Table 4

Advertising Effectiveness Regression Results

Coefficient	Name	Estimate	Standard Error	t-statistic	p
Model excluding clutter in decay
α	Constant	−0.44	0.00383	−114.81	.000
β	Linear predictor—no advertising (LP)	0.156	0.00172	90.37	.000
ε	Square root of advertising spending (A)	−0.0016	0.000068	−23.53	.000
σ	Clutter × square root of advertising spending (CA)	0.000058	0.00000294	19.7	.000
τ	Sales percentage difference (SPD)	−3.08	0.0794	−38.76	.000
ρ	Familiarity × square root of advertising spending (FA)	−0.0001	0.000046	−2.17	.038
λ	Constant for advertising	2.92	0	45.38
Model including clutter in decay
α	Constant	−0.438	0.00369	−118.73	.000
β	Linear predictor—no advertising (LP)	−0.00163	0.0000637	−25.59	.000
ε	Square root of advertising spending (A)	−0.00163	0.0000637	−25.59	.000
σ	Clutter × square root of advertising spending (CA)	0.0000591	0.00000325	18.2	.000
τ	Sales percentage difference (SPD)	−3.17	0.0782	−40.5
ρ	Familiarity × square root of advertising spending (FA)	0.0000917	0.0000434	−2.11	.043
λ	Constant for advertising	2.95	0	45.58
Final model including Herfindahl–Hirschman index
α	Constant	−0.439	0.003826	−114.92	.000
β	Linear predictor—no advertising (LP)	−0.1558	0.001723	−90.42	.000
ε	Square root of advertising spending (A)	−0.00163	0.000071	−22.96	.000
σ	Clutter × square root of advertising spending (CA)	0.000058	0.000002935	19.76	.000
τ	Sales percentage difference (SPD)	−3.07	0.07932	−38.71	.000
ρ	Familiarity × square root of advertising spending (FA)	−0.00011	0.000047	−2.34	.026
η	Herfindahl–Hirschman index × square root of advertising spending (HA)	0.000044	0.000032	1.38	.155
λ	Constant for advertising	2.914	0.06415	45.42	.000

Clutter excluded: pseudo R² = .922; Clutter included: pseudo R² = .922; Final model: pseudo R² = .922