Table 3. Intervention Effects at the Individual and Organization Levels Adjusteda for Participant and State Characteristics in 12 States, Study of Evidence-Based Decision Making (EBDM),2014–2016.
| Dependent Variable | Intervention Effect Parameter Estimateb |
|||
|---|---|---|---|---|
| β (SE) | 95% Confidence Interval | t | P Valueb | |
| Individual | ||||
| EBDM skill gaps (10-item sum) | −5.56 (1.59) | −9.32 to −1.80 | −3.50 | .01 |
| Prioritization | −0.58 (0.20) | −1.07 to −0.09 | −2.89 | .03 |
| Adapting interventions | −0.69 (0.22) | −1.21 to −0.17 | −3.13 | .02 |
| Quantifying the issue | −0.59 (0.22) | −1.09 to −0.08 | −2.69 | .03 |
| Evaluation designs | −0.43 (0.24) | −1.00 to 0.14 | −1.79 | .11 |
| Quantitative evaluation | −0.23 (0.21) | −0.77 to 0.26 | −1.21 | .33 |
| Qualitative evaluation | −0.59 (0.24) | −1.19 to 0.02 | −2.48 | .05 |
| Economic evaluation | 0.18 (0.28) | −0.51 to 0.87 | 0.65 | .54 |
| Action planning | −0.35 (0.24) | 0.91 to 0.20 | −1.50 | .18 |
| Community assessment | −0.59 (0.22) | −1.11 to −0.06 | −2.65 | .03 |
| Communicating research to policy makers | −0.96 (0.28) | −1.63 to −0.29 | −3.41 | .01 |
| Use of research evidence for job tasks (6-item mean) | 0.12 (0.07) | −0.04 to 0.28 | 1.74 | .12 |
|
| ||||
| Organization | ||||
| Access to evidence and skilled staff (4- item factor) | 0.37 (0.14) | 0.02 to 0.72 | 2.73 | .04 |
| Program evaluation factor (3-item factor) | 0.03 (0.10) | −0.21 to 0.26 | 0.28 | .78 |
| Supervisory expectations for EBDM (3-item factor) | −0.06 (0.26) | −0.73 to 0.62 | −0.21 | .84 |
| Participatory decision making (3-item factor) | −0.06 (0.12) | −0.36 to 0.23 | −0.57 | .59 |
Abbreviation: SE, standard error.
a
Participant characteristics were sex, agency, job position, age group, having a public health master’s or doctoral degree, and having a master’s or doctoral degree in any field; state characteristics were accreditation status, chronic disease revenue from the Centers for Disease Control and Prevention to the state public health department, tertile of state population size, percentage of state population living in urban area, percentage of state population living in poverty, and state party control of the governorship, state senate, and state house.
b
Mixed analysis of covariance (ANCOVA) models with state as a random effect; parameter estimate P values are fixed solution 2-sided t tests within mixed ANCOVA.