Skip to content
Rigor, made runnable.

Methods

Every analysis Explonia runs in your browser, explained in plain English — what it is, when to use it, the thresholds reviewers expect, and how to report it in APA. Then run it on your own data, free to start.

Measurement & validity

Build, validate and report your measurement model — factors, reliability and discriminant validity.

confirmatory factor analysis

Confirmatory Factor Analysis

Test a measurement model you already have — loadings, fit, AVE and CR — in your browser, then export your data to SPSS/.sav or AMOS, and the model syntax to R (lavaan).

Open method
exploratory factor analysis

Exploratory Factor Analysis

Exploratory factor analysis (EFA) finds the structure behind a set of items when you do not know it in advance: how many factors there are and which items belong to each. Common guidelines are a KMO of .60 or higher (.80 is good), a significant Bartlett's test (p < .05), and loadings of at least .40 to .50 for interpretation. You can run EFA online in Explonia, in your browser, on data you collect or upload.

Open method
efa vs cfa difference

EFA vs CFA

EFA vs CFA — when to explore and when to confirm — with a decision table, and both run in Explonia.

Open method
cronbach alpha reliability

Reliability (alpha, omega, CR, AVE)

Cronbach's alpha (α), McDonald's omega (ω) and composite reliability (CR) all estimate internal consistency, and .70 is the commonly used cutoff for each; AVE measures convergent validity, with .50 as the commonly used cutoff. Omega and CR do not assume equal item loadings, so they are often preferred over alpha.

Open method
htmt discriminant validity

HTMT Discriminant Validity

HTMT (the heterotrait-monotrait ratio of correlations) assesses discriminant validity between reflective constructs. This guide explains what HTMT measures, when to use it, its assumptions and how to interpret the result. For choosing between the .85 and .90 criteria, see the HTMT threshold guide (.85 vs .90).

Open method common method bias

Common Method Bias

Common method bias (CMB), or common method variance, is artificial covariance that arises when the same respondents answer all your questions with the same method at the same time. It can inflate or deflate the correlations between constructs. Harman's single-factor test flags a problem when one factor explains 50% or more of the variance, but the test is weak and should not be your only evidence. Stronger checks model the method variance directly: a common latent factor (ULMC) or, preferably, a measured marker variable.

Open method
Structural models

Test relationships between constructs and compare groups with SEM.

cb-sem covariance based sem

Covariance-Based SEM

Covariance-based SEM (CB-SEM) tests a full theoretical model, both the measurement part and the paths between constructs, by reproducing the observed covariance matrix. It gives global fit indices, so you can confirm or reject a theory. Commonly used cutoffs are CFI and TLI of .95 or higher, RMSEA of .06 or lower and SRMR of .08 or lower (Hu & Bentler, 1999). Looser lines are often treated as acceptable: CFI/TLI of .90 (Bentler & Bonett, 1980), RMSEA of .08 (Browne & Cudeck, 1993) and SRMR of .10 (Schermelleh-Engel et al., 2003). CB-SEM suits mature theory with reflective constructs; for prediction or formative constructs, consider PLS-SEM.

Open method
pls-sem analysis

PLS-SEM

Run PLS-SEM on your own data — path coefficients, bootstrap confidence intervals and R² — with no SmartPLS license needed.

Open method
mediation analysis

Mediation Analysis

Test indirect effects with bootstrapped confidence intervals — and get the APA sentence drafted for you to edit.

Open method
measurement invariance

Measurement Invariance

Measurement invariance tests whether a scale means the same thing across groups, in three nested steps: configural (same pattern of loadings), metric (equal loadings) and scalar (equal loadings and intercepts). You need scalar invariance before you compare latent means across groups. Each step adds constraints, and invariance is supported when fit does not get meaningfully worse: a change in CFI (ΔCFI) of .01 or less (Cheung & Rensvold, 2002), often read together with ΔRMSEA of .015 or less (Chen, 2007).

Open method
Design & comparison

Design studies and compare conditions, segments or feature trade-offs.

Classical analyses

Describe your data and test simple relationships: summaries, cross-tabs, regression and repeated measures.

Planning & diagnostics

Plan the sample you need before you collect data, and check that the assumptions behind your model hold.

Run your whole study in one place

Design it, collect it, and get publication-grade analysis with a draft APA paragraph you can edit, free to start.

Start free