The role of media in influencing students’ STEM career interest

International Journal of STEM Education

Table 4 Coefficients for the structural components of the model

	β	SE	Sig
STEM career interest ON
READ	− 0.042	0.015
TV/GAME	0.030	0.017
APP	0.006	0.016
IDENTITY	0.519	0.013	***
SELFDEV	0.526	0.026	***
WORKLIFE	0.116	0.018	***
MATERIAL	− 0.399	0.024	***
INTERPERSO	− 0.709	0.022	***
Male	0.122	0.013	***
Parent edu	− 0.017	0.013
Beginning of HS STEM career interest	0.618	0.012	***
IDENTITY ON
READ	0.031	0.012	**
TV/GAME	0.221	0.013	***
APP	0.069	0.012	***
Male	0.111	0.01	***
Parent edu	0.081	0.01	***
Beginning of HS STEM career interest	0.522	0.01	***
SELFDEV ON
READ	0.001	0.012
TV/GAME	0.024	0.015	***
APP	0.067	0.014	***
Male	− 0.029	0.011	***
Parent edu	− 0.035	0.011	***
Beginning of HS STEM career interest	0.016	0.011
WORKLIFE ON
READ	− 0.046	0.013	***
TV/GAME	− 0.022	0.015
APP	0.048	0.014	***
Male	− 0.049	0.011	***
Parent edu	− 0.067	0.011	***
Beginning of HS STEM career interest	− 0.063	0.011	***
MATERIAL ON
READ	− 0.052	0.014	***
TV/GAME	− 0.024	0.016	***
APP	0.084	0.015	***
Male	0.083	0.012	***
Parent edu	− 0.041	0.012	***
Beginning of HS STEM career interest	− 0.040	0.012	***
INTERPER ON
READ	0.028	0.014	**
TV/GAME	− 0.026	0.017	*
APP	0.050	0.016	***
Male	− 0.256	0.012	***
Parent edu	− 0.033	0.012	***
Beginning of HS STEM career interest	− 0.274	0.012	***

When the outcome of the prediction is STEM career interest (a binary variable), the coefficients are probit coefficients. Other coefficients are linear regression coefficients. The measurement components of the model are not shown in this model, they are nearly identical to the coefficients shown in Tables 2 and 3. Capitalized variables are latent variables. *p < 0.05; **p < 0.01; ***p < 0.001 after Bonferroni p value adjustment due to multiple comparisons