Reporting Variables
Grade Range
Teachers were classified by grade range (elementary, middle, and high) according to the information they provided about their teaching schedule. Elementary was defined as grades K–5 plus 6th grade self-contained; middle was defined as 6th grade non-self-contained and grades 7–8; high was defined as grades 9–12.
Percentage of Students in School Eligible for Free/Reduced-Price Lunch
Each teacher was classified into 1 of 4 categories based on the proportion of students in their school eligible for free/reduced-price lunch (FRL). The categories were defined as quartiles within groups of schools serving the same grades—e.g., schools with grades K–5, schools with grades 6–8 (see Table C-1).
Percentage of Students from Race/Ethnicity Groups Historically Underrepresented in STEM in Class
Each teacher was classified into 1 of 4 categories based on the proportion of students in their school identified as being from underrepresented minority (URM) groups in STEM (i.e., American Indian or Alaskan Native, Black or African American, Hispanic or Latino, Native Hawaiian or Other Pacific Islander, multi-racial). As this proportion is similar in schools regardless of grades served, the categories were defined as quartiles across all classes (see Table C-2).
Community Type
Each teacher was classified as belonging to 1 of 3 types of communities based on the location of their school:
- Urban: Central city;
- Suburban: Area surrounding a central city, but still located within the counties constituting a Metropolitan Statistical Area (MSA); or
- Rural: Area outside any MSA.
Political Leaning of County
Teachers were coded into 1 of 2 categories based on whether the majority of voters in their school’s county voted for the Democratic presidential candidate or Republican presidential candidate in the 2020 election.
Overview of Composites
To facilitate the reporting of large amounts of survey data, and because individual questionnaire items are potentially unreliable, HRI used factor analysis to identify survey questions that could be combined into “composites.” Each composite represents an important construct related to COVID in science education.
Each composite is calculated by summing the responses to the items associated with that composite and then dividing by the total points possible. In order for the composites to be on a 100-point scale, the lowest response option on each scale was set to 0 and the others were adjusted accordingly. For example, an item with a scale ranging from 1 to 4 was re-coded to have a scale of 0 to 3. By doing this, someone who marks the lowest point on every item in a composite receives a composite score of 0 rather than some positive number. It also assures that 50 is the true mid-point. The denominator for each composite is determined by computing the maximum possible sum of responses for a series of items and dividing by 100; e.g., a 9-item composite where each item is on a scale of 0–3 would have a denominator of 0.27. Composites values were not computed for participants who respond to fewer than two-thirds of the items that form the composite.
The composites were derived through a multi-stage process. As a first step, to test whether the items intended to target the same underlying construct indeed showed similar response patterns, an exploratory factor analysis was conducted on a subset of the data. (The complete dataset was split randomly into two subsets to allow for independent exploratory and confirmatory factor analyses.) Using Mplus version 8.1, several different factor solutions were produced and scree plots, eigenvalues, and factor patterns were examined. Based on item fit and conceptual coherence, preliminary composite definitions were created. Next, the preliminary composite definitions were applied to a different subset of the data and a confirmatory factor analysis was performed, again using Mplus. Mplus provides two fit indices to evaluate the model: the root mean square error of approximation (RMSEA) and the standardized root mean square residual (SRMR). The psychometric literature provides multiple criteria for judging acceptable model fit using this index, ranging from 0.05–0.10.9 The obtained values from final models are presented in the tables, allowing the reader to apply their preferred criteria for evaluating fit. Lastly, to further aid in the assessment of the composites, Cronbach’s coefficient alpha, a common measure of reliability, was calculated and is presented in the tables. An alpha of 0.6–0.8 is evidence of moderate reliability and a value over 0.8 is considered evidence of strong reliability.
Definitions of Composites
Composite definitions are presented below with the item numbers from the questionnaire, along with the reliability and fit indices.
Sources of Information About COVID
These composites estimate the extent to which teachers used various sources for their own information about COVID.
Teacher Decision Making
These composites estimate the extent to which various factors influenced whether teachers addressed COVID in their instruction.
Teacher Feelings
These composites estimate the range of feelings science teachers may have experienced prior to the pandemic, during the 2020-21 school year, and during the 2021-22 school year.
9Hu, L., & Bentler, P.M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55.