 'brak'\n",
    "            pod3_źle_małe = 'brak'\n",
    "            pod3_niewiadomo_małe = 'brak'\n",
    "\n",
    "\n",
    "        duże1.append(pod1_dobrze_duże)\n",
    "        małe1.append(pod1_dobrze_małe)\n",
    "        duże2.append(pod2_dobrze_duże)\n",
    "        duże3.append(pod3_dobrze_duże)\n",
    "        małe2.append(pod2_dobrze_małe)\n",
    "        małe3.append(pod3_dobrze_małe)\n",
    "\n",
    "        źle_duże1.append(pod1_źle_duże)\n",
    "        źle_duże2.append(pod2_źle_duże)\n",
    "        źle_duże3.append(pod3_źle_duże)\n",
    "        niewiadomo_duże1.append(pod1_niewiadomo_duże)\n",
    "        niewiadomo_duże2.append(pod2_niewiadomo_duże)\n",
    "        niewiadomo_duże3.append(pod3_niewiadomo_duże)\n",
    "\n",
    "        źle_małe1.append(pod1_źle_małe)\n",
    "        źle_małe2.append(pod2_źle_małe)\n",
    "        źle_małe3.append(pod3_źle_małe)\n",
    "        niewiadomo_małe1.append(pod1_niewiadomo_małe)\n",
    "        niewiadomo_małe2.append(pod2_niewiadomo_małe)\n",
    "        niewiadomo_małe3.append(pod3_niewiadomo_małe)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "sdss = sds*25"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>wielkość próby</th>\n",
       "      <th>sd</th>\n",
       "      <th>efekt duży, podejście1</th>\n",
       "      <th>efekt duży, podejście2</th>\n",
       "      <th>efekt duży, podejście3</th>\n",
       "      <th>mały_podejście1</th>\n",
       "      <th>mały_podejście2</th>\n",
       "      <th>mały_podejście3</th>\n",
       "      <th>źle_duże1</th>\n",
       "      <th>źle_duże2</th>\n",
       "      <th>źle_duże3</th>\n",
       "      <th>nie wiadomo duże1</th>\n",
       "      <th>nie wiadomo duże2</th>\n",
       "      <th>nie wiadomo duże3</th>\n",
       "      <th>źle_małe1</th>\n",
       "      <th>źle_małe2</th>\n",
       "      <th>źle_małe3</th>\n",
       "      <th>niewiadomo_małe1</th>\n",
       "      <th>niewiadomo_małe2</th>\n",
       "      <th>niewiadomo_małe3</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>20</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.473684</td>\n",
       "      <td>0.905263</td>\n",
       "      <td>0.473684</td>\n",
       "      <td>0.16</td>\n",
       "      <td>0.82</td>\n",
       "      <td>0.16</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.073684</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.526316</td>\n",
       "      <td>0.021053</td>\n",
       "      <td>0.84</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.18</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.84</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.84</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>20</td>\n",
       "      <td>2.1</td>\n",
       "      <td>0.536842</td>\n",
       "      <td>0.852632</td>\n",
       "      <td>0.494737</td>\n",
       "      <td>0.22</td>\n",
       "      <td>0.78</td>\n",
       "      <td>0.18</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.105263</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.463158</td>\n",
       "      <td>0.042105</td>\n",
       "      <td>0.82</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.78</td>\n",
       "      <td>0.02</td>\n",
       "      <td>0.82</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>20</td>\n",
       "      <td>2.2</td>\n",
       "      <td>0.536842</td>\n",
       "      <td>0.873684</td>\n",
       "      <td>0.463158</td>\n",
       "      <td>0.38</td>\n",
       "      <td>0.84</td>\n",
       "      <td>0.28</td>\n",
       "      <td>0.010526</td>\n",
       "      <td>0.094737</td>\n",
       "      <td>0.021053</td>\n",
       "      <td>0.452632</td>\n",
       "      <td>0.031579</td>\n",
       "      <td>0.66</td>\n",
       "      <td>0.08</td>\n",
       "      <td>0.12</td>\n",
       "      <td>0.06</td>\n",
       "      <td>0.54</td>\n",
       "      <td>0.04</td>\n",
       "      <td>0.66</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>20</td>\n",
       "      <td>2.3</td>\n",
       "      <td>0.505263</td>\n",
       "      <td>0.800000</td>\n",
       "      <td>0.421053</td>\n",
       "      <td>0.34</td>\n",
       "      <td>0.80</td>\n",
       "      <td>0.22</td>\n",
       "      <td>0.021053</td>\n",
       "      <td>0.136842</td>\n",
       "      <td>0.021053</td>\n",
       "      <td>0.473684</td>\n",
       "      <td>0.063158</td>\n",
       "      <td>0.78</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.18</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.66</td>\n",
       "      <td>0.02</td>\n",
       "      <td>0.78</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>20</td>\n",
       "      <td>2.4</td>\n",
       "      <td>0.484211</td>\n",
       "      <td>0.800000</td>\n",
       "      <td>0.410526</td>\n",
       "      <td>0.26</td>\n",
       "      <td>0.72</td>\n",
       "      <td>0.14</td>\n",
       "      <td>0.010526</td>\n",
       "      <td>0.136842</td>\n",
       "      <td>0.010526</td>\n",
       "      <td>0.505263</td>\n",
       "      <td>0.063158</td>\n",
       "      <td>0.84</td>\n",
       "      <td>0.02</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.02</td>\n",
       "      <td>0.72</td>\n",
       "      <td>0.08</td>\n",
       "      <td>0.84</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>715</th>\n",
       "      <td>135</td>\n",
       "      <td>4.5</td>\n",
       "      <td>0.505263</td>\n",
       "      <td>0.926316</td>\n",
       "      <td>0.568421</td>\n",
       "      <td>0.12</td>\n",
       "      <td>0.80</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.052632</td>\n",
       "      <td>0.021053</td>\n",
       "      <td>0.494737</td>\n",
       "      <td>0.021053</td>\n",
       "      <td>0.80</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.08</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.88</td>\n",
       "      <td>0.12</td>\n",
       "      <td>0.80</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>716</th>\n",
       "      <td>135</td>\n",
       "      <td>4.6</td>\n",
       "      <td>0.515789</td>\n",
       "      <td>0.936842</td>\n",
       "      <td>0.547368</td>\n",
       "      <td>0.28</td>\n",
       "      <td>0.76</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.063158</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.484211</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.70</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.22</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.72</td>\n",
       "      <td>0.02</td>\n",
       "      <td>0.70</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>717</th>\n",
       "      <td>135</td>\n",
       "      <td>4.7</td>\n",
       "      <td>0.526316</td>\n",
       "      <td>0.915789</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>0.36</td>\n",
       "      <td>0.82</td>\n",
       "      <td>0.36</td>\n",
       "      <td>0.010526</td>\n",
       "      <td>0.084211</td>\n",
       "      <td>0.010526</td>\n",
       "      <td>0.463158</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.64</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.16</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.64</td>\n",
       "      <td>0.02</td>\n",
       "      <td>0.64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>718</th>\n",
       "      <td>135</td>\n",
       "      <td>4.8</td>\n",
       "      <td>0.473684</td>\n",
       "      <td>0.905263</td>\n",
       "      <td>0.494737</td>\n",
       "      <td>0.28</td>\n",
       "      <td>0.86</td>\n",
       "      <td>0.32</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.084211</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.526316</td>\n",
       "      <td>0.010526</td>\n",
       "      <td>0.66</td>\n",
       "      <td>0.02</td>\n",
       "      <td>0.14</td>\n",
       "      <td>0.02</td>\n",
       "      <td>0.70</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.66</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>719</th>\n",
       "      <td>135</td>\n",
       "      <td>4.9</td>\n",
       "      <td>0.547368</td>\n",
       "      <td>0.863158</td>\n",
       "      <td>0.568421</td>\n",
       "      <td>0.36</td>\n",
       "      <td>0.74</td>\n",
       "      <td>0.36</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.126316</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.452632</td>\n",
       "      <td>0.010526</td>\n",
       "      <td>0.64</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.24</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.64</td>\n",
       "      <td>0.02</td>\n",
       "      <td>0.64</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>720 rows × 20 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     wielkość próby   sd  efekt duży, podejście1  efekt duży, podejście2  \\\n",
       "0                20  2.0                0.473684                0.905263   \n",
       "1                20  2.1                0.536842                0.852632   \n",
       "2                20  2.2                0.536842                0.873684   \n",
       "3                20  2.3                0.505263                0.800000   \n",
       "4                20  2.4                0.484211                0.800000   \n",
       "..              ...  ...                     ...                     ...   \n",
       "715             135  4.5                0.505263                0.926316   \n",
       "716             135  4.6                0.515789                0.936842   \n",
       "717             135  4.7                0.526316                0.915789   \n",
       "718             135  4.8                0.473684                0.905263   \n",
       "719             135  4.9                0.547368                0.863158   \n",
       "\n",
       "     efekt duży, podejście3  mały_podejście1  mały_podejście2  \\\n",
       "0                  0.473684             0.16             0.82   \n",
       "1                  0.494737             0.22             0.78   \n",
       "2                  0.463158             0.38             0.84   \n",
       "3                  0.421053             0.34             0.80   \n",
       "4                  0.410526             0.26             0.72   \n",
       "..                      ...              ...              ...   \n",
       "715                0.568421             0.12             0.80   \n",
       "716                0.547368             0.28             0.76   \n",
       "717                0.600000             0.36             0.82   \n",
       "718                0.494737             0.28             0.86   \n",
       "719                0.568421             0.36             0.74   \n",
       "\n",
       "     mały_podejście3  źle_duże1  źle_duże2  źle_duże3  nie wiadomo duże1  \\\n",
       "0               0.16   0.000000   0.073684   0.000000           0.526316   \n",
       "1               0.18   0.000000   0.105263   0.000000           0.463158   \n",
       "2               0.28   0.010526   0.094737   0.021053           0.452632   \n",
       "3               0.22   0.021053   0.136842   0.021053           0.473684   \n",
       "4               0.14   0.010526   0.136842   0.010526           0.505263   \n",
       "..               ...        ...        ...        ...                ...   \n",
       "715             0.20   0.000000   0.052632   0.021053           0.494737   \n",
       "716             0.30   0.000000   0.063158   0.000000           0.484211   \n",
       "717             0.36   0.010526   0.084211   0.010526           0.463158   \n",
       "718             0.32   0.000000   0.084211   0.000000           0.526316   \n",
       "719             0.36   0.000000   0.126316   0.000000           0.452632   \n",
       "\n",
       "     nie wiadomo duże2  nie wiadomo duże3  źle_małe1  źle_małe2  źle_małe3  \\\n",
       "0             0.021053               0.84       0.00       0.18       0.00   \n",
       "1             0.042105               0.82       0.00       0.20       0.00   \n",
       "2             0.031579               0.66       0.08       0.12       0.06   \n",
       "3             0.063158               0.78       0.00       0.18       0.00   \n",
       "4             0.063158               0.84       0.02       0.20       0.02   \n",
       "..                 ...                ...        ...        ...        ...   \n",
       "715           0.021053               0.80       0.00       0.08       0.00   \n",
       "716           0.000000               0.70       0.00       0.22       0.00   \n",
       "717           0.000000               0.64       0.00       0.16       0.00   \n",
       "718           0.010526               0.66       0.02       0.14       0.02   \n",
       "719           0.010526               0.64       0.00       0.24       0.00   \n",
       "\n",
       "     niewiadomo_małe1  niewiadomo_małe2  niewiadomo_małe3  \n",
       "0                0.84              0.00              0.84  \n",
       "1                0.78              0.02              0.82  \n",
       "2                0.54              0.04              0.66  \n",
       "3                0.66              0.02              0.78  \n",
       "4                0.72              0.08              0.84  \n",
       "..                ...               ...               ...  \n",
       "715              0.88              0.12              0.80  \n",
       "716              0.72              0.02              0.70  \n",
       "717              0.64              0.02              0.64  \n",
       "718              0.70              0.00              0.66  \n",
       "719              0.64              0.02              0.64  \n",
       "\n",
       "[720 rows x 20 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trzy_podejścia = pd.DataFrame(zip(przedziały_n,sdss,duże1,duże2,duże3,małe1,małe2,małe3,źle_duże1,źle_duże2,źle_duże3,niewiadomo_duże1,niewiadomo_duże2,niewiadomo_małe3,źle_małe1,źle_małe2,źle_małe3,niewiadomo_małe1,\n",
    "                                  niewiadomo_małe2,niewiadomo_małe3), columns = ['wielkość próby','sd','efekt duży, podejście1','efekt duży, podejście2','efekt duży, podejście3','mały_podejście1','mały_podejście2','mały_podejście3',\n",
    "                                                                                 'źle_duże1','źle_duże2','źle_duże3','nie wiadomo duże1','nie wiadomo duże2','nie wiadomo duże3','źle_małe1','źle_małe2','źle_małe3',\n",
    "                                                                                 'niewiadomo_małe1','niewiadomo_małe2','niewiadomo_małe3'])\n",
    "trzy_podejścia"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 11.6 s\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def rysuj():\n",
    "    fig, (ax1,ax2) = plt.subplots(2,sharex=True)\n",
    "    n1 = 50\n",
    "    n2 = 50\n",
    "    sd1 = 2\n",
    "    sd2 = 2\n",
    "    delta = 3\n",
    "    mean1 = 0\n",
    "    ax1.set_xlim(-6,6)\n",
    "    diffs = [i*0.1 for i in range(-60,61)]\n",
    "    for diff in diffs:\n",
    "        if diff >= 0:\n",
    "            delta = 3\n",
    "        else:\n",
    "            delta = -3\n",
    "        sample1 = get_normal_sample(n1,mean1+diff,sd1)\n",
    "        sample2 = get_normal_sample(n2,mean1,sd2)\n",
    "        res2=[independent_2(sample1,sample2,i*0.01,delta) for i in range(1,101)]\n",
    "        res3=[independent_3(sample1,sample2,i*0.01,delta) for i in range(1,101)]\n",
    "        result = diff\n",
    "        res2_vals = {'m':[1-(i+1)*0.01 for i,x in enumerate(res2) if x == 'm'],'d':[1-(i+1)*0.01 for i,x in enumerate(res2) if x == 'd'],'nwd':[1-(i+1)*0.01 for i,x in enumerate(res2) if x == 'nwd']}\n",
    "        res3_vals = {'m':[1-(i+1)*0.01 for i,x in enumerate(res3) if x == 'm'],'d':[1-(i+1)*0.01 for i,x in enumerate(res3) if x == 'd'],'nwd':[1-(i+1)*0.01 for i,x in enumerate(res3) if x == 'nwd']}\n",
    "        \n",
    "        \n",
    "        if np.abs(result) >= delta :\n",
    "            if len(res2_vals['m']) != 0:\n",
    "                ax1.axvline(diff,ymin=res2_vals['m'][-1],ymax=res2_vals['m'][0], color = 'red')\n",
    "            if len(res2_vals['d']) != 0:\n",
    "                ax1.axvline(diff,ymin=res2_vals['d'][-1],ymax=res2_vals['d'][0], color = 'green')#,label = 'efekt duży')\n",
    "            if len(res2_vals['nwd']) != 0:\n",
    "                ax1.axvline(diff,ymin=res2_vals['nwd'][-1],ymax=res2_vals['nwd'][0], color = 'black')#,label = 'nie wiadomo')            \n",
    "            if len(res3_vals['m']) != 0:\n",
    "                ax2.axvline(diff,ymin=res3_vals['m'][-1],ymax=res3_vals['m'][0], color = 'red')\n",
    "            if len(res3_vals['d']) != 0:\n",
    "                ax2.axvline(diff,ymin=res3_vals['d'][-1],ymax=res3_vals['d'][0], color = 'green')#,label = 'efekt duży')\n",
    "            if len(res3_vals['nwd']) != 0:\n",
    "                ax2.axvline(diff,ymin=res3_vals['nwd'][-1],ymax=res3_vals['nwd'][0], color = 'black')#,label = 'nie wiadomo')\n",
    "\n",
    "        else:\n",
    "\n",
    "            if len(res2_vals['m']) != 0:\n",
    "                ax1.axvline(diff,ymin=res2_vals['m'][-1],ymax=res2_vals['m'][0], color = 'blue')#,label = 'efekt mały')\n",
    "            if len(res2_vals['d']) != 0:\n",
    "                ax1.axvline(diff,ymin=res2_vals['d'][-1],ymax=res2_vals['d'][0], color = 'red')\n",
    "            if len(res2_vals['nwd']) != 0:\n",
    "                ax1.axvline(diff,ymin=res2_vals['nwd'][-1],ymax=res2_vals['nwd'][0], color = 'black')#,label = 'nie wiadomo')            \n",
    "            if len(res3_vals['m']) != 0:\n",
    "                ax2.axvline(diff,ymin=res3_vals['m'][-1],ymax=res3_vals['m'][0], color = 'blue')#,label = 'efekt mały')\n",
    "            if len(res3_vals['d']) != 0:\n",
    "                ax2.axvline(diff,ymin=res3_vals['d'][-1],ymax=res3_vals['d'][0], color = 'red')\n",
    "            if len(res3_vals['nwd']) != 0:\n",
    "                ax2.axvline(diff,ymin=res3_vals['nwd'][-1],ymax=res3_vals['nwd'][0], color = 'black')#,label = 'nie wiadomo')\n",
    "    \n",
    "    ax1.axvline(0,ymin=0,ymax=0,color='red',label = 'błędna odpowiedź')\n",
    "    ax1.axvline(0,ymin=0,ymax=0,color='green',label = 'efekt duży')\n",
    "    ax1.axvline(0,ymin=0,ymax=0,color='blue',label = 'efekt mały')\n",
    "    ax1.axvline(0,ymin=0,ymax=0,color='black',label = 'nie wiadomo')\n",
    "    ax1.set_ylabel('poziom ufności')\n",
    "    ax2.set_xlabel('poziom ufności dla poszczególnych odpowiedzi w zależności od średniej populacyjnej, delta=3')\n",
    "    mlt.rcParams.update(mlt.rcParamsDefault)\n",
    "\n",
    "    fig.legend(frameon=False, loc = 'center', bbox_to_anchor = (0.5,1))\n",
    "        \n",
    "        \n",
    "%time rysuj()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0002223307845566902"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dane = pd.read_csv('data.csv')\n",
    "x1 = dane['x1']\n",
    "x2 = dane['x2']\n",
    "#np.mean(x1) - np.mean(x2)\n",
    "t_test_for_independent_groups(x1,x2,diff=4,alternative='greater')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}