Factor Analysis Services

EFA vs CFA

Which Factor Analysis Do You Need?

Understanding the difference is critical choosing the wrong method invalidates your validity evidence.

Exploratory Factor Analysis

Used when the factor structure is unknown. Lets the data determine how items cluster into latent constructs.

  • No prior theory about factor structure required
  • Identifies natural item groupings from data
  • Determines optimal number of factors (scree plot, eigenvalues, parallel analysis)
  • Rotation methods: Varimax (orthogonal), Promax/Oblimin (oblique)
  • Reports factor loadings, communalities, and explained variance
  • Used in scale development & pilot studies

Best for: New questionnaire development, item reduction, initial construct exploration

Confirmatory Factor Analysis

Used when the`` factor structure is theoretically specified. Tests whether data fits a pre-defined measurement model.

  • Tests a theoretically-driven factor model
  • Reports model fit: CFI, TLI, RMSEA, SRMR, χ²
  • Provides standardised factor loadings & AVE
  • Assesses convergent & discriminant validity
  • Modification indices for model improvement
  • Required before SEM / path modelling

Best for: Validating existing scales, multi-group comparisons, pre-SEM measurement model

Software & Methods

Tools & Techniques We Use

SPSS Exploratory FA

IBM SPSS used for EFA with Principal Axis Factoring, scree plots, Varimax/Promax rotation, and parallel analysis.

IBM Amos CFA & SEM

Full Amos path diagrams with ML estimation, modification indices, and complete fit index reporting.

R lavaan & psych

Open-source CFA and EFA using R's lavaan and psych packages fully reproducible with syntax provided.

Mplus Advanced CFA

Mplus for ordinal data (WLSMV), ESEM, bi-factor models, and mixture factor analysis.

Higher-Order & Bi-Factor Models

Second-order CFA and bi-factor models tested when a general factor underlies multiple specific factors.

Multi-Group & Invariance Testing

Configural, metric, scalar, and strict invariance testing across groups with ΔCFI-based evaluation.

CFA Fit Indices

Acceptable Model Fit Thresholds We Report

Every CFA we deliver includes all five key fit indices verified against published cutoffs.

χ² / df Ratio

Acceptable: < 3.0
Good: < 2.0

RMSEA

Acceptable: < .08
Good: < .05

CFI & TLI

Acceptable: > .90
Good: > .95

SRMR

Acceptable: < .08
Good: < .05

AVE & CR

AVE > .50
CR > .70

Our Benchmarks for Reliable Outcomes

KMO > 0.6

Ensuring data factorability through Kaiser-Meyer-Olkin measure.

Loadings > 0.5

Suppressing small coefficients and mitigating cross-loadings.

CFI > 0.90

Pursuing robust and widely accepted Comparative Fit Indices in CFA.

Cronbach's α

Validating internal consistency reliability for newly emerged factors.