Full Walkthrough
A complete example of a QuantAI analysis — from raw CSV upload through smart test recommendations, assumption checking, APA-formatted results, and the AI-drafted Results paragraph. The dataset is a real anxiety intervention study with 247 participants.
Upload your dataset
The researcher exports their Qualtrics survey and drags the CSV directly into QuantAI. No reformatting, no manual variable coding — QuantAI parses the file, detects the platform, strips metadata rows, and identifies every column's type automatically.
anxiety_intervention_study.csv
247 rows · 18 columns · 42 KB · parsed in 0.31s
Detected Variables
QuantAI recommends the right test
After upload, QuantAI reads the variable types and instantly presents plain-language research questions the data can answer — no need to know statistical method names in advance. One click selects the test and pre-fills the variable assignments. Experienced researchers can skip straight to any test via the wizard or full test list.
anxiety_intervention_study.csv
247 rows · 18 columns detected
What QuantAI found in your data
Research questions your data can answer
How do the two treatment_group groups compare on anxiety_post?
Independent t-test — compares anxiety_post between control and treatment
Which variables independently predict anxiety_post?
Multiple regression — shows each factor's unique effect after controlling for others
What relationships exist among your numeric variables?
Spearman correlation matrix across 4 continuous variables
What are the summary statistics for my variables?
Means, SDs, skewness, kurtosis — standard starting point
Assign Variables
Dependent variable
Continuous numeric outcome
Selected: anxiety_post
Grouping variable
Categorical variable with exactly 2 groups
Selected: treatment_group
Results
QuantAI returns the full results panel in seconds. Every element below is real output — the assumption flags ran before the test, the table follows APA 7th edition formatting exactly, and the narrative was written by Claude using only the computed statistics.
Independent Samples t-Test
Dataset: anxiety_intervention_study.csv · 247 participants · Listwise deletion: 0 cases removed
Assumption Checks
Missing Data
No missing data detected across analysis variables (anxiety_post, treatment_group).
Normality — Control group (Shapiro-Wilk)
W = 0.974, p = .341. Normality assumption met for the control group.
Normality — Treatment group (Shapiro-Wilk)
W = 0.981, p = .512. Normality assumption met for the treatment group.
Homogeneity of Variance (Levene's test)
F = 4.821, p = .029. Unequal variances detected between groups.
Recommendation: Welch's correction has been applied automatically. Report this in your Methods section: "Levene's test indicated unequal variances (F = 4.82, p = .029); therefore, a Welch's t-test was conducted."
Distribution Plot
Figure 1. Distribution of post-intervention anxiety scores by group.
APA Table Output
Table 1
Independent Samples t-Test Results for Post-Intervention Anxiety Scores
| Group | n | M | SD |
|---|---|---|---|
| Control | 124 | 3.42 | 0.81 |
| Treatment | 123 | 4.17 | 0.74 |
AI-Drafted Results Paragraph
APA 7th edition · 127 words"An independent samples t-test was conducted to examine whether post-intervention anxiety scores differed between the control and treatment groups. Levene's test for equality of variances indicated unequal variances (F = 4.82, p = .029); therefore, a Welch'st-test was conducted. Results revealed a statistically significant difference between the control group (M = 3.42, SD = 0.81, n = 124) and the treatment group (M = 4.17, SD = 0.74, n = 123), t(241.8) = −7.39,p < .001, Cohen's d = 0.98. The large effect size indicates that the treatment group reported substantially higher post-intervention anxiety scores compared to the control group. These findings should be interpreted in light of the unequal group variances noted above."
Drafted by Claude · Reviewed by you before submission
This paragraph is a well-informed first draft. Review it for accuracy, adjust to your specific research context, and take ownership of the language before submitting to a journal or committee.
Verify independently
Every results page includes step-by-step instructions to reproduce the analysis in R or SPSS. You don't have to take QuantAI's word for it — the same computation is one command away in any standard statistical tool. Results match to reported decimal precision.
QuantAI uses scipy.stats and statsmodels for all computations. The R code below runs the identical analysis. Results should match to the decimal precision shown in the table above.
library(readr)
library(effsize)
df <- read_csv("anxiety_intervention_study.csv")
# Welch's t-test (default in R — same as QuantAI when Levene's p ≤ .05)
result <- t.test(anxiety_post ~ treatment_group, data = df)
print(result)
# Expected output:
# t = -7.385, df = 241.8, p-value < 2.2e-16
# 95% CI: [-0.951, -0.549]
# Cohen's d (effect size)
cohen.d(df$anxiety_post ~ df$treatment_group)
# Expected: d = 0.981 (large)If your independently-computed results differ from QuantAI, email hello@quantai.study with your dataset and we will investigate within 24 hours.
Export to Word
Scholar Pass and Doctoral Pro users click one button to download a formatted .docx file. The document contains the APA table and the Results paragraph — formatted and ready to paste directly into Chapter 4 of a dissertation or a manuscript Methods/Results section.
What the .docx contains
APA-formatted table
Properly ruled, titled, and noted per APA 7th edition §7.13. Paste directly into your manuscript.
Results paragraph
The AI-drafted paragraph with full statistical notation. Edit as needed, then paste into your Results section.
Assumption report
A plain-language summary of every assumption check with flag status and recommendations. Include in your Appendix or Methods.
Standard formatting
12pt Times New Roman, 1-inch margins, double-spaced — dissertation-ready defaults.
Analysis complete
Independent t-Test · anxiety_post · 247 participants
Export Options