{ "cells": [ { "cell_type": "markdown", "id": "46fcf466", "metadata": {}, "source": [ "# Career Guidance Participation, Academic Characteristics and Placement Status\n", "## Reproducible analysis notebook (Analyses 1–12)\n", "\n", "**Dataset:** `Career_Guidance_Data_Final.xlsx` (N = 500 final-year undergraduates)\n", "\n", "**Research questions**\n", "- RQ1: What are the participation patterns in institutional engagement activities, and what are the placement outcomes among final-year students?\n", "- RQ2: Is participation in institutional engagement activities significantly associated with placement status?\n", "- RQ3: Are attendance and academic performance associated with placement status when participation is taken into consideration?\n", "\n", "**Hypotheses**\n", "- H1: Participation is significantly associated with placement status.\n", "- H2: Academic performance (CGPA band) is significantly associated with placement status.\n", "- H3: Attendance is significantly associated with placement status.\n", "\n", "**Coding:** Yes = 1, No = 0. Academic performance: 1 = CGPA 5–6, 2 = 6–7, 3 = 7–8, 4 = 8–9, 5 = 9–10.\n", "\n", "All analyses are associational (cross-sectional, observational). Random seed = 2026 for all resampling procedures.\n" ] }, { "cell_type": "code", "execution_count": 1, "id": "123931e9", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:11:40.742857Z", "iopub.status.busy": "2026-09-27T19:11:40.742235Z", "iopub.status.idle": "2026-09-27T19:11:44.782753Z", "shell.execute_reply": "2026-09-27T19:11:44.781430Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Times New Roman installed: False | fallback Liberation Serif: True\n", "0.15.0 scipy.stats 2.4.4 3.0.2\n" ] } ], "source": [ "import warnings, json, os\n", "warnings.filterwarnings(\"ignore\")\n", "import numpy as np, pandas as pd\n", "import scipy.stats as st\n", "import statsmodels.api as sm\n", "import statsmodels.formula.api as smf\n", "from statsmodels.stats.outliers_influence import variance_inflation_factor\n", "from statsmodels.stats.proportion import proportion_confint\n", "from sklearn.metrics import roc_auc_score, brier_score_loss\n", "import matplotlib, matplotlib.pyplot as plt\n", "from matplotlib import font_manager\n", "\n", "SEED = 2026\n", "rng = np.random.default_rng(SEED)\n", "FIG = \"figures\"; OUT = \"outputs\"\n", "os.makedirs(FIG, exist_ok=True); os.makedirs(OUT, exist_ok=True)\n", "\n", "# ---- Figure style: Times New Roman, large legible fonts ----\n", "# 'Times New Roman' is used when installed; 'Liberation Serif' (metric-identical) is the fallback.\n", "matplotlib.rcParams.update({\n", " \"font.family\": \"serif\",\n", " \"font.serif\": [\"Times New Roman\", \"Liberation Serif\", \"Tinos\", \"DejaVu Serif\"],\n", " \"mathtext.fontset\": \"custom\", \"mathtext.rm\": \"Times New Roman\", \"mathtext.it\": \"Times New Roman:italic\",\n", " \"font.size\": 18, \"axes.titlesize\": 19, \"axes.labelsize\": 19,\n", " \"xtick.labelsize\": 17, \"ytick.labelsize\": 17, \"legend.fontsize\": 16,\n", " \"axes.linewidth\": 1.3, \"xtick.major.width\": 1.3, \"ytick.major.width\": 1.3,\n", " \"axes.spines.top\": False, \"axes.spines.right\": False,\n", " \"savefig.dpi\": 600, \"savefig.bbox\": \"tight\", \"pdf.fonttype\": 42, \"svg.fonttype\": \"none\",\n", "})\n", "avail = {f.name for f in font_manager.fontManager.ttflist}\n", "print(\"Times New Roman installed:\", \"Times New Roman\" in avail, \"| fallback Liberation Serif:\", \"Liberation Serif\" in avail)\n", "print(sm.__version__, st.__name__, np.__version__, pd.__version__)\n", "\n", "def savefig(fig, name):\n", " for ext in [\"png\", \"tiff\", \"pdf\"]:\n", " kw = {\"pil_kwargs\": {\"compression\": \"tiff_lzw\"}} if ext == \"tiff\" else {}\n", " fig.savefig(f\"{FIG}/{name}.{ext}\", **kw)\n", " plt.show(); plt.close(fig)\n", "\n", "def fmt_p(p):\n", " return \"< .001\" if p < .001 else f\"{p:.3f}\".replace(\"0.\", \".\", 1)\n", "\n", "R = {} # container for every number reported in the manuscript\n" ] }, { "cell_type": "markdown", "id": "955de758", "metadata": {}, "source": [ "## Analysis 1 — Data screening and coding audit" ] }, { "cell_type": "code", "execution_count": 2, "id": "d2d334e7", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:11:44.785214Z", "iopub.status.busy": "2026-09-27T19:11:44.784885Z", "iopub.status.idle": "2026-09-27T19:11:45.156691Z", "shell.execute_reply": "2026-09-27T19:11:45.155634Z" } }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
CheckResultNote
0Records500500 expected (89.3% of 560 invited)
1Variables28
2Missing values (all cells)0
3Exact duplicate records0
4Out-of-range values: Age [18-30]0observed 20-23
5Out-of-range values: Attendance (%) [0-100]0observed 50-90
6Out-of-range values: Academic Performance [1-5]0observed 1-5
7Out-of-range values: Participated in Engagemen...0observed 0-1
8Out-of-range values: Secured Job [0-1]0observed 0-1
9Participants reporting a substantive non-parti...00 = consistent skip logic
10Non-participants coded 'Other' reason6'Other' also used as not-applicable code for p...
11Academic Year fully determined by Field of Stu...11 = yes; Academic Year cannot enter a model wi...
12Rows where the five global perception items ar...500500 = items are response-for-response duplicates
\n", "
" ], "text/plain": [ " Check Result \\\n", "0 Records 500 \n", "1 Variables 28 \n", "2 Missing values (all cells) 0 \n", "3 Exact duplicate records 0 \n", "4 Out-of-range values: Age [18-30] 0 \n", "5 Out-of-range values: Attendance (%) [0-100] 0 \n", "6 Out-of-range values: Academic Performance [1-5] 0 \n", "7 Out-of-range values: Participated in Engagemen... 0 \n", "8 Out-of-range values: Secured Job [0-1] 0 \n", "9 Participants reporting a substantive non-parti... 0 \n", "10 Non-participants coded 'Other' reason 6 \n", "11 Academic Year fully determined by Field of Stu... 1 \n", "12 Rows where the five global perception items ar... 500 \n", "\n", " Note \n", "0 500 expected (89.3% of 560 invited) \n", "1 \n", "2 \n", "3 \n", "4 observed 20-23 \n", "5 observed 50-90 \n", "6 observed 1-5 \n", "7 observed 0-1 \n", "8 observed 0-1 \n", "9 0 = consistent skip logic \n", "10 'Other' also used as not-applicable code for p... \n", "11 1 = yes; Academic Year cannot enter a model wi... \n", "12 500 = items are response-for-response duplicates " ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
Field of StudyArtsCommerceEngineeringITScience
Academic Year
3rd Year11612904539
4th Year0017100
\n", "
" ], "text/plain": [ "Field of Study Arts Commerce Engineering IT Science\n", "Academic Year \n", "3rd Year 116 129 0 45 39\n", "4th Year 0 0 171 0 0" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
minmaxmeanstdcount
Academic Performance
1506457.274.2252
2556962.974.5176
3657972.314.37160
4759081.934.16145
5759082.634.5967
\n", "
" ], "text/plain": [ " min max mean std count\n", "Academic Performance \n", "1 50 64 57.27 4.22 52\n", "2 55 69 62.97 4.51 76\n", "3 65 79 72.31 4.37 160\n", "4 75 90 81.93 4.16 145\n", "5 75 90 82.63 4.59 67" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "raw = pd.read_excel(\"Supplementary_Material_1_revised.xlsx\", sheet_name=\"Data\")\n", "raw.columns = [c.strip() for c in raw.columns]\n", "df = raw.copy()\n", "if \"Respondent_ID\" not in df.columns:\n", " df.insert(0, \"Respondent_ID\", [f\"R{i:03d}\" for i in range(1, len(df)+1)])\n", "\n", "audit = []\n", "audit.append((\"Records\", len(df), \"500 expected (89.3% of 560 invited)\"))\n", "audit.append((\"Variables\", df.shape[1]-1, \"\"))\n", "audit.append((\"Missing values (all cells)\", int(df.isna().sum().sum()), \"\"))\n", "audit.append((\"Exact duplicate records\", int(raw.duplicated().sum()), \"\"))\n", "\n", "# range checks\n", "chk = {\"Age\": (18, 30), \"Attendance (%)\": (0, 100), \"Academic Performance\": (1, 5),\n", " \"Participated in Engagement\": (0, 1), \"Secured Job\": (0, 1)}\n", "for v, (lo, hi) in chk.items():\n", " bad = int(((df[v] < lo) | (df[v] > hi)).sum())\n", " audit.append((f\"Out-of-range values: {v} [{lo}-{hi}]\", bad, f\"observed {df[v].min()}-{df[v].max()}\"))\n", "\n", "# logical consistency: non-participation reason should be substantive only for non-participants\n", "subst = df[\"Non-Participation Reason\"] != \"Other\"\n", "audit.append((\"Participants reporting a substantive non-participation reason\", int((subst & (df[\"Participated in Engagement\"]==1)).sum()), \"0 = consistent skip logic\"))\n", "audit.append((\"Non-participants coded 'Other' reason\", int(((~subst) & (df[\"Participated in Engagement\"]==0)).sum()), \"'Other' also used as not-applicable code for participants\"))\n", "\n", "# structural aliasing: Academic Year vs Field of Study\n", "ct_yf = pd.crosstab(df[\"Academic Year\"], df[\"Field of Study\"])\n", "aliased = int(((ct_yf > 0).sum(axis=0) == 1).all())\n", "audit.append((\"Academic Year fully determined by Field of Study (Engineering = 4th year)\", aliased, \"1 = yes; Academic Year cannot enter a model with Field\"))\n", "\n", "# redundancy of global perception items\n", "m5 = {\"Excellent\":5,\"Good\":4,\"Average\":3,\"Fair\":2,\"Poor\":1}\n", "ms = {\"Completely\":5,\"Very much\":4,\"Moderately\":3,\"Slightly\":2,\"Not at all\":1}\n", "me = {\"Extremely effective\":5,\"Very effective\":4,\"Moderately effective\":3,\"Somewhat effective\":2,\"Ineffective\":1}\n", "glob = pd.DataFrame({\"Availability\": df[\"Career Guidance Availability\"].map(m5),\n", " \"Support\": df[\"Career Guidance Support\"].map(ms),\n", " \"Effectiveness\": df[\"Career Guidance Effectiveness\"].map(me),\n", " \"Impact\": df[\"Career Guidance Impact\"],\n", " \"Placement Effectiveness\": df[\"Placement Effectiveness\"]})\n", "ident = int((glob.nunique(axis=1) == 1).sum())\n", "audit.append((\"Rows where the five global perception items are identical\", ident, \"500 = items are response-for-response duplicates\"))\n", "\n", "# attendance ranges by CGPA band (overlap check)\n", "rng_tab = df.groupby(\"Academic Performance\")[\"Attendance (%)\"].agg([\"min\",\"max\",\"mean\",\"std\",\"count\"]).round(2)\n", "audit_df = pd.DataFrame(audit, columns=[\"Check\", \"Result\", \"Note\"])\n", "R[\"audit\"] = audit_df.to_dict(\"records\")\n", "display(audit_df); display(ct_yf); display(rng_tab)\n" ] }, { "cell_type": "code", "execution_count": 3, "id": "12f97ef9", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:11:45.159098Z", "iopub.status.busy": "2026-09-27T19:11:45.158413Z", "iopub.status.idle": "2026-09-27T19:11:45.183226Z", "shell.execute_reply": "2026-09-27T19:11:45.182447Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " count mean std min 25% 50% 75% max\n", "placed 500.0 0.900 0.300300 0.0 1.0 1.0 1.0 1.0\n", "part 500.0 0.886 0.318130 0.0 1.0 1.0 1.0 1.0\n", "cgpa 500.0 3.198 1.165123 1.0 2.0 3.0 4.0 5.0\n", "att 500.0 73.498 9.766307 50.0 67.0 75.0 81.0 90.0\n" ] } ], "source": [ "# Analytical variables\n", "df[\"placed\"] = df[\"Secured Job\"].astype(int)\n", "df[\"part\"] = df[\"Participated in Engagement\"].astype(int)\n", "df[\"cgpa\"] = df[\"Academic Performance\"].astype(int)\n", "df[\"att\"] = df[\"Attendance (%)\"].astype(float)\n", "df[\"att10\"] = df[\"att\"] / 10 # OR per 10 percentage points\n", "df[\"male\"] = (df[\"Gender\"] == \"Male\").astype(int)\n", "df[\"field\"] = pd.Categorical(df[\"Field of Study\"], [\"Arts\",\"Commerce\",\"Engineering\",\"IT\",\"Science\"])\n", "CG = {1:\"5–6\",2:\"6–7\",3:\"7–8\",4:\"8–9\",5:\"9–10\"}\n", "df[\"cgpa_lab\"] = df[\"cgpa\"].map(CG)\n", "df.to_csv(f\"{OUT}/analytic_dataset.csv\", index=False)\n", "print(df[[\"placed\",\"part\",\"cgpa\",\"att\"]].describe().T)\n" ] }, { "cell_type": "markdown", "id": "9d43ecb7", "metadata": {}, "source": [ "## Analysis 2 — Descriptive statistics (RQ1)" ] }, { "cell_type": "code", "execution_count": 4, "id": "dbb9d095", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:11:45.185771Z", "iopub.status.busy": "2026-09-27T19:11:45.185121Z", "iopub.status.idle": "2026-09-27T19:11:45.237470Z", "shell.execute_reply": "2026-09-27T19:11:45.236601Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Gender\n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
n%
Gender
Male27354.6
Female22745.4
\n", "
" ], "text/plain": [ " n %\n", "Gender \n", "Male 273 54.6\n", "Female 227 45.4" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Field of Study\n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
n%
Field of Study
Engineering17134.2
Commerce12925.8
Arts11623.2
IT459.0
Science397.8
\n", "
" ], "text/plain": [ " n %\n", "Field of Study \n", "Engineering 171 34.2\n", "Commerce 129 25.8\n", "Arts 116 23.2\n", "IT 45 9.0\n", "Science 39 7.8" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Academic Year\n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
n%
Academic Year
4th Year17134.2
3rd Year32965.8
\n", "
" ], "text/plain": [ " n %\n", "Academic Year \n", "4th Year 171 34.2\n", "3rd Year 329 65.8" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "cgpa_lab\n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
n%
cgpa_lab
5–65210.4
6–77615.2
7–816032.0
8–914529.0
9–106713.4
\n", "
" ], "text/plain": [ " n %\n", "cgpa_lab \n", "5–6 52 10.4\n", "6–7 76 15.2\n", "7–8 160 32.0\n", "8–9 145 29.0\n", "9–10 67 13.4" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Participated in Engagement\n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
n%
Participated in Engagement
144388.6
05711.4
\n", "
" ], "text/plain": [ " n %\n", "Participated in Engagement \n", "1 443 88.6\n", "0 57 11.4" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Secured Job\n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
n%
Secured Job
145090.0
05010.0
\n", "
" ], "text/plain": [ " n %\n", "Secured Job \n", "1 450 90.0\n", "0 50 10.0" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
n% of non-participants
Non-Participation Reason
Lack of awareness2442.1
Time constraints1729.8
Lack of interest1017.5
Other610.5
\n", "
" ], "text/plain": [ " n % of non-participants\n", "Non-Participation Reason \n", "Lack of awareness 24 42.1\n", "Time constraints 17 29.8\n", "Lack of interest 10 17.5\n", "Other 6 10.5" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "{'age': {'mean': np.float64(21.23), 'sd': np.float64(0.8663724386225647), 'min': 20, 'max': 23}, 'att': {'mean': np.float64(73.498), 'sd': np.float64(9.766307260936962), 'median': np.float64(75.0), 'q1': np.float64(67.0), 'q3': np.float64(81.0), 'min': np.float64(50.0), 'max': np.float64(90.0), 'skew': np.float64(-0.3793841026613706), 'kurt': np.float64(-0.7026914739746242)}, 'n_part': 443, 'n_placed': 450, 'placed_rate': 0.9, 'placed_ci': (0.8705780903680599, 0.9233224372626984), 'part_rate': 0.886, 'part_ci': (0.8551509101576845, 0.9109630990059973)}\n" ] } ], "source": [ "def freq(v, order=None):\n", " t = df[v].value_counts()\n", " if order: t = t.reindex(order)\n", " return pd.DataFrame({\"n\": t, \"%\": (100*t/len(df)).round(1)})\n", "\n", "desc = {}\n", "for v, o in [(\"Gender\",[\"Male\",\"Female\"]),(\"Field of Study\",[\"Engineering\",\"Commerce\",\"Arts\",\"IT\",\"Science\"]),\n", " (\"Academic Year\",[\"4th Year\",\"3rd Year\"]),(\"cgpa_lab\",list(CG.values())),\n", " (\"Participated in Engagement\",[1,0]),(\"Secured Job\",[1,0])]:\n", " desc[v] = freq(v, o); print(v); display(desc[v])\n", "\n", "R[\"age\"] = dict(mean=df[\"Age\"].mean(), sd=df[\"Age\"].std(), min=int(df[\"Age\"].min()), max=int(df[\"Age\"].max()))\n", "R[\"att\"] = dict(mean=df[\"att\"].mean(), sd=df[\"att\"].std(), median=df[\"att\"].median(),\n", " q1=df[\"att\"].quantile(.25), q3=df[\"att\"].quantile(.75), min=df[\"att\"].min(), max=df[\"att\"].max(),\n", " skew=st.skew(df[\"att\"]), kurt=st.kurtosis(df[\"att\"]))\n", "R[\"cgpa_med\"] = float(df[\"cgpa\"].median())\n", "R[\"n_part\"] = int(df[\"part\"].sum()); R[\"n_nonpart\"] = int((1-df[\"part\"]).sum())\n", "R[\"n_placed\"] = int(df[\"placed\"].sum()); R[\"n_notplaced\"] = int((1-df[\"placed\"]).sum())\n", "lo, hi = proportion_confint(R[\"n_placed\"], len(df), method=\"wilson\")\n", "R[\"placed_rate\"] = R[\"n_placed\"]/len(df); R[\"placed_ci\"] = (lo, hi)\n", "lo, hi = proportion_confint(R[\"n_part\"], len(df), method=\"wilson\")\n", "R[\"part_rate\"] = R[\"n_part\"]/len(df); R[\"part_ci\"] = (lo, hi)\n", "\n", "np_reason = df.loc[df.part==0, \"Non-Participation Reason\"].value_counts()\n", "np_reason = pd.DataFrame({\"n\": np_reason, \"% of non-participants\": (100*np_reason/np_reason.sum()).round(1)})\n", "R[\"np_reason\"] = np_reason.reset_index().to_dict(\"records\")\n", "display(np_reason)\n", "print({k: R[k] for k in [\"age\",\"att\",\"n_part\",\"n_placed\",\"placed_rate\",\"placed_ci\",\"part_rate\",\"part_ci\"]})\n" ] }, { "cell_type": "code", "execution_count": 5, "id": "c850a609", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:11:45.239659Z", "iopub.status.busy": "2026-09-27T19:11:45.238970Z", "iopub.status.idle": "2026-09-27T19:11:45.318815Z", "shell.execute_reply": "2026-09-27T19:11:45.317703Z" } }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
variablelevelnplacedratechi2dfpfisher_pVmin_exp
0GenderFemale2272050.9030.04410.8340.8820.00922.7
1GenderMale2732450.8970.04410.8340.8820.00922.7
2Field of StudyArts1161030.8887.46740.113NaN0.1223.9
3Field of StudyCommerce1291200.9307.46740.113NaN0.1223.9
4Field of StudyEngineering1711570.9187.46740.113NaN0.1223.9
5Field of StudyIT45390.8677.46740.113NaN0.1223.9
6Field of StudyScience39310.7957.46740.113NaN0.1223.9
7Academic Year3rd Year3292930.8910.94910.3300.3510.04417.1
8Academic Year4th Year1711570.9180.94910.3300.3510.04417.1
\n", "
" ], "text/plain": [ " variable level n placed rate chi2 df p \\\n", "0 Gender Female 227 205 0.903 0.044 1 0.834 \n", "1 Gender Male 273 245 0.897 0.044 1 0.834 \n", "2 Field of Study Arts 116 103 0.888 7.467 4 0.113 \n", "3 Field of Study Commerce 129 120 0.930 7.467 4 0.113 \n", "4 Field of Study Engineering 171 157 0.918 7.467 4 0.113 \n", "5 Field of Study IT 45 39 0.867 7.467 4 0.113 \n", "6 Field of Study Science 39 31 0.795 7.467 4 0.113 \n", "7 Academic Year 3rd Year 329 293 0.891 0.949 1 0.330 \n", "8 Academic Year 4th Year 171 157 0.918 0.949 1 0.330 \n", "\n", " fisher_p V min_exp \n", "0 0.882 0.009 22.7 \n", "1 0.882 0.009 22.7 \n", "2 NaN 0.122 3.9 \n", "3 NaN 0.122 3.9 \n", "4 NaN 0.122 3.9 \n", "5 NaN 0.122 3.9 \n", "6 NaN 0.122 3.9 \n", "7 0.351 0.044 17.1 \n", "8 0.351 0.044 17.1 " ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "{'att': {'m1': np.float64(74.79683972911964), 'sd1': np.float64(9.238630725513381), 'm0': np.float64(63.40350877192982), 'sd0': np.float64(7.692435173418306), 'med1': np.float64(76.0), 'med0': np.float64(62.0), 't': np.float64(10.26967529755028), 'p_t': np.float64(3.635327780880866e-16), 'U': np.float64(20750.5), 'p_U': np.float64(2.4183554286296056e-15), 'r_rb': np.float64(0.6435388697477327)}, 'cgpa': {'m1': np.float64(3.36117381489842), 'sd1': np.float64(1.103482957340458), 'm0': np.float64(1.9298245614035088), 'sd0': np.float64(0.7986517460876311), 'med1': np.float64(3.0), 'med0': np.float64(2.0), 't': np.float64(12.123568427015824), 'p_t': np.float64(2.5155094766158667e-20), 'U': np.float64(21137.5), 'p_U': np.float64(1.0410198248838405e-17), 'r_rb': np.float64(0.674191121143717)}}\n" ] } ], "source": [ "def cramers_v(x, y):\n", " t = pd.crosstab(x, y).values\n", " c2 = st.chi2_contingency(t, correction=False)[0]\n", " return np.sqrt(c2 / (t.sum() * (min(t.shape) - 1)))\n", "\n", "# Placement by demographic characteristics, and academic profile by participation (context for confounding)\n", "demo = []\n", "for v in [\"Gender\", \"Field of Study\", \"Academic Year\"]:\n", " t = pd.crosstab(df[v], df.placed)\n", " c2, p_, dof_, e_ = st.chi2_contingency(t.values, correction=False)\n", " p_exact = st.fisher_exact(t.values)[1] if t.shape == (2, 2) else np.nan\n", " for lvl in t.index:\n", " demo.append(dict(variable=v, level=lvl, n=int(t.loc[lvl].sum()), placed=int(t.loc[lvl, 1]),\n", " rate=t.loc[lvl, 1]/t.loc[lvl].sum(), chi2=c2, df=dof_, p=p_, fisher_p=p_exact,\n", " V=cramers_v(df[v], df.placed), min_exp=e_.min()))\n", "demo = pd.DataFrame(demo); R[\"demo_placement\"] = demo.to_dict(\"records\"); display(demo.round(3))\n", "\n", "prof = {}\n", "for v in [\"att\", \"cgpa\"]:\n", " x1 = df.loc[df.part==1, v]; x0 = df.loc[df.part==0, v]\n", " U_, pU_ = st.mannwhitneyu(x1, x0, alternative=\"two-sided\")\n", " tw, pw = st.ttest_ind(x1, x0, equal_var=False)\n", " prof[v] = dict(m1=x1.mean(), sd1=x1.std(), m0=x0.mean(), sd0=x0.std(), med1=x1.median(), med0=x0.median(),\n", " t=tw, p_t=pw, U=U_, p_U=pU_, r_rb=2*U_/(len(x1)*len(x0))-1)\n", "R[\"profile_by_part\"] = prof; print(prof)\n" ] }, { "cell_type": "markdown", "id": "a069297f", "metadata": {}, "source": [ "## Analysis 3 — Participation × placement cross-tabulation (RQ2 / H1)" ] }, { "cell_type": "code", "execution_count": 6, "id": "aea82646", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:11:45.321015Z", "iopub.status.busy": "2026-09-27T19:11:45.320393Z", "iopub.status.idle": "2026-09-27T19:11:46.698494Z", "shell.execute_reply": "2026-09-27T19:11:46.697327Z" } }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
groupnplacednot_placedratelohi
0Participated443426170.9616250.9394090.975905
1Did not participate5724330.4210530.3018760.550199
\n", "
" ], "text/plain": [ " group n placed not_placed rate lo hi\n", "0 Participated 443 426 17 0.961625 0.939409 0.975905\n", "1 Did not participate 57 24 33 0.421053 0.301876 0.550199" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "{'rd': 0.5405726505880956, 'rd_lo': np.float64(0.41115445797937145), 'rd_hi': np.float64(0.6699908431968197), 'rr': 2.283860045146727, 'rr_lo': np.float64(1.683510820480279), 'rr_hi': np.float64(3.0982971076654935), 'or_': 34.455882352941174, 'or_lo': np.float64(16.852622095734194), 'or_hi': np.float64(70.44647544907767)}\n" ] }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "ct = pd.crosstab(df[\"part\"], df[\"placed\"]).reindex(index=[1,0], columns=[1,0])\n", "ct.index = [\"Participated\", \"Did not participate\"]; ct.columns = [\"Placed\", \"Not placed\"]\n", "rates = []\n", "for g, row in ct.iterrows():\n", " n = row.sum(); k = row[\"Placed\"]; lo, hi = proportion_confint(k, n, method=\"wilson\")\n", " rates.append(dict(group=g, n=int(n), placed=int(k), not_placed=int(n-k), rate=k/n, lo=lo, hi=hi))\n", "rates = pd.DataFrame(rates); R[\"h1_rates\"] = rates.to_dict(\"records\"); display(rates)\n", "\n", "a, b = ct.loc[\"Participated\"]; c, d_ = ct.loc[\"Did not participate\"]\n", "p1, p0 = a/(a+b), c/(c+d_)\n", "rd = p1 - p0; se_rd = np.sqrt(p1*(1-p1)/(a+b) + p0*(1-p0)/(c+d_))\n", "rr = p1/p0; se_lrr = np.sqrt(b/(a*(a+b)) + d_/(c*(c+d_)))\n", "orr = (a*d_)/(b*c); se_lor = np.sqrt(1/a+1/b+1/c+1/d_)\n", "R[\"h1_effects\"] = dict(rd=rd, rd_lo=rd-1.96*se_rd, rd_hi=rd+1.96*se_rd,\n", " rr=rr, rr_lo=np.exp(np.log(rr)-1.96*se_lrr), rr_hi=np.exp(np.log(rr)+1.96*se_lrr),\n", " or_=orr, or_lo=np.exp(np.log(orr)-1.96*se_lor), or_hi=np.exp(np.log(orr)+1.96*se_lor))\n", "print(R[\"h1_effects\"])\n", "\n", "fig, ax = plt.subplots(figsize=(8, 6))\n", "cols = [\"#3a3a3a\", \"#a6a6a6\"]\n", "for i, r in rates.iterrows():\n", " ax.bar(i, 100*r.rate, width=0.55, color=cols[i], edgecolor=\"black\", linewidth=1.2)\n", " ax.errorbar(i, 100*r.rate, yerr=[[100*(r.rate-r.lo)], [100*(r.hi-r.rate)]], color=\"black\", capsize=10, lw=1.8, capthick=1.8)\n", " ax.text(i, 100*r.hi + 2.5, f\"{100*r.rate:.1f}%\\n({r.placed}/{r.n})\", ha=\"center\", va=\"bottom\", fontsize=17)\n", "ax.set_xticks([0, 1]); ax.set_xticklabels([\"Participated\", \"Did not participate\"])\n", "ax.set_ylabel(\"Students placed (%)\"); ax.set_ylim(0, 118); ax.set_yticks(range(0, 101, 20))\n", "ax.set_xlabel(\"Participation in institutional engagement activities\")\n", "savefig(fig, \"Fig1_placement_by_participation\")\n" ] }, { "cell_type": "markdown", "id": "1cd547fc", "metadata": {}, "source": [ "## Analysis 4 — Chi-square test of independence (H1)" ] }, { "cell_type": "code", "execution_count": 7, "id": "ad2bc3b5", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:11:46.700886Z", "iopub.status.busy": "2026-09-27T19:11:46.700021Z", "iopub.status.idle": "2026-09-27T19:11:46.708158Z", "shell.execute_reply": "2026-09-27T19:11:46.707369Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "{'chi2': np.float64(163.97370401172233), 'df': 1, 'p': np.float64(1.5328730811569313e-37), 'chi2_yates': np.float64(158.0223445496108), 'p_yates': np.float64(3.060261602676487e-36), 'fisher_p': np.float64(6.714669667657038e-24), 'min_expected': 5.7, 'cells_lt5': 0}\n", "Expected counts:\n", " [[398.7 44.3]\n", " [ 51.3 5.7]]\n" ] } ], "source": [ "obs = ct.values\n", "chi2, p, dof, exp = st.chi2_contingency(obs, correction=False)\n", "chi2_y, p_y, _, _ = st.chi2_contingency(obs, correction=True)\n", "fisher_or, fisher_p = st.fisher_exact(obs)\n", "R[\"h1_chi\"] = dict(chi2=chi2, df=dof, p=p, chi2_yates=chi2_y, p_yates=p_y, fisher_p=fisher_p,\n", " min_expected=float(exp.min()), cells_lt5=int((exp < 5).sum()))\n", "print(R[\"h1_chi\"]); print(\"Expected counts:\\n\", exp.round(2))\n" ] }, { "cell_type": "markdown", "id": "84f43bcf", "metadata": {}, "source": [ "## Analysis 5 — Effect size (Cramér's V / φ) with 95% bootstrap CI (H1)" ] }, { "cell_type": "code", "execution_count": 8, "id": "647279b0", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:11:46.710377Z", "iopub.status.busy": "2026-09-27T19:11:46.709659Z", "iopub.status.idle": "2026-09-27T19:12:05.206679Z", "shell.execute_reply": "2026-09-27T19:12:05.205192Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "{'V': np.float64(0.5726669259032205), 'lo': np.float64(0.45192214701544303), 'hi': np.float64(0.6873479330592546)}\n" ] } ], "source": [ "def cramers_v(x, y):\n", " t = pd.crosstab(x, y).values\n", " if min(t.shape) < 2: return np.nan\n", " c2 = st.chi2_contingency(t, correction=False)[0]\n", " return np.sqrt(c2 / (t.sum() * (min(t.shape) - 1)))\n", "\n", "def boot_v(x, y, B=5000):\n", " x = np.asarray(x); y = np.asarray(y); n = len(x); out = np.empty(B)\n", " for b in range(B):\n", " i = rng.integers(0, n, n); out[b] = cramers_v(x[i], y[i])\n", " return np.nanpercentile(out, [2.5, 97.5])\n", "\n", "V = cramers_v(df.part, df.placed); ci = boot_v(df.part, df.placed)\n", "R[\"h1_V\"] = dict(V=V, lo=ci[0], hi=ci[1])\n", "print(R[\"h1_V\"]) # Cohen (1988) df*=1: .10 small, .30 medium, .50 large\n" ] }, { "cell_type": "markdown", "id": "1132269f", "metadata": {}, "source": [ "## Analysis 6 — Academic performance (CGPA band) and placement (H2)" ] }, { "cell_type": "code", "execution_count": 9, "id": "fe9f10c1", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:12:05.209713Z", "iopub.status.busy": "2026-09-27T19:12:05.208934Z", "iopub.status.idle": "2026-09-27T19:12:25.696988Z", "shell.execute_reply": "2026-09-27T19:12:25.696039Z" } }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
PlacedNot placednratelohi
5–62329520.4423080.3159510.576602
6–75917760.7763160.6706540.855388
7–815731600.9812500.9463300.993603
8–914411450.9931030.9619720.998782
9–10670671.0000000.9457741.000000
\n", "
" ], "text/plain": [ " Placed Not placed n rate lo hi\n", "5–6 23 29 52 0.442308 0.315951 0.576602\n", "6–7 59 17 76 0.776316 0.670654 0.855388\n", "7–8 157 3 160 0.981250 0.946330 0.993603\n", "8–9 144 1 145 0.993103 0.961972 0.998782\n", "9–10 67 0 67 1.000000 0.945774 1.000000" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "{'chi2': np.float64(167.09838948609365), 'df': 4, 'p': np.float64(4.386874712698728e-35), 'min_expected': 5.2, 'V': np.float64(0.578097551432444), 'V_lo': np.float64(0.4838169144614329), 'V_hi': np.float64(0.6773298464063722), 'z_trend': np.float64(10.745305095149556), 'p_trend': np.float64(6.235542138601229e-27), 'U': np.float64(20481.5), 'p_U': np.float64(7.169751314411539e-23), 'r_rb': np.float64(0.8205777777777779), 'med_placed': 3.0, 'med_not': 1.0, 'mean_placed': np.float64(3.3844444444444446), 'mean_not': np.float64(1.52), 'rho': np.float64(0.44076800434711255), 'p_rho': np.float64(3.5131526682440223e-25)}\n" ] }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "ct2 = pd.crosstab(df[\"cgpa\"], df[\"placed\"]).reindex(columns=[1,0]); ct2.columns = [\"Placed\",\"Not placed\"]\n", "ct2[\"n\"] = ct2.sum(1); ct2[\"rate\"] = ct2[\"Placed\"]/ct2[\"n\"]\n", "ci2 = [proportion_confint(k, n, method=\"wilson\") for k, n in zip(ct2[\"Placed\"], ct2[\"n\"])]\n", "ct2[\"lo\"] = [c[0] for c in ci2]; ct2[\"hi\"] = [c[1] for c in ci2]; ct2.index = [CG[i] for i in ct2.index]\n", "display(ct2)\n", "chi2b, pb, dofb, expb = st.chi2_contingency(pd.crosstab(df.cgpa, df.placed).values, correction=False)\n", "Vb = cramers_v(df.cgpa, df.placed); cib = boot_v(df.cgpa, df.placed)\n", "\n", "# Cochran–Armitage test for linear trend\n", "def cochran_armitage(k, n, s):\n", " k, n, s = map(np.asarray, (k, n, s)); N = n.sum(); pbar = k.sum()/N\n", " T = np.sum(s*(k - n*pbar)); var = pbar*(1-pbar)*(np.sum(n*s**2) - np.sum(n*s)**2/N)\n", " z = T/np.sqrt(var); return z, 2*st.norm.sf(abs(z))\n", "z_ca, p_ca = cochran_armitage(ct2[\"Placed\"], ct2[\"n\"], np.arange(1, 6))\n", "\n", "# Mann–Whitney U with rank-biserial correlation (placed vs not placed)\n", "g1 = df.loc[df.placed==1, \"cgpa\"]; g0 = df.loc[df.placed==0, \"cgpa\"]\n", "U, p_u = st.mannwhitneyu(g1, g0, alternative=\"two-sided\")\n", "rrb = 2*U/(len(g1)*len(g0)) - 1\n", "rho, p_rho = st.spearmanr(df.cgpa, df.placed)\n", "R[\"h2\"] = dict(chi2=chi2b, df=dofb, p=pb, min_expected=float(expb.min()), V=Vb, V_lo=cib[0], V_hi=cib[1],\n", " z_trend=z_ca, p_trend=p_ca, U=U, p_U=p_u, r_rb=rrb,\n", " med_placed=float(g1.median()), med_not=float(g0.median()),\n", " mean_placed=g1.mean(), mean_not=g0.mean(), rho=rho, p_rho=p_rho,\n", " rates=ct2.reset_index().rename(columns={\"index\":\"band\"}).to_dict(\"records\"))\n", "print({k: v for k, v in R[\"h2\"].items() if k != \"rates\"})\n", "\n", "fig, ax = plt.subplots(figsize=(9, 6))\n", "x = np.arange(5)\n", "ax.bar(x, 100*ct2[\"rate\"], width=0.6, color=\"#7f7f7f\", edgecolor=\"black\", linewidth=1.2)\n", "ax.errorbar(x, 100*ct2[\"rate\"], yerr=[np.clip(100*(ct2[\"rate\"]-ct2[\"lo\"]),0,None), np.clip(100*(ct2[\"hi\"]-ct2[\"rate\"]),0,None)],\n", " fmt=\"none\", color=\"black\", capsize=8, lw=1.8, capthick=1.8)\n", "for i, (r, h, k, n) in enumerate(zip(ct2[\"rate\"], ct2[\"hi\"], ct2[\"Placed\"], ct2[\"n\"])):\n", " ax.text(i, 100*h + 2.5, f\"{100*r:.1f}%\\n({k}/{n})\", ha=\"center\", va=\"bottom\", fontsize=15)\n", "ax.set_xticks(x); ax.set_xticklabels(ct2.index); ax.set_xlabel(\"CGPA band\")\n", "ax.set_ylabel(\"Students placed (%)\"); ax.set_ylim(0, 122); ax.set_yticks(range(0, 101, 20))\n", "savefig(fig, \"Fig2_placement_by_cgpa\")\n" ] }, { "cell_type": "markdown", "id": "581e54dd", "metadata": {}, "source": [ "## Analysis 7 — Attendance and placement (H3)" ] }, { "cell_type": "code", "execution_count": 10, "id": "adb9fd05", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:12:25.699241Z", "iopub.status.busy": "2026-09-27T19:12:25.698618Z", "iopub.status.idle": "2026-09-27T19:12:26.977729Z", "shell.execute_reply": "2026-09-27T19:12:26.976547Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "{'m1': np.float64(75.03111111111112), 'sd1': np.float64(8.849592549658798), 'm0': np.float64(59.7), 'sd0': np.float64(6.151986571466993), 'med1': np.float64(76.0), 'med0': np.float64(59.0), 'diff': np.float64(15.331111111111113), 'diff_lo': np.float64(13.40843576956672), 'diff_hi': np.float64(17.253786452655508), 't': np.float64(15.889313969941748), 'df': np.float64(73.69708156992728), 'p': np.float64(1.6477827485369927e-25), 'g': np.float64(1.7755240895977322), 'g_lo': np.float64(1.4633070524099996), 'g_hi': np.float64(2.0877411267854646), 'levene': np.float64(9.217160610517116), 'p_levene': np.float64(0.0025231215725024358), 'shapiro_p_placed': np.float64(8.095809293669744e-08), 'shapiro_p_not': np.float64(0.018968824839386656), 'U': np.float64(20437.0), 'p_U': np.float64(2.4410247880891892e-21), 'r_rb': np.float64(0.8166222222222221), 'r_pb': np.float64(0.4714104783084066), 'p_rpb': np.float64(4.999948814396894e-29)}\n" ] }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "a1 = df.loc[df.placed==1, \"att\"]; a0 = df.loc[df.placed==0, \"att\"]\n", "t_w, p_w = st.ttest_ind(a1, a0, equal_var=False)\n", "n1, n0 = len(a1), len(a0); s1, s0 = a1.var(ddof=1), a0.var(ddof=1)\n", "df_w = (s1/n1 + s0/n0)**2 / ((s1/n1)**2/(n1-1) + (s0/n0)**2/(n0-1))\n", "diff = a1.mean() - a0.mean(); se_d = np.sqrt(s1/n1 + s0/n0); tcrit = st.t.ppf(.975, df_w)\n", "sp = np.sqrt(((n1-1)*s1 + (n0-1)*s0)/(n1+n0-2)); d_c = diff/sp\n", "J = 1 - 3/(4*(n1+n0)-9); g = J*d_c\n", "se_g = np.sqrt((n1+n0)/(n1*n0) + g**2/(2*(n1+n0)))\n", "lev, p_lev = st.levene(a1, a0, center=\"median\")\n", "sw1, sw0 = st.shapiro(a1).pvalue, st.shapiro(a0).pvalue\n", "U3, pU3 = st.mannwhitneyu(a1, a0, alternative=\"two-sided\"); rrb3 = 2*U3/(n1*n0) - 1\n", "rpb, p_rpb = st.pointbiserialr(df.placed, df.att)\n", "R[\"h3\"] = dict(m1=a1.mean(), sd1=a1.std(), m0=a0.mean(), sd0=a0.std(), med1=a1.median(), med0=a0.median(),\n", " diff=diff, diff_lo=diff-tcrit*se_d, diff_hi=diff+tcrit*se_d, t=t_w, df=df_w, p=p_w,\n", " g=g, g_lo=g-1.96*se_g, g_hi=g+1.96*se_g, levene=lev, p_levene=p_lev,\n", " shapiro_p_placed=sw1, shapiro_p_not=sw0, U=U3, p_U=pU3, r_rb=rrb3, r_pb=rpb, p_rpb=p_rpb)\n", "print(R[\"h3\"])\n", "\n", "fig, ax = plt.subplots(figsize=(8, 6.5))\n", "data = [a1.values, a0.values]\n", "bp = ax.boxplot(data, positions=[0, 1], widths=0.5, patch_artist=True, showfliers=False,\n", " medianprops=dict(color=\"black\", lw=2.2), whiskerprops=dict(lw=1.6), capprops=dict(lw=1.6),\n", " boxprops=dict(lw=1.6))\n", "for patch, c in zip(bp[\"boxes\"], [\"#bfbfbf\", \"#ffffff\"]): patch.set_facecolor(c)\n", "for i, dat in enumerate(data):\n", " jit = rng.uniform(-0.17, 0.17, len(dat))\n", " ax.scatter(np.full(len(dat), i) + jit, dat, s=14, color=\"black\", alpha=0.35, zorder=3)\n", " ax.scatter(i, dat.mean(), marker=\"D\", s=110, color=\"white\", edgecolor=\"black\", lw=1.8, zorder=4)\n", "ax.set_xticks([0, 1]); ax.set_xticklabels([f\"Placed\\n(n = {n1})\", f\"Not placed\\n(n = {n0})\"])\n", "ax.set_ylabel(\"Attendance (%)\"); ax.set_ylim(45, 95)\n", "savefig(fig, \"Fig3_attendance_by_placement\")\n" ] }, { "cell_type": "markdown", "id": "c8b5fe9c", "metadata": {}, "source": [ "## Analysis 8 — Correlation and multicollinearity diagnostics (H2/H3)" ] }, { "cell_type": "code", "execution_count": 11, "id": "1f704ff9", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:12:26.980364Z", "iopub.status.busy": "2026-09-27T19:12:26.979535Z", "iopub.status.idle": "2026-09-27T19:12:28.637962Z", "shell.execute_reply": "2026-09-27T19:12:28.637007Z" } }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
attcgpapartplaced
att1.0000.8690.3710.471
cgpa0.8691.0000.3910.481
part0.3710.3911.0000.573
placed0.4710.4810.5731.000
\n", "
" ], "text/plain": [ " att cgpa part placed\n", "att 1.000 0.869 0.371 0.471\n", "cgpa 0.869 1.000 0.391 0.481\n", "part 0.371 0.391 1.000 0.573\n", "placed 0.471 0.481 0.573 1.000" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
attcgpapartplaced
att1.0000.8700.3540.425
cgpa0.8701.0000.3840.441
part0.3540.3841.0000.573
placed0.4250.4410.5731.000
\n", "
" ], "text/plain": [ " att cgpa part placed\n", "att 1.000 0.870 0.354 0.425\n", "cgpa 0.870 1.000 0.384 0.441\n", "part 0.354 0.384 1.000 0.573\n", "placed 0.425 0.441 0.573 1.000" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
PredictorVIFTolerance
0Participation1.1860.843
1CGPA band4.1740.240
2Attendance (per 10 pts)4.1010.244
\n", "
" ], "text/plain": [ " Predictor VIF Tolerance\n", "0 Participation 1.186 0.843\n", "1 CGPA band 4.174 0.240\n", "2 Attendance (per 10 pts) 4.101 0.244" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stderr", "output_type": "stream", "text": [ "findfont: Font family ['cursive'] not found. Falling back to DejaVu Sans.\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "findfont: Generic family 'cursive' not found because none of the following families were found: Apple Chancery, Textile, Zapf Chancery, Sand, Script MT, Felipa, Comic Neue, Comic Sans MS, cursive\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "findfont: Font family ['Times New Roman'] not found. Falling back to DejaVu Sans.\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "findfont: Font family ['Times New Roman'] not found. Falling back to DejaVu Sans.\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Condition index: 4.03\n" ] }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "vars_ = [\"att\", \"cgpa\", \"part\", \"placed\"]\n", "lab = {\"att\":\"Attendance (%)\", \"cgpa\":\"CGPA band\", \"part\":\"Participation\", \"placed\":\"Placed\"}\n", "pear = df[vars_].corr(); spear = df[vars_].corr(\"spearman\")\n", "pp = pd.DataFrame(index=vars_, columns=vars_, dtype=float)\n", "for i in vars_:\n", " for j in vars_: pp.loc[i, j] = st.pearsonr(df[i], df[j])[1]\n", "r_ac, p_ac = st.pearsonr(df.att, df.cgpa)\n", "z = np.arctanh(r_ac); se = 1/np.sqrt(len(df)-3)\n", "R[\"corr_att_cgpa\"] = dict(r=r_ac, lo=np.tanh(z-1.96*se), hi=np.tanh(z+1.96*se), p=p_ac,\n", " rho=st.spearmanr(df.att, df.cgpa)[0])\n", "R[\"corr_matrix_pearson\"] = pear.round(3).to_dict(); R[\"corr_matrix_spearman\"] = spear.round(3).to_dict()\n", "R[\"corr_matrix_p\"] = pp.to_dict()\n", "display(pear.round(3)); display(spear.round(3))\n", "\n", "X = sm.add_constant(df[[\"part\", \"cgpa\", \"att10\"]])\n", "vif = pd.DataFrame({\"Predictor\": [\"Participation\", \"CGPA band\", \"Attendance (per 10 pts)\"],\n", " \"VIF\": [variance_inflation_factor(X.values, i) for i in range(1, 4)]})\n", "vif[\"Tolerance\"] = 1/vif[\"VIF\"]\n", "Xs = df[[\"part\",\"cgpa\",\"att10\"]]; Xs = (Xs - Xs.mean())/Xs.std()\n", "ev = np.linalg.eigvalsh(np.corrcoef(Xs.T.values)); cond_index = np.sqrt(ev.max()/ev.min())\n", "R[\"vif\"] = vif.to_dict(\"records\"); R[\"cond_index\"] = cond_index\n", "display(vif.round(3)); print(\"Condition index:\", round(cond_index, 2))\n", "\n", "fig, ax = plt.subplots(figsize=(9, 6.5))\n", "for k in range(1, 6):\n", " dat = df.loc[df.cgpa==k, \"att\"]; jit = rng.uniform(-0.2, 0.2, len(dat))\n", " ax.scatter(np.full(len(dat), k) + jit, dat, s=16, color=\"black\", alpha=0.35)\n", " ax.plot([k-0.3, k+0.3], [dat.mean()]*2, color=\"black\", lw=3)\n", "ax.set_xticks(range(1, 6)); ax.set_xticklabels([CG[k] for k in range(1, 6)])\n", "ax.set_xlabel(\"CGPA band\"); ax.set_ylabel(\"Attendance (%)\"); ax.set_ylim(45, 95)\n", "ax.text(0.03, 0.96, f\"Pearson $r$ = {r_ac:.2f}\", transform=ax.transAxes, fontsize=18, va=\"top\")\n", "savefig(fig, \"Fig4_attendance_by_cgpa\")\n" ] }, { "cell_type": "markdown", "id": "24944f30", "metadata": {}, "source": [ "## Analysis 9 — Binary logistic regression (RQ3)\n", "Nested models: M1 participation only; M2 + CGPA; M3 + attendance; M4 (primary) participation + CGPA + attendance." ] }, { "cell_type": "code", "execution_count": 12, "id": "ce90bb02", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:12:28.640197Z", "iopub.status.busy": "2026-09-27T19:12:28.639568Z", "iopub.status.idle": "2026-09-27T19:12:28.749875Z", "shell.execute_reply": "2026-09-27T19:12:28.748882Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "M1 placed ~ part\n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
BSEWaldpOROR_loOR_hi
Intercept-0.3180.2681.4090.2350.7270.4301.230
part3.5400.36594.1060.00034.45616.85370.446
\n", "
" ], "text/plain": [ " B SE Wald p OR OR_lo OR_hi\n", "Intercept -0.318 0.268 1.409 0.235 0.727 0.430 1.230\n", "part 3.540 0.365 94.106 0.000 34.456 16.853 70.446" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "M2 placed ~ part + cgpa\n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
BSEWaldpOROR_loOR_hi
Intercept-4.1090.67537.0880.00.0160.0040.062
part2.8650.45040.5220.017.5407.26142.372
cgpa1.8920.28643.6960.06.6303.78411.617
\n", "
" ], "text/plain": [ " B SE Wald p OR OR_lo OR_hi\n", "Intercept -4.109 0.675 37.088 0.0 0.016 0.004 0.062\n", "part 2.865 0.450 40.522 0.0 17.540 7.261 42.372\n", "cgpa 1.892 0.286 43.696 0.0 6.630 3.784 11.617" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "M3 placed ~ part + att10\n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
BSEWaldpOROR_loOR_hi
Intercept-13.1022.00342.7760.00.0000.0000.000
part2.8980.43644.1170.018.1387.71242.654
att102.0030.30941.9780.07.4114.04313.583
\n", "
" ], "text/plain": [ " B SE Wald p OR OR_lo OR_hi\n", "Intercept -13.102 2.003 42.776 0.0 0.000 0.000 0.000\n", "part 2.898 0.436 44.117 0.0 18.138 7.712 42.654\n", "att10 2.003 0.309 41.978 0.0 7.411 4.043 13.583" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "M4 placed ~ part + cgpa + att10\n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
BSEWaldpOROR_loOR_hi
Intercept-8.1742.52910.4490.0010.0000.0000.040
part2.8420.45439.2650.00017.1547.05141.731
cgpa1.3030.4398.7970.0033.6801.5568.705
att100.8250.4852.8860.0892.2810.8815.906
\n", "
" ], "text/plain": [ " B SE Wald p OR OR_lo OR_hi\n", "Intercept -8.174 2.529 10.449 0.001 0.000 0.000 0.040\n", "part 2.842 0.454 39.265 0.000 17.154 7.051 41.731\n", "cgpa 1.303 0.439 8.797 0.003 3.680 1.556 8.705\n", "att10 0.825 0.485 2.886 0.089 2.281 0.881 5.906" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
minus2LLLRLR_dfLR_pAICBICCoxSnellNagelkerkeMcFadden
M1221.783103.3001.00.0225.783234.2120.1870.3900.318
M2148.702176.3812.00.0154.702167.3450.2970.6220.543
M3155.168169.9152.00.0161.168173.8110.2880.6030.523
M4145.748179.3353.00.0153.748170.6070.3010.6300.552
\n", "
" ], "text/plain": [ " minus2LL LR LR_df LR_p AIC BIC CoxSnell Nagelkerke \\\n", "M1 221.783 103.300 1.0 0.0 225.783 234.212 0.187 0.390 \n", "M2 148.702 176.381 2.0 0.0 154.702 167.345 0.297 0.622 \n", "M3 155.168 169.915 2.0 0.0 161.168 173.811 0.288 0.603 \n", "M4 145.748 179.335 3.0 0.0 153.748 170.607 0.301 0.630 \n", "\n", " McFadden \n", "M1 0.318 \n", "M2 0.543 \n", "M3 0.523 \n", "M4 0.552 " ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
LRdfp
M2 vs M173.08141.00.0000
M3 vs M166.61531.00.0000
M4 vs M22.95321.00.0857
M4 vs M39.41931.00.0021
\n", "
" ], "text/plain": [ " LR df p\n", "M2 vs M1 73.0814 1.0 0.0000\n", "M3 vs M1 66.6153 1.0 0.0000\n", "M4 vs M2 2.9532 1.0 0.0857\n", "M4 vs M3 9.4193 1.0 0.0021" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "{'events': 50, 'params_M4': 3, 'epv': np.float64(16.666666666666668)}\n" ] } ], "source": [ "forms = {\"M1\": \"placed ~ part\", \"M2\": \"placed ~ part + cgpa\", \"M3\": \"placed ~ part + att10\",\n", " \"M4\": \"placed ~ part + cgpa + att10\"}\n", "fits = {k: smf.logit(f, df).fit(disp=0) for k, f in forms.items()}\n", "null = smf.logit(\"placed ~ 1\", df).fit(disp=0)\n", "n = len(df)\n", "\n", "def nagelkerke(m):\n", " cs = 1 - np.exp((2/n)*(null.llf - m.llf)); mx = 1 - np.exp((2/n)*null.llf); return cs, cs/mx\n", "\n", "def coef_table(m):\n", " ci = m.conf_int(); t = pd.DataFrame({\"B\": m.params, \"SE\": m.bse, \"Wald\": (m.params/m.bse)**2,\n", " \"p\": m.pvalues, \"OR\": np.exp(m.params), \"OR_lo\": np.exp(ci[0]), \"OR_hi\": np.exp(ci[1])})\n", " return t\n", "\n", "tabs, summ = {}, {}\n", "for k, m in fits.items():\n", " tabs[k] = coef_table(m); cs, nk = nagelkerke(m)\n", " lr = 2*(m.llf - null.llf)\n", " summ[k] = dict(minus2LL=-2*m.llf, LR=lr, LR_df=int(m.df_model), LR_p=st.chi2.sf(lr, m.df_model),\n", " AIC=m.aic, BIC=m.bic, CoxSnell=cs, Nagelkerke=nk, McFadden=m.prsquared)\n", " print(k, forms[k]); display(tabs[k].round(3))\n", "summ = pd.DataFrame(summ).T; display(summ.round(3))\n", "# nested LR tests\n", "lrt = {}\n", "for big, small in [(\"M2\",\"M1\"),(\"M3\",\"M1\"),(\"M4\",\"M2\"),(\"M4\",\"M3\")]:\n", " lr = 2*(fits[big].llf - fits[small].llf); dfd = fits[big].df_model - fits[small].df_model\n", " lrt[f\"{big} vs {small}\"] = dict(LR=lr, df=int(dfd), p=st.chi2.sf(lr, dfd))\n", "display(pd.DataFrame(lrt).T.round(4))\n", "R[\"models\"] = {k: tabs[k].reset_index().rename(columns={\"index\":\"term\"}).to_dict(\"records\") for k in tabs}\n", "R[\"model_fit\"] = summ.to_dict(\"index\"); R[\"lrt\"] = lrt\n", "R[\"epv\"] = dict(events=int(min(df.placed.sum(), (1-df.placed).sum())), params_M4=3,\n", " epv=min(df.placed.sum(), (1-df.placed).sum())/3)\n", "print(R[\"epv\"])\n" ] }, { "cell_type": "code", "execution_count": 13, "id": "d5d00498", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:12:28.752189Z", "iopub.status.busy": "2026-09-27T19:12:28.751553Z", "iopub.status.idle": "2026-09-27T19:12:31.250639Z", "shell.execute_reply": "2026-09-27T19:12:31.249394Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "2000 valid replicates\n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
termOR_boot_loOR_boot_hiprop_B_gt0
0Intercept0.0000.0350.000
1part8.03554.4561.000
2cgpa1.40112.1930.994
3att100.8418.0180.950
\n", "
" ], "text/plain": [ " term OR_boot_lo OR_boot_hi prop_B_gt0\n", "0 Intercept 0.000 0.035 0.000\n", "1 part 8.035 54.456 1.000\n", "2 cgpa 1.401 12.193 0.994\n", "3 att10 0.841 8.018 0.950" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# Bootstrap percentile CIs for primary model M4 (B = 2000)\n", "B = 2000; boot = []\n", "Xb = sm.add_constant(df[[\"part\",\"cgpa\",\"att10\"]]).values; yb = df.placed.values\n", "for b in range(B):\n", " i = rng.integers(0, n, n)\n", " try:\n", " r = sm.Logit(yb[i], Xb[i]).fit(disp=0, maxiter=200)\n", " if r.mle_retvals[\"converged\"]: boot.append(r.params)\n", " except Exception: pass\n", "boot = np.array(boot)\n", "bt = pd.DataFrame({\"term\": [\"Intercept\",\"part\",\"cgpa\",\"att10\"],\n", " \"OR_boot_lo\": np.exp(np.percentile(boot, 2.5, axis=0)),\n", " \"OR_boot_hi\": np.exp(np.percentile(boot, 97.5, axis=0)),\n", " \"prop_B_gt0\": (boot > 0).mean(axis=0)})\n", "R[\"boot_M4\"] = dict(valid=len(boot), table=bt.to_dict(\"records\")); print(len(boot), \"valid replicates\"); display(bt.round(3))\n" ] }, { "cell_type": "code", "execution_count": 14, "id": "03c480b1", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:12:31.252816Z", "iopub.status.busy": "2026-09-27T19:12:31.252631Z", "iopub.status.idle": "2026-09-27T19:12:31.274801Z", "shell.execute_reply": "2026-09-27T19:12:31.273799Z" } }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
partcgpaatt10problohi
0015.7270.1040.0460.220
1026.2970.4070.2610.572
2037.2310.8450.6920.930
3048.1930.9780.9170.994
4058.2630.9940.9650.999
5115.7270.6670.5090.794
6126.2970.9220.8670.955
7137.2310.9890.9720.996
8148.1930.9990.9941.000
9158.2631.0000.9971.000
\n", "
" ], "text/plain": [ " part cgpa att10 prob lo hi\n", "0 0 1 5.727 0.104 0.046 0.220\n", "1 0 2 6.297 0.407 0.261 0.572\n", "2 0 3 7.231 0.845 0.692 0.930\n", "3 0 4 8.193 0.978 0.917 0.994\n", "4 0 5 8.263 0.994 0.965 0.999\n", "5 1 1 5.727 0.667 0.509 0.794\n", "6 1 2 6.297 0.922 0.867 0.955\n", "7 1 3 7.231 0.989 0.972 0.996\n", "8 1 4 8.193 0.999 0.994 1.000\n", "9 1 5 8.263 1.000 0.997 1.000" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# Predicted probabilities from M4 (for interpretation; covariates at observed means)\n", "m4 = fits[\"M4\"]\n", "grid = []\n", "for p_ in [0, 1]:\n", " for k in range(1, 6):\n", " grid.append(dict(part=p_, cgpa=k, att10=df.loc[df.cgpa==k, \"att10\"].mean()))\n", "grid = pd.DataFrame(grid); pr = m4.get_prediction(grid).summary_frame()\n", "grid[\"prob\"] = pr[\"predicted\"].values if \"predicted\" in pr else m4.predict(grid).values\n", "grid[\"lo\"] = pr[\"ci_lower\"].values; grid[\"hi\"] = pr[\"ci_upper\"].values\n", "R[\"pred_M4\"] = grid.to_dict(\"records\"); display(grid.round(3))\n" ] }, { "cell_type": "markdown", "id": "49a65379", "metadata": {}, "source": [ "## Analysis 10 — Model diagnostics (RQ3)" ] }, { "cell_type": "code", "execution_count": 15, "id": "eb39958e", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:12:31.277130Z", "iopub.status.busy": "2026-09-27T19:12:31.276516Z", "iopub.status.idle": "2026-09-27T19:12:31.342081Z", "shell.execute_reply": "2026-09-27T19:12:31.341104Z" } }, "outputs": [ { "data": { "text/plain": [ "HL 12.0603\n", "HL_df 8\n", "HL_p 0.1485\n", "BT_att_p 0.6814\n", "BT_cgpa_p 0.4047\n", "cook_max 0.1518\n", "n_cook_gt_4n 35\n", "n_cook_gt_1 0\n", "lev_max 0.0561\n", "n_lev_gt_3k 82\n", "n_absres_gt2 9\n", "n_absres_gt3 4\n", "max_abs_dev 3.6224\n", "converged 1\n", "iterations 9\n", "separation_check_max_abs_B 8.1741\n", "max_SE 2.5287\n", "dtype: object" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
obsexpn
grp
01715.5950
13739.2950
24746.8650
35049.2450
45049.5050
55049.7250
64949.9150
75049.9450
85049.9650
95049.9950
\n", "
" ], "text/plain": [ " obs exp n\n", "grp \n", "0 17 15.59 50\n", "1 37 39.29 50\n", "2 47 46.86 50\n", "3 50 49.24 50\n", "4 50 49.50 50\n", "5 50 49.72 50\n", "6 49 49.91 50\n", "7 50 49.94 50\n", "8 50 49.96 50\n", "9 50 49.99 50" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
Respondent_IDpartcgpaattplacedp_hat_M4cookleveragestd_pearson_resid
0R0011481.000.9990.1520.001-26.579
136R1370162.010.1470.0510.0332.450
257R2580160.010.1270.0450.0252.649
228R2290255.010.2630.0440.0561.725
238R2391256.000.8690.0420.024-2.607
196R1970366.000.7640.0390.043-1.842
309R3100367.000.7790.0370.038-1.914
227R2280369.000.8060.0350.031-2.071
181R1821163.000.7620.0330.038-1.827
374R3751163.000.7620.0330.038-1.827
379R3801268.000.9470.0330.007-4.239
442R4431268.000.9470.0330.007-4.239
460R4611268.000.9470.0330.007-4.239
244R2451258.000.8870.0310.015-2.819
433R4341162.000.7470.0280.035-1.750
378R3791162.000.7470.0280.035-1.750
466R4670260.010.3500.0160.0331.387
92R0931159.000.6980.0160.027-1.540
203R2041151.000.5440.0150.045-1.118
72R0731151.000.5440.0150.045-1.118
403R4041158.000.6800.0140.025-1.477
7R0080269.000.5300.0140.045-1.088
393R3941152.000.5640.0140.039-1.161
175R1761150.010.5240.0130.0520.980
219R2201157.000.6620.0130.025-1.417
493R4941154.000.6040.0120.030-1.255
311R3120262.010.3880.0120.0281.274
431R4320268.000.5100.0110.039-1.041
444R4450269.010.5300.0110.0450.963
259R2600267.010.4890.0100.0351.040
239R2400266.010.4690.0100.0311.082
237R2380265.010.4480.0100.0291.126
184R1850265.010.4480.0100.0291.126
280R2810265.010.4480.0100.0291.126
17R0181152.010.5640.0080.0390.896
\n", "
" ], "text/plain": [ " Respondent_ID part cgpa att placed p_hat_M4 cook leverage \\\n", "0 R001 1 4 81.0 0 0.999 0.152 0.001 \n", "136 R137 0 1 62.0 1 0.147 0.051 0.033 \n", "257 R258 0 1 60.0 1 0.127 0.045 0.025 \n", "228 R229 0 2 55.0 1 0.263 0.044 0.056 \n", "238 R239 1 2 56.0 0 0.869 0.042 0.024 \n", "196 R197 0 3 66.0 0 0.764 0.039 0.043 \n", "309 R310 0 3 67.0 0 0.779 0.037 0.038 \n", "227 R228 0 3 69.0 0 0.806 0.035 0.031 \n", "181 R182 1 1 63.0 0 0.762 0.033 0.038 \n", "374 R375 1 1 63.0 0 0.762 0.033 0.038 \n", "379 R380 1 2 68.0 0 0.947 0.033 0.007 \n", "442 R443 1 2 68.0 0 0.947 0.033 0.007 \n", "460 R461 1 2 68.0 0 0.947 0.033 0.007 \n", "244 R245 1 2 58.0 0 0.887 0.031 0.015 \n", "433 R434 1 1 62.0 0 0.747 0.028 0.035 \n", "378 R379 1 1 62.0 0 0.747 0.028 0.035 \n", "466 R467 0 2 60.0 1 0.350 0.016 0.033 \n", "92 R093 1 1 59.0 0 0.698 0.016 0.027 \n", "203 R204 1 1 51.0 0 0.544 0.015 0.045 \n", "72 R073 1 1 51.0 0 0.544 0.015 0.045 \n", "403 R404 1 1 58.0 0 0.680 0.014 0.025 \n", "7 R008 0 2 69.0 0 0.530 0.014 0.045 \n", "393 R394 1 1 52.0 0 0.564 0.014 0.039 \n", "175 R176 1 1 50.0 1 0.524 0.013 0.052 \n", "219 R220 1 1 57.0 0 0.662 0.013 0.025 \n", "493 R494 1 1 54.0 0 0.604 0.012 0.030 \n", "311 R312 0 2 62.0 1 0.388 0.012 0.028 \n", "431 R432 0 2 68.0 0 0.510 0.011 0.039 \n", "444 R445 0 2 69.0 1 0.530 0.011 0.045 \n", "259 R260 0 2 67.0 1 0.489 0.010 0.035 \n", "239 R240 0 2 66.0 1 0.469 0.010 0.031 \n", "237 R238 0 2 65.0 1 0.448 0.010 0.029 \n", "184 R185 0 2 65.0 1 0.448 0.010 0.029 \n", "280 R281 0 2 65.0 1 0.448 0.010 0.029 \n", "17 R018 1 1 52.0 1 0.564 0.008 0.039 \n", "\n", " std_pearson_resid \n", "0 -26.579 \n", "136 2.450 \n", "257 2.649 \n", "228 1.725 \n", "238 -2.607 \n", "196 -1.842 \n", "309 -1.914 \n", "227 -2.071 \n", "181 -1.827 \n", "374 -1.827 \n", "379 -4.239 \n", "442 -4.239 \n", "460 -4.239 \n", "244 -2.819 \n", "433 -1.750 \n", "378 -1.750 \n", "466 1.387 \n", "92 -1.540 \n", "203 -1.118 \n", "72 -1.118 \n", "403 -1.477 \n", "7 -1.088 \n", "393 -1.161 \n", "175 0.980 \n", "219 -1.417 \n", "493 -1.255 \n", "311 1.274 \n", "431 -1.041 \n", "444 0.963 \n", "259 1.040 \n", "239 1.082 \n", "237 1.126 \n", "184 1.126 \n", "280 1.126 \n", "17 0.896 " ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "def hosmer_lemeshow(y, p, g=10):\n", " d = pd.DataFrame({\"y\": y, \"p\": p}); d[\"grp\"] = pd.qcut(d.p.rank(method=\"first\"), g, labels=False)\n", " agg = d.groupby(\"grp\").agg(obs=(\"y\",\"sum\"), exp=(\"p\",\"sum\"), n=(\"y\",\"size\"))\n", " hl = (((agg.obs-agg.exp)**2)/(agg.exp*(1-agg.exp/agg.n))).sum(); return hl, g-2, st.chi2.sf(hl, g-2), agg\n", "\n", "p_hat = m4.predict(df)\n", "hl, hl_df, hl_p, hl_tab = hosmer_lemeshow(df.placed, p_hat)\n", "# Box–Tidwell linearity-in-the-logit\n", "bt_df = df.assign(att_ln=df.att*np.log(df.att), cg_ln=df.cgpa*np.log(df.cgpa))\n", "bt_fit = smf.logit(\"placed ~ part + cgpa + att + att_ln + cg_ln\", bt_df).fit(disp=0)\n", "# influence\n", "glm = sm.GLM(df.placed, sm.add_constant(df[[\"part\",\"cgpa\",\"att10\"]]), family=sm.families.Binomial()).fit()\n", "inf = glm.get_influence(); cook = inf.cooks_distance[0]; lev = inf.hat_matrix_diag\n", "spr = inf.resid_studentized; dev = glm.resid_deviance\n", "k_par = 4\n", "diag = dict(HL=hl, HL_df=hl_df, HL_p=hl_p,\n", " BT_att_p=bt_fit.pvalues[\"att_ln\"], BT_cgpa_p=bt_fit.pvalues[\"cg_ln\"],\n", " cook_max=cook.max(), n_cook_gt_4n=int((cook > 4/n).sum()), n_cook_gt_1=int((cook > 1).sum()),\n", " lev_max=lev.max(), n_lev_gt_3k=int((lev > 3*k_par/n).sum()),\n", " n_absres_gt2=int((np.abs(spr) > 2).sum()), n_absres_gt3=int((np.abs(spr) > 3).sum()),\n", " max_abs_dev=np.abs(dev).max(), converged=bool(m4.mle_retvals[\"converged\"]),\n", " iterations=int(m4.mle_retvals[\"iterations\"]),\n", " separation_check_max_abs_B=float(np.abs(m4.params).max()), max_SE=float(m4.bse.max()))\n", "R[\"diag\"] = diag; R[\"hl_table\"] = hl_tab.reset_index().to_dict(\"records\")\n", "display(pd.Series(diag).round(4)); display(hl_tab.round(2))\n", "df[\"cook\"] = cook; df[\"leverage\"] = lev; df[\"std_pearson_resid\"] = spr; df[\"p_hat_M4\"] = p_hat\n", "infl_cases = df.loc[cook > 4/n, [\"Respondent_ID\",\"part\",\"cgpa\",\"att\",\"placed\",\"p_hat_M4\",\"cook\",\"leverage\",\"std_pearson_resid\"]]\n", "R[\"infl_cases\"] = infl_cases.round(4).to_dict(\"records\"); display(infl_cases.sort_values(\"cook\", ascending=False).round(3))\n" ] }, { "cell_type": "markdown", "id": "fc1f9e12", "metadata": {}, "source": [ "## Analysis 11 — Sensitivity analyses (RQ2/RQ3)" ] }, { "cell_type": "code", "execution_count": 16, "id": "00941978", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:12:31.344747Z", "iopub.status.busy": "2026-09-27T19:12:31.343835Z", "iopub.status.idle": "2026-09-27T19:12:31.434502Z", "shell.execute_reply": "2026-09-27T19:12:31.433410Z" } }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
analysistermORlohip
0S0 Primary model (M4)part17.1547.05141.7310.000
1S0 Primary model (M4)cgpa3.6801.5568.7050.003
2S0 Primary model (M4)att102.2810.8815.9060.089
3S1 Firth penalised likelihoodpart15.6416.58037.1810.000
4S1 Firth penalised likelihoodcgpa3.5131.5178.1340.003
5S1 Firth penalised likelihoodatt102.2040.8685.5930.096
6S2 Bootstrap percentile CI (B = 2000)part17.1548.03554.456NaN
7S2 Bootstrap percentile CI (B = 2000)cgpa3.6801.40112.193NaN
8S2 Bootstrap percentile CI (B = 2000)att102.2810.8418.018NaN
9S3 + Gender and field of studypart20.5577.68854.9640.000
10S3 + Gender and field of studycgpa4.1881.68510.4060.002
11S3 + Gender and field of studyatt102.4760.9156.7030.074
12S4 Excluding |std. residual| > 3 (n = 496; Firth)part27.8719.83578.9770.000
13S4 Excluding |std. residual| > 3 (n = 496; Firth)cgpa3.7061.3989.8270.008
14S4 Excluding |std. residual| > 3 (n = 496; Firth)att103.8191.28111.3890.016
15S5 Composite academic index replaces CGPA and ...part17.1047.06041.4370.000
16S5 Composite academic index replaces CGPA and ...acad_idx9.9685.00919.8370.000
17S6 CGPA as categories (ref. 5–6)C(cg3)[T.6–7]4.0381.33712.2000.013
18S6 CGPA as categories (ref. 5–6)C(cg3)[T.≥7]18.9882.725132.3060.003
19S6 CGPA as categories (ref. 5–6)part19.1097.51948.5650.000
20S6 CGPA as categories (ref. 5–6)att102.3030.9095.8360.079
21S7 Fisher's exact test (unadjusted participation)part34.456NaNNaN0.000
22S8 cgpa entered without the other academic var...cgpa6.6303.78411.6170.000
23S8 att10 entered without the other academic va...att107.4114.04313.5830.000
\n", "
" ], "text/plain": [ " analysis term OR lo hi p\n", "0 S0 Primary model (M4) part 17.154 7.051 41.731 0.000\n", "1 S0 Primary model (M4) cgpa 3.680 1.556 8.705 0.003\n", "2 S0 Primary model (M4) att10 2.281 0.881 5.906 0.089\n", "3 S1 Firth penalised likelihood part 15.641 6.580 37.181 0.000\n", "4 S1 Firth penalised likelihood cgpa 3.513 1.517 8.134 0.003\n", "5 S1 Firth penalised likelihood att10 2.204 0.868 5.593 0.096\n", "6 S2 Bootstrap percentile CI (B = 2000) part 17.154 8.035 54.456 NaN\n", "7 S2 Bootstrap percentile CI (B = 2000) cgpa 3.680 1.401 12.193 NaN\n", "8 S2 Bootstrap percentile CI (B = 2000) att10 2.281 0.841 8.018 NaN\n", "9 S3 + Gender and field of study part 20.557 7.688 54.964 0.000\n", "10 S3 + Gender and field of study cgpa 4.188 1.685 10.406 0.002\n", "11 S3 + Gender and field of study att10 2.476 0.915 6.703 0.074\n", "12 S4 Excluding |std. residual| > 3 (n = 496; Firth) part 27.871 9.835 78.977 0.000\n", "13 S4 Excluding |std. residual| > 3 (n = 496; Firth) cgpa 3.706 1.398 9.827 0.008\n", "14 S4 Excluding |std. residual| > 3 (n = 496; Firth) att10 3.819 1.281 11.389 0.016\n", "15 S5 Composite academic index replaces CGPA and ... part 17.104 7.060 41.437 0.000\n", "16 S5 Composite academic index replaces CGPA and ... acad_idx 9.968 5.009 19.837 0.000\n", "17 S6 CGPA as categories (ref. 5–6) C(cg3)[T.6–7] 4.038 1.337 12.200 0.013\n", "18 S6 CGPA as categories (ref. 5–6) C(cg3)[T.≥7] 18.988 2.725 132.306 0.003\n", "19 S6 CGPA as categories (ref. 5–6) part 19.109 7.519 48.565 0.000\n", "20 S6 CGPA as categories (ref. 5–6) att10 2.303 0.909 5.836 0.079\n", "21 S7 Fisher's exact test (unadjusted participation) part 34.456 NaN NaN 0.000\n", "22 S8 cgpa entered without the other academic var... cgpa 6.630 3.784 11.617 0.000\n", "23 S8 att10 entered without the other academic va... att10 7.411 4.043 13.583 0.000" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "def firth_logit(y, X, maxit=100, tol=1e-8):\n", " \"\"\"Firth (1993) bias-reduced logistic regression via modified score (Heinze & Schemper, 2002).\"\"\"\n", " y = np.asarray(y, float); X = np.asarray(X, float); b = np.zeros(X.shape[1])\n", " def pll(b):\n", " p = 1/(1+np.exp(-X@b)); W = p*(1-p); I = X.T@(X*W[:, None])\n", " return np.sum(y*np.log(p)+(1-y)*np.log(1-p)) + 0.5*np.linalg.slogdet(I)[1]\n", " for it in range(maxit):\n", " p = 1/(1+np.exp(-X@b)); W = p*(1-p); I = X.T@(X*W[:, None]); Iinv = np.linalg.inv(I)\n", " h = np.einsum(\"ij,jk,ik->i\", X*np.sqrt(W)[:, None], Iinv, X*np.sqrt(W)[:, None])\n", " U = X.T@(y - p + h*(0.5 - p)); step = Iinv@U; s = 1.0; old = pll(b)\n", " while pll(b + s*step) < old - 1e-12 and s > 1e-4: s /= 2\n", " b = b + s*step\n", " if np.max(np.abs(s*step)) < tol: break\n", " p = 1/(1+np.exp(-X@b)); W = p*(1-p); se = np.sqrt(np.diag(np.linalg.inv(X.T@(X*W[:, None]))))\n", " zz = b/se; return b, se, 2*st.norm.sf(np.abs(zz))\n", "\n", "def or_row(label, b, se, pv, names):\n", " rows = []\n", " for nm, bb, ss, pp in zip(names, b, se, pv):\n", " rows.append(dict(analysis=label, term=nm, OR=np.exp(bb), lo=np.exp(bb-1.96*ss), hi=np.exp(bb+1.96*ss), p=pp))\n", " return rows\n", "\n", "sens = []\n", "# S0 primary\n", "t = tabs[\"M4\"]; sens += [dict(analysis=\"S0 Primary model (M4)\", term=i, OR=t.loc[i,\"OR\"], lo=t.loc[i,\"OR_lo\"], hi=t.loc[i,\"OR_hi\"], p=t.loc[i,\"p\"]) for i in [\"part\",\"cgpa\",\"att10\"]]\n", "# S1 Firth\n", "Xf = sm.add_constant(df[[\"part\",\"cgpa\",\"att10\"]]); b, se, pv = firth_logit(df.placed, Xf)\n", "sens += or_row(\"S1 Firth penalised likelihood\", b[1:], se[1:], pv[1:], [\"part\",\"cgpa\",\"att10\"])\n", "# S2 bootstrap CIs\n", "for r in R[\"boot_M4\"][\"table\"][1:]:\n", " sens.append(dict(analysis=\"S2 Bootstrap percentile CI (B = 2000)\", term=r[\"term\"], OR=t.loc[r[\"term\"],\"OR\"], lo=r[\"OR_boot_lo\"], hi=r[\"OR_boot_hi\"], p=np.nan))\n", "# S3 add gender + field (academic year omitted: aliased with field)\n", "m_s3 = smf.logit(\"placed ~ part + cgpa + att10 + male + C(field)\", df).fit(disp=0); t3 = coef_table(m_s3)\n", "sens += [dict(analysis=\"S3 + Gender and field of study\", term=i, OR=t3.loc[i,\"OR\"], lo=t3.loc[i,\"OR_lo\"], hi=t3.loc[i,\"OR_hi\"], p=t3.loc[i,\"p\"]) for i in [\"part\",\"cgpa\",\"att10\"]]\n", "R[\"S3_full\"] = t3.reset_index().rename(columns={\"index\":\"term\"}).to_dict(\"records\")\n", "R[\"S3_fit\"] = dict(LR_add=2*(m_s3.llf-m4.llf), df_add=int(m_s3.df_model-m4.df_model), p_add=st.chi2.sf(2*(m_s3.llf-m4.llf), m_s3.df_model-m4.df_model))\n", "# S4 exclude poorly fitted cases (|standardised Pearson residual| > 3).\n", "# Note: removing all cases with Cook's D > 4/n (35 cases) removes most non-placed misfits and\n", "# produces complete separation, so the residual-based rule is used; the Cook's rule is reported in the diagnostics.\n", "keep = np.abs(df.std_pearson_resid) <= 3\n", "b4, se4, pv4 = firth_logit(df.placed[keep], sm.add_constant(df.loc[keep, [\"part\",\"cgpa\",\"att10\"]]))\n", "m_s4 = smf.logit(forms[\"M4\"], df[keep]).fit(disp=0, maxiter=200)\n", "R[\"S4_ml\"] = dict(converged=bool(m_s4.mle_retvals[\"converged\"]), OR=np.exp(m_s4.params).to_dict(), p=m_s4.pvalues.to_dict(), n=int(keep.sum()))\n", "sens += or_row(f\"S4 Excluding |std. residual| > 3 (n = {keep.sum()}; Firth)\", b4[1:], se4[1:], pv4[1:], [\"part\",\"cgpa\",\"att10\"])\n", "R[\"S4_excluded\"] = df.loc[~keep, [\"Respondent_ID\",\"Field of Study\",\"part\",\"cgpa\",\"att\",\"placed\",\"p_hat_M4\",\"std_pearson_resid\"]].round(3).to_dict(\"records\")\n", "# S5 composite academic-engagement index (mean of z-scores)\n", "df[\"acad_idx\"] = ((df.att-df.att.mean())/df.att.std() + (df.cgpa-df.cgpa.mean())/df.cgpa.std())/2\n", "m_s5 = smf.logit(\"placed ~ part + acad_idx\", df).fit(disp=0); t5 = coef_table(m_s5)\n", "sens += [dict(analysis=\"S5 Composite academic index replaces CGPA and attendance\", term=i, OR=t5.loc[i,\"OR\"], lo=t5.loc[i,\"OR_lo\"], hi=t5.loc[i,\"OR_hi\"], p=t5.loc[i,\"p\"]) for i in [\"part\",\"acad_idx\"]]\n", "# S6 CGPA categorical (5–6 ref; 6–7; >=7) + attendance\n", "df[\"cg3\"] = pd.Categorical(np.select([df.cgpa==1, df.cgpa==2], [\"5–6\",\"6–7\"], \"≥7\"), [\"5–6\",\"6–7\",\"≥7\"])\n", "m_s6 = smf.logit(\"placed ~ part + C(cg3) + att10\", df).fit(disp=0); t6 = coef_table(m_s6)\n", "for i in t6.index[1:]:\n", " sens.append(dict(analysis=\"S6 CGPA as categories (ref. 5–6)\", term=i, OR=t6.loc[i,\"OR\"], lo=t6.loc[i,\"OR_lo\"], hi=t6.loc[i,\"OR_hi\"], p=t6.loc[i,\"p\"]))\n", "# S7 Participation: exact test of association (unadjusted)\n", "sens.append(dict(analysis=\"S7 Fisher's exact test (unadjusted participation)\", term=\"part\", OR=st.fisher_exact(ct.values)[0], lo=np.nan, hi=np.nan, p=R[\"h1_chi\"][\"fisher_p\"]))\n", "# S8 Attendance-only and CGPA-only adjusted for participation (M3, M2)\n", "for k, term in [(\"M2\",\"cgpa\"),(\"M3\",\"att10\")]:\n", " tt = tabs[k]; sens.append(dict(analysis=f\"S8 {term} entered without the other academic variable ({k})\", term=term, OR=tt.loc[term,\"OR\"], lo=tt.loc[term,\"OR_lo\"], hi=tt.loc[term,\"OR_hi\"], p=tt.loc[term,\"p\"]))\n", "sens = pd.DataFrame(sens); R[\"sens\"] = sens.to_dict(\"records\")\n", "pd.set_option(\"display.width\", 200); display(sens.round(3))\n" ] }, { "cell_type": "code", "execution_count": 17, "id": "505e1049", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:12:31.436320Z", "iopub.status.busy": "2026-09-27T19:12:31.436112Z", "iopub.status.idle": "2026-09-27T19:12:34.618879Z", "shell.execute_reply": "2026-09-27T19:12:34.617694Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# Forest plot: adjusted ORs across primary and sensitivity models (single panel, grouped rows)\n", "def get(label_prefix, term):\n", " r = sens[(sens.analysis.str.startswith(label_prefix)) & (sens.term == term)].iloc[0]; return r.OR, r.lo, r.hi\n", "blocks = [(\"Participation (yes vs no)\", \"part\", [(\"Unadjusted (M1)\", tabs[\"M1\"].loc[\"part\", [\"OR\",\"OR_lo\",\"OR_hi\"]].values)]),\n", " (\"CGPA (per band)\", \"cgpa\", [(\"Participation-adjusted (M2)\", tabs[\"M2\"].loc[\"cgpa\", [\"OR\",\"OR_lo\",\"OR_hi\"]].values)]),\n", " (\"Attendance (per 10 points)\", \"att10\", [(\"Participation-adjusted (M3)\", tabs[\"M3\"].loc[\"att10\", [\"OR\",\"OR_lo\",\"OR_hi\"]].values)])]\n", "common = [(\"Primary model (M4)\", \"S0\"), (\"Firth penalised\", \"S1\"), (\"+ Gender and field\", \"S3\"), (\"Misfits excluded\", \"S4\")]\n", "rows = []\n", "for title, term, first in blocks:\n", " rows.append((title, None, True))\n", " for lab_, vals in first: rows.append((lab_, vals, False))\n", " for lab_, pre in common: rows.append((lab_, get(pre, term), False))\n", "fig, ax = plt.subplots(figsize=(10.5, 11.5))\n", "y = len(rows)\n", "yt, yl, bold = [], [], []\n", "for lab_, vals, hdr in rows:\n", " y -= 1; yt.append(y); yl.append(lab_); bold.append(hdr)\n", " if vals is not None:\n", " o, lo, hi = vals\n", " ax.errorbar(o, y, xerr=[[o-lo], [hi-o]], fmt=\"s\" if lab_.startswith(\"Primary\") else \"o\",\n", " color=\"black\", mfc=\"black\" if lab_.startswith(\"Primary\") else \"white\", ms=11, capsize=6, lw=2, mew=2)\n", "ax.axvline(1, ls=\"--\", color=\"grey\", lw=1.6); ax.set_xscale(\"log\"); ax.set_xlim(0.5, 120)\n", "ax.set_xticks([0.5, 1, 2, 5, 10, 20, 50, 100]); ax.xaxis.set_major_formatter(matplotlib.ticker.FuncFormatter(lambda v, _: f\"{v:g}\"))\n", "ax.set_yticks(yt); ax.set_yticklabels(yl)\n", "for t, b in zip(ax.get_yticklabels(), bold):\n", " if b: t.set_fontweight(\"bold\")\n", "ax.tick_params(axis=\"y\", length=0); ax.spines[\"left\"].set_visible(False)\n", "ax.set_xlabel(\"Odds ratio with 95% CI (log scale)\")\n", "savefig(fig, \"Fig5_forest_sensitivity\")\n" ] }, { "cell_type": "markdown", "id": "2646107e", "metadata": {}, "source": [ "## Analysis 11b — E-value for unmeasured confounding (participation)\n", "VanderWeele & Ding (2017). Because placement is common (90%), the OR is converted to an approximate risk ratio using RR ≈ √OR (conservative common-outcome approximation)." ] }, { "cell_type": "code", "execution_count": 18, "id": "df4fa041", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:12:34.621324Z", "iopub.status.busy": "2026-09-27T19:12:34.620584Z", "iopub.status.idle": "2026-09-27T19:12:34.627411Z", "shell.execute_reply": "2026-09-27T19:12:34.626442Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "{'OR': np.float64(17.154082398043798), 'OR_lo': np.float64(7.051435673697116), 'RR_approx': np.float64(4.141748712566203), 'RR_lo_approx': np.float64(2.6554539487057793), 'E_point': np.float64(7.749009956983857), 'E_ci': np.float64(4.752113607542029), 'method': 'RR approximated by sqrt(OR) for a common outcome (VanderWeele & Ding 2017)'}\n" ] } ], "source": [ "def evalue_rr(rr):\n", " rr = max(rr, 1/rr); return rr + np.sqrt(rr*(rr-1))\n", "t = tabs[\"M4\"].loc[\"part\"]\n", "rr_pt, rr_lo = np.sqrt(t[\"OR\"]), np.sqrt(t[\"OR_lo\"])\n", "R[\"evalue\"] = dict(OR=t[\"OR\"], OR_lo=t[\"OR_lo\"], RR_approx=rr_pt, RR_lo_approx=rr_lo,\n", " E_point=evalue_rr(rr_pt), E_ci=evalue_rr(rr_lo) if rr_lo > 1 else 1.0,\n", " method=\"RR approximated by sqrt(OR) for a common outcome (VanderWeele & Ding 2017)\")\n", "print(R[\"evalue\"])\n" ] }, { "cell_type": "markdown", "id": "9e009f8f", "metadata": {}, "source": [ "## Analysis 12 — Predictive performance (supplementary only; not central to RQs)\n", "Reported only in the Supplementary Information, with optimism correction, because the RQs are associational rather than predictive." ] }, { "cell_type": "code", "execution_count": 19, "id": "086f78bd", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:12:34.629681Z", "iopub.status.busy": "2026-09-27T19:12:34.628945Z", "iopub.status.idle": "2026-09-27T19:12:36.735045Z", "shell.execute_reply": "2026-09-27T19:12:36.733682Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "{'AUC_apparent': 0.9532222222222223, 'AUC_optimism_corrected': np.float64(0.9511142065957384), 'Brier': 0.043269404881893306, 'accuracy': 0.936, 'sensitivity': np.float64(0.9777777777777777), 'specificity': np.float64(0.56), 'null_accuracy': np.float64(0.9), 'boot_reps': 500}\n" ] } ], "source": [ "auc_app = roc_auc_score(df.placed, p_hat); brier = brier_score_loss(df.placed, p_hat)\n", "opt = []\n", "for b in range(500):\n", " i = rng.integers(0, n, n)\n", " try:\n", " r = sm.Logit(yb[i], Xb[i]).fit(disp=0)\n", " opt.append(roc_auc_score(yb[i], r.predict(Xb[i])) - roc_auc_score(yb, r.predict(Xb)))\n", " except Exception: pass\n", "auc_corr = auc_app - np.mean(opt)\n", "cls = (p_hat >= 0.5).astype(int)\n", "tp = int(((cls==1)&(df.placed==1)).sum()); tn = int(((cls==0)&(df.placed==0)).sum())\n", "R[\"pred\"] = dict(AUC_apparent=auc_app, AUC_optimism_corrected=auc_corr, Brier=brier,\n", " accuracy=(tp+tn)/n, sensitivity=tp/df.placed.sum(), specificity=tn/(1-df.placed).sum(),\n", " null_accuracy=df.placed.mean(), boot_reps=len(opt))\n", "print(R[\"pred\"])\n" ] }, { "cell_type": "markdown", "id": "ab472e57", "metadata": {}, "source": [ "## Export results" ] }, { "cell_type": "code", "execution_count": 20, "id": "1ec03ec2", "metadata": { "execution": { "iopub.execute_input": "2026-09-27T19:12:36.737857Z", "iopub.status.busy": "2026-09-27T19:12:36.736982Z", "iopub.status.idle": "2026-09-27T19:12:36.758366Z", "shell.execute_reply": "2026-09-27T19:12:36.757431Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Saved: ['Supplementary_Material_1.xlsx', 'analytic_dataset_with_diagnostics.csv', 'results.json', 'analytic_dataset.csv'] ['Fig3_attendance_by_placement.tiff', 'Fig1_placement_by_participation.pdf', 'Fig2_placement_by_cgpa.pdf', 'Fig4_attendance_by_cgpa.png', 'Fig5_forest_sensitivity.tiff', 'Fig1_placement_by_participation.png', 'Fig5_forest_sensitivity.png', 'Fig5_forest_sensitivity.pdf', 'Fig3_attendance_by_placement.png', 'Fig3_attendance_by_placement.pdf', 'Fig1_placement_by_participation.tiff', 'Fig4_attendance_by_cgpa.pdf', 'Fig2_placement_by_cgpa.tiff', 'Fig4_attendance_by_cgpa.tiff', 'Fig2_placement_by_cgpa.png']\n" ] } ], "source": [ "def conv(o):\n", " if isinstance(o, (np.integer,)): return int(o)\n", " if isinstance(o, (np.floating,)): return float(o)\n", " if isinstance(o, (np.bool_,)): return bool(o)\n", " if isinstance(o, tuple): return list(o)\n", " return str(o)\n", "with open(f\"{OUT}/results.json\", \"w\") as f: json.dump(R, f, default=conv, indent=1)\n", "df.drop(columns=[\"field\",\"cg3\"]).to_csv(f\"{OUT}/analytic_dataset_with_diagnostics.csv\", index=False)\n", "print(\"Saved:\", os.listdir(OUT), os.listdir(FIG))\n" ] } ], "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": "3.12.3" } }, "nbformat": 4, "nbformat_minor": 5 }