Back to What do extra predictors add?

Saved evidence · NLSS 2.0.1

Saved run 20260918T102612-25f365e6489

The saved calculation record complements the authored report. It records the numerical work; it does not establish that a method or interpretation is appropriate. Original NLSS figure numbering is preserved.

Assumption Checks (Regression)

Analysis

  • Analysis: regression
  • Dependent variable: exhaustion
  • Blocks: age;workload,resources
  • Alpha: 0.05
  • Rounding digits: 2
  • Normality test: shapiro
  • Vif Warn: 5
  • Vif High: 10
  • Outlier Z: 3
  • Cook Multiplier: 4
  • Max Shapiro N: 5000
  • Linearity: Yes
  • Homoscedasticity: Yes
  • VIF: Yes
  • Durbin Watson: No
  • Outliers: Yes
  • Influence: Yes

Table 1

Regression Assumption Checks

Model Assumption Test Target Statistic df p Value Status Screen n
Block 1 Normality Shapiro-Wilk Residuals .99 = .324 available ok 180
Block 1 Linearity Quadratic added-term F age .01 1.00, 177.00 = .937 available ok 180
Block 1 Homoscedasticity Breusch-Pagan Residuals .13 1.00 = .721 available ok 180
Block 1 Outliers Std. residuals Residuals 2.84 available ok 180
Block 1 Influence Cook's distance Residuals .05 available flag 180
Block 1 Multicollinearity VIF age 1.00 available ok 180
Block 2 Normality Shapiro-Wilk Residuals .99 = .169 available ok 180
Block 2 Linearity Quadratic added-term F age .04 1.00, 175.00 = .834 available ok 180
Block 2 Linearity Quadratic added-term F workload .13 1.00, 175.00 = .717 available ok 180
Block 2 Linearity Quadratic added-term F resources .06 1.00, 175.00 = .811 available ok 180
Block 2 Homoscedasticity Breusch-Pagan Residuals .19 3.00 = .980 available ok 180
Block 2 Outliers Std. residuals Residuals 2.92 available ok 180
Block 2 Influence Cook's distance Residuals .04 available flag 180
Block 2 Multicollinearity VIF age 1.01 available ok 180
Block 2 Multicollinearity VIF workload 1.07 available ok 180
Block 2 Multicollinearity VIF resources 1.06 available ok 180

Note. Normality assessed with Shapiro-Wilk (alpha = 0.05). Homoscedasticity assessed with Breusch-Pagan. VIF thresholds: 5 (moderate), 10 (high). Outliers flagged at |std resid| > 3. Cook's D threshold = 4/n. Availability is explicit. Diagnostic flags are screens, not a blanket analysis approval or rejection. Linearity uses separate centered quadratic added-term F tests. Hierarchical blocks use common complete cases. VIF is per design column, not term GVIF.

Narrative

Block 1: Normality (Shapiro-Wilk) for Residuals, statistic = .99, p = .324, n = 180. Status: available.

Block 1: Linearity (Quadratic added-term F) for age, df = 1.00, 177.00, statistic = .01, p = .937, n = 180. Status: available. One centered quadratic term added to the fitted block; a limited curvature screen, not proof of linearity.

Block 1: Homoscedasticity (Breusch-Pagan) for Residuals, df = 1.00, statistic = .13, p = .721, n = 180. Status: available. Koenker studentized n R²; auxiliary degrees of freedom use model-matrix rank minus the intercept.

Block 1: Outliers (Std. residuals) for Residuals, value = 2.84, n = 180. Status: available. |std resid| > 3: 0

Block 1: Influence (Cook's distance) for Residuals, value = .05, n = 180. Status: available. Cook's D > 0.0222222222222222: 8

Block 1: Multicollinearity (VIF) for age, value = 1.00, n = 180. Status: available. Design-column VIF; factor dummy columns are not term-level GVIF.

Block 2: Normality (Shapiro-Wilk) for Residuals, statistic = .99, p = .169, n = 180. Status: available.

Block 2: Linearity (Quadratic added-term F) for age, df = 1.00, 175.00, statistic = .04, p = .834, n = 180. Status: available. One centered quadratic term added to the fitted block; a limited curvature screen, not proof of linearity.

Block 2: Linearity (Quadratic added-term F) for workload, df = 1.00, 175.00, statistic = .13, p = .717, n = 180. Status: available. One centered quadratic term added to the fitted block; a limited curvature screen, not proof of linearity.

Block 2: Linearity (Quadratic added-term F) for resources, df = 1.00, 175.00, statistic = .06, p = .811, n = 180. Status: available. One centered quadratic term added to the fitted block; a limited curvature screen, not proof of linearity.

Block 2: Homoscedasticity (Breusch-Pagan) for Residuals, df = 3.00, statistic = .19, p = .980, n = 180. Status: available. Koenker studentized n R²; auxiliary degrees of freedom use model-matrix rank minus the intercept.

Block 2: Outliers (Std. residuals) for Residuals, value = 2.92, n = 180. Status: available. |std resid| > 3: 0

Block 2: Influence (Cook's distance) for Residuals, value = .04, n = 180. Status: available. Cook's D > 0.0222222222222222: 11

Block 2: Multicollinearity (VIF) for age, value = 1.01, n = 180. Status: available. Design-column VIF; factor dummy columns are not term-level GVIF.

Block 2: Multicollinearity (VIF) for workload, value = 1.07, n = 180. Status: available. Design-column VIF; factor dummy columns are not term-level GVIF.

Block 2: Multicollinearity (VIF) for resources, value = 1.06, n = 180. Status: available. Design-column VIF; factor dummy columns are not term-level GVIF.