I'll help you analyze the survey data on student anxiety levels using Python. Since you haven't provided the raw data, I'll first create a realistic dataset based on common anxiety survey scenarios and then perform a complete analysis. Here's a comprehensive solution:

```python
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from scipy import stats
from scipy.stats import shapiro, levene
import statsmodels.api as sm
from statsmodels.stats.anova import anova_lm
from statsmodels.stats.multicomp import pairwise_tukeyhsd
from statsmodels.formula.api import ols
import warnings
warnings.filterwarnings('ignore')

Set style for publication-ready plots

plt.style.use('seaborn-v0_8')
sns.set_palette("colorblind")

Create realistic survey data for 300 participants

np.random.seed(42) # For reproducibility

n = 300
data = {
'participant_id': range(1, n+1),
'age': np.random.normal(20, 2, n).astype(int),
'gender': np.random.choice(['Male', 'Female', 'Non-binary'], n, p=[0.48, 0.48, 0.04]),
'year': np.random.choice(['Freshman', 'Sophomore', 'Junior', 'Senior'], n, p=[0.25, 0.25, 0.25, 0.25]),
'gpa': np.clip(np.random.normal(3.2, 0.5, n), 0, 4),

# Anxiety measures (GAD-7 scale: 0-21, higher = more anxiety)
'anxiety_score': np.random.normal(10, 4, n),
'sleep_quality': np.random.choice([1, 2, 3, 4, 5], n, p=[0.1, 0.15, 0.3, 0.25, 0.2]), # 1=poor, 5=excellent

# Categorical variables for additional analysis
'major_stress': np.random.choice(['Low', 'Moderate', 'High'], n, p=[0.3, 0.4, 0.3]),
'academic_load': np.random.choice(['Light', 'Moderate', 'Heavy'], n, p=[0.2, 0.5, 0.3]),
'social_support': np.random.choice(['Low', 'Moderate', 'High'], n, p=[0.25, 0.5, 0.25])
}

Create the dataset

df = pd.DataFrame(data)

Introduce some realistic missing values and outliers

Missing data (5% missing completely at random)

mask_missing = np.random.random(n) < 0.05
df.loc[maskmissing, ['anxietyscore', 'gpa', 'sleep_quality']] = np.nan

Add some outliers for anxiety scores (beyond reasonable range)

outlier_indices = np.random.choice(df.index, 15, replace=False)
df.loc[outlierindices[:5], 'anxietyscore'] = np.random.uniform(25, 35, 5) # High outliers
df.loc[outlierindices[5:10], 'anxietyscore'] = np.random.uniform(-5, -1, 5) # Low outliers

print("Dataset created with 300 participants")
print(f"Initial shape: {df.shape}")
print("\nFirst few rows:")
print(df.head())

=============================================================================

1. DATA CLEANING AND PREPARATION

=============================================================================

def clean_data(df):
"""Clean and prepare the dataset"""
df_clean = df.copy()

print("\n" + "="*60)
print("1. DATA CLEANING AND PREPARATION")
print("="*60)

# Remove duplicate participants
initialn = len(dfclean)
dfclean = dfclean.dropduplicates(subset=['participantid'], keep='first')
print(f"Removed {initialn - len(dfclean)} duplicates")

# Handle outliers in anxiety scores using IQR method
Q1 = dfclean['anxietyscore'].quantile(0.25)
Q3 = dfclean['anxietyscore'].quantile(0.75)
IQR = Q3 - Q1
lower_bound = Q1 - 1.5 * IQR
upper_bound = Q3 + 1.5 * IQR

# Clip outliers instead of removing (more conservative)
noutliersbefore = ((dfclean['anxietyscore'] < lower_bound) |
(dfclean['anxietyscore'] > upper_bound)).sum()

dfclean['anxietyscore'] = dfclean['anxietyscore'].clip(lower=0, upper=21) # GAD-7 max=21
dfclean['anxietyscore'] = np.clip(dfclean['anxietyscore'], lowerbound, upperbound)

print(f"Clipped {noutliersbefore} outliers in anxiety scores (IQR method)")
print(f"Anxiety score range after cleaning: {dfclean['anxietyscore'].min():.1f} - {dfclean['anxietyscore'].max():.1f}")

# Handle age outliers
dfclean = dfclean[(dfclean['age'] >= 16) & (dfclean['age'] <= 30)]
print(f"Filtered age outliers: kept participants aged 16-30 ({len(df_clean)} remaining)")

# Handle missing values
print("\nMissing value summary:")
missingsummary = dfclean.isnull().sum()
print(missingsummary[missingsummary > 0])

# Impute continuous variables with median, categorical with mode
for col in ['anxiety_score', 'gpa']:
if col in df_clean.columns:
medianval = dfclean[col].median()
nmissing = dfclean[col].isnull().sum()
dfclean[col] = dfclean[col].fillna(median_val)
print(f" Imputed {nmissing} missing {col} values with median ({medianval:.1f})")

for col in ['sleep_quality', 'gender', 'year']:
if col in df_clean.columns:
modeval = dfclean[col].mode()[0]
nmissing = dfclean[col].isnull().sum()
dfclean[col] = dfclean[col].fillna(mode_val)
print(f" Imputed {nmissing} missing {col} values with mode ({modeval})")

# Convert sleep_quality to numeric if it's categorical
if dfclean['sleepquality'].dtype == 'object':
dfclean['sleepquality'] = pd.tonumeric(dfclean['sleep_quality'], errors='coerce')

# Create anxiety categories for chi-square analysis
dfclean['anxietycategory'] = pd.cut(dfclean['anxietyscore'],
bins=[-1, 5, 10, 15, 21],
labels=['Minimal', 'Mild', 'Moderate', 'Severe'])

print(f"\nFinal cleaned dataset: {len(df_clean)} participants")
print(f"Missing values after cleaning: {df_clean.isnull().sum().sum()}")

return df_clean

Clean the data

dfclean = cleandata(df)

=============================================================================

2. DESCRIPTIVE STATISTICS

=============================================================================

def descriptive_stats(df):
"""Generate descriptive statistics"""
print("\n" + "="*60)
print("2. DESCRIPTIVE STATISTICS")
print("="*60)

# Overall descriptive statistics
print("\nAnxiety Score Descriptive Statistics:")
descstats = df[['anxietyscore', 'age', 'gpa', 'sleep_quality']].describe()
print(desc_stats.round(2))

# Categorical variable frequencies
print("\nCategorical Variable Frequencies:")
catvars = ['gender', 'year', 'majorstress', 'academicload', 'socialsupport', 'anxiety_category']
for var in cat_vars:
if var in df.columns:
print(f"\n{var}:")
print(df[var].valuecounts().sortindex() if df[var].dtype == 'object' else
df[var].valuecounts().sortindex())

Generate descriptive statistics

descriptivestats(dfclean)

=============================================================================

3. STATISTICAL TESTS WITH ASSUMPTION CHECKING

=============================================================================

def runstatisticaltests(df):
"""Run appropriate statistical tests with assumption checking"""
print("\n" + "="*60)
print("3. STATISTICAL TESTS WITH ASSUMPTION CHECKING")
print("="*60)

results = {}

# 3.1 Independent t-test: Gender differences in anxiety
print("\n3.1 INDEPENDENT T-TEST: Gender differences in anxiety")

# Prepare data (exclude Non-binary due to small sample)
df_ttest = df[df['gender'].isin(['Male', 'Female'])].copy()

maleanxiety = dfttest[dfttest['gender'] == 'Male']['anxietyscore']
femaleanxiety = dfttest[dfttest['gender'] == 'Female']['anxietyscore']

# Assumption 1: Normality (Shapiro-Wilk test)
print(f"\nAssumption Checking:")
shapiromale = shapiro(maleanxiety.dropna())
shapirofemale = shapiro(femaleanxiety.dropna())
print(f"Shapiro-Wilk normality test - Male: W={shapiromale.statistic:.3f}, p={shapiromale.pvalue:.3f}")
print(f"Shapiro-Wilk normality test - Female: W={