Read-only sample study. A complete walkthrough — from problem statement to a validated report — using first_year_attrition.csv (400 rows × 13 cols). Sign in to run your own.

Sample study · survival design

First-Year Attrition Among New Hires

Research information

Statement of the problem

This people-analytics study follows 400 new hires from a 30-day engagement survey through 12-month exit outcomes, using time-to-event (survival) methods on months_to_exit to examine how engagement, function, and manager tenure relate to who leaves and when.

Objectives

To describe engagement_score, manager_tenure_months, and demographics across the four functions.
To assess the reliability of the engagement items engagement_1, engagement_2, engagement_3, engagement_4, engagement_5, and engagement_6.
To determine whether engagement_score differs across the four functions.
To determine the relationship between engagement_score and manager_tenure_months.
To estimate Kaplan-Meier survival curves for months_to_exit by function and compare using the log-rank test.
To model months_to_exit using Cox regression on engagement_score and manager_tenure_months.

Dataset

first_year_attrition.csv400 rows × 13 cols

Column	Kind	Distinct
months_to_exit	numeric	96
event_indicator	numeric	2
employee_id	text	400
function	categorical	4
role_level	categorical	3
manager_tenure_months	numeric	68
engagement_1	numeric	6
engagement_2	numeric	6
engagement_3	numeric	6
engagement_4	numeric	6
engagement_5	numeric	7
engagement_6	numeric	6
engagement_score	numeric	29

Feasibility & estimateFree

✓ Feasibledesign: survival · via heuristic

Detected a survival design with 12 analyzable variable(s) and 6 analytical question(s). Recommended 12 statistical test(s). Estimated cost: 94 credits.

Recommended analysis

To describe engagement_score, manager_tenure_months, and demographics across the four functions.

Descriptive Statistics(manager_tenure_months, engagement_score)

Frequency & Percentage Distribution(function)

To assess the reliability of the engagement items engagement_1, engagement_2, engagement_3, engagement_4, engagement_5, and engagement_6.

Cronbach's Alpha(engagement_1, engagement_2, engagement_3, engagement_4, engagement_5, engagement_6)

McDonald's Omega(engagement_1, engagement_2, engagement_3, engagement_4, engagement_5, engagement_6)

To determine whether engagement_score differs across the four functions.

One-Way ANOVA(engagement_score, function)

Tukey HSD Post-Hoc(engagement_score, function)

Games–Howell Post-Hoc(engagement_score, function)

Dunn's Test (post-hoc)(engagement_score, function)

To determine the relationship between engagement_score and manager_tenure_months.

Pearson Correlation(manager_tenure_months, engagement_score)

To estimate Kaplan-Meier survival curves for months_to_exit by function and compare using the log-rank test.

Kaplan–Meier Survival Estimate(months_to_exit, event_indicator, function)

Log-Rank Test(months_to_exit, event_indicator, function)

Cox Proportional Hazards Regression(months_to_exit, event_indicator, function)

To model months_to_exit using Cox regression on engagement_score and manager_tenure_months.

Kaplan–Meier Survival Estimate(months_to_exit, event_indicator)

Cox Proportional Hazards Regression(months_to_exit, event_indicator, manager_tenure_months, engagement_score)

Dataset Complexity18

Statistical Tests41

AI Interpretation18

Visualizations12

Report Generation5

Total to run94 credits

Estimated runtime: ~2-3 minutes

ResultsComputed

12 tests94 credits

Descriptive Statistics (manager_tenure_months, engagement_score)

n	max	min	std	mean	mode	range	median	variable	variance
400	71	3	20.020	36.828	25	68	37	manager_tenure_months	400.815
400	7	2.170	0.963	4.990	5.170	4.830	5	engagement_score	0.927

Frequency & Percentage Distribution (function)

function · Engineering43.0%

function · Sales23.0%

function · Operations17.3%

function · Marketing16.8%

Cronbach's Alpha (engagement_1, engagement_2, engagement_3, engagement_4, engagement_5, engagement_6)

α0.939N400

Internal consistency is excellent (alpha = 0.939).

McDonald's Omega (engagement_1, engagement_2, engagement_3, engagement_4, engagement_5, engagement_6)

N400

item	loading
engagement_1	0.841
engagement_2	0.847
engagement_3	0.861
engagement_4	0.847
engagement_5	0.839
engagement_6	0.855

Internal consistency (McDonald's omega) is excellent (omega = 0.939).

One-Way ANOVA (engagement_score, function)

significant

p< .001

Eta squared0.141

negligible

n	std	mean	group
172	0.861	5.310	Engineering
67	0.931	5.209	Marketing
69	0.949	4.648	Operations
92	0.893	4.489	Sales

The result is statistically significant (p = 0.0000 < 0.05).

Tukey HSD Post-Hoc (engagement_score, function)

pair	p value	mean diff	significant
Engineering vs Marketing	0.862	0.101	false
Engineering vs Operations	0.000	0.662	true
Engineering vs Sales	0.000	0.821	true
Marketing vs Operations	0.002	0.561	true
Marketing vs Sales	0.000	0.720	true
Operations vs Sales	0.680	0.159	false

Pairwise comparisons with family-wise error control (Tukey HSD).

Spearman Rank Correlation (manager_tenure_months, engagement_score)

n.s.

r-.046df398R².002N400p.357

Spearman r-0.046

negligible

The result is not statistically significant (p = 0.3566 >= 0.05).

Spearman Rank Correlation was used instead of the Pearson Correlation because the latter's assumption(s) were not met (Normality of manager_tenure_months (Shapiro-Wilk), Normality of engagement_score (Shapiro-Wilk)).

Kaplan–Meier Survival Estimate (months_to_exit, event_indicator, function)

N400

n	group	events	censored	median survival	restricted mean
172	Engineering	114	58	5.800	0.595
67	Marketing	47	20	5.900	0.626
69	Operations	58	11	2.500	0.472
92	Sales	78	14	1.500	0.460

Observed 297 event(s) in 400 subject(s).

Log-Rank Test (months_to_exit, event_indicator, function)

significant

χ²29.64df3N400p< .001

n	group	events
172	Engineering	114
67	Marketing	47
69	Operations	58
92	Sales	78

The result is statistically significant (p = 0.0000 < 0.05).

Cox Proportional Hazards Regression (months_to_exit, event_indicator, function)

significant

N400p< .001

Concordance0.586

medium

coef	p value	hr ci low	predictor	hr ci high	hazard ratio
0.052	0.765	0.750	function_Marketing	1.479	1.053
0.592	0.000	1.316	function_Operations	2.481	1.807
0.680	0.000	1.477	function_Sales	2.637	1.973

The result is statistically significant (p = 0.0000 < 0.05).

Kaplan–Meier Survival Estimate (months_to_exit, event_indicator)

N400

n	group	events	censored	median survival	restricted mean
400	overall	297	103	4	0.487

Observed 297 event(s) in 400 subject(s).

Cox Proportional Hazards Regression (months_to_exit, event_indicator, manager_tenure_months, engagement_score)

significant

N400p< .001

Concordance0.737

medium

coef	p value	hr ci low	predictor	hr ci high	hazard ratio
-0.003	0.361	0.991	manager_tenure_months	1.003	0.997
-0.967	0.000	0.332	engagement_score	0.435	0.380

The result is statistically significant (p = 0.0000 < 0.05).

Report

Publication-ready report

Tests used · Results · Discussion — generated from the computed statistics.

Every claim checked · 45 numbers verified · 2 findings cross-checked · prose: deterministic

Open report ↗Download DOCX Download PDF

Reproducible code

Download the exact Jupyter notebook or Python script that reproduces these results with widely-used libraries (scipy, statsmodels, lifelines).

Download Jupyter notebook Download Python script

Run your own study

Bring your problem statement, objectives, and a dataset — thericerca does the rest.

Get started →