Impacting secondary students’ STEM knowledge through collaborative STEM teacher partnerships

Abstract

Integrated Science, Technology, Engineering, and Mathematics (STEM) teaching provides an opportunity for students to learn STEM knowledge across two or more domains. The following study presents students’ STEM content knowledge achievement after learning an integrated STEM unit taught by science and engineering technology teachers. After completing a two-week teacher professional development workshop, science and engineering technology teachers implemented an exemplar STEM unit called D-BAIT. The integrated STEM unit included entomology, biology, biomimicry, physics, and engineering design content. The researchers constructed a STEM knowledge multiple-choice pre/post-test assessment to assess students’ understanding of these concepts. This study employed a quasi-experimental nonequivalent comparison group design and collected a total of 1,345 pre/post-test assessments. The data were analyzed through the independent samples t-test. The results indicate that the integrated STEM unit implemented by teacher collaboration increased students’ overall STEM content knowledge. The comparison between science and engineering students’ knowledge gain showed that the integrated STEM unit significantly impacts students’ content knowledge. The comparisons between domain and cross-domain knowledge in science and engineering content found no significant differences; however, the mean score gain in cross-domain was higher than within the subject domain. The results of this study indicate that students can learn domain content outside of their course of study.

Appendix

A. Item Analysis.

Instrument Item Difficulty Index. The item difficulty test was employed to test if the individual items have an appropriate level of difficulty. The instrument item difficulty, denoted by p, is measured by the proportion of participants responding to each item correctly. Item difficulty is calculated by:

$$p=\frac{{X}_{i}}{N}, where$$

${\text{X}}_{\text{i}}$ = Number of students responding correctly to item i

N = number of students taking the assessment.

The result of item analysis shows the difficulty ranges from 0.10 to 0.78, and confirmed that instrument items 3, 4, and 24 results were computed to around 0.10 difficulty, meaning 90% of students did not correctly answer those items.

• 1. Item Discrimination. Instrument item discrimination analysis detects whether each instrument item is efficiently constructed by comparing answer rates between students ranked over 66% and below 33% on the test. Item discrimination was identified by the following formula,

$$\text{d}=\frac{{\text{U}}_{\text{i}}}{{\text{n}}_{\text{i}\text{U}}}-\frac{{\text{L}}_{\text{i}}}{{\text{n}}_{\text{i}\text{L}}},\text{w}\text{h}\text{e}\text{r}\text{e}$$

${U}_{i}$=number of students who have total scores in the upper range and who also have item i correct,

${L}_{i}$=number of students who have total scores in the lower range and who also have item i correct,

${n}_{iU}$=number of students who have total scores in the upper range of total test scores, and

${n}_{iL}$=number of students who have total scores in the lower range of total test scores.

The item discrimination index has from − 1.0 to 1.0. A negative discrimination index means that lower scoring students performed better on the item than higher scoring students. A low discrimination index (< 0.3) indicates an item was equally difficult for both groups. A high index value means the item discriminates well between low and high scorers. An index value of 0.3 and above is good and 0.6 and above is very good (Ebel, 1973; Ovwigho, 2014). The analysis for item discrimination indicated that item 3 and 24 had negative indexes, which imply that lower ranked students were more likely to answer the question correctly than higher ranked students. Also, items 4 and 23 were very low on the discrimination scale in the “Poor” index score range, another justification to remove these items from the instrument.

• 2. Internal consistency and reliability. To assess internal consistency of the test instrument, the researchers calculated Cronbach’s Alpha using SPSS 23 software. The overall Cronbach’s Alpha was 0.69 which is a marginal range of score. The statistical item analysis indicated that if item 3, 4, 23, and 24 were deleted, the overall Cronbach’s Alpha score would be over 0.7, which is in the acceptable index range. Lastly, test reliability was assessed through split-half reliability using the adjusted Spearman-Brown prophecy formula (Brown, 1910; Spearman, 1910; Remmers & Ewart, 1941). The reliability score was computed in R statistics software and resulted in a reliable score range of 0.876.

$$reliability=(2\times {r}_{half-test})/(1+{r}_{half-test})$$

The results of item analysis are shown in Table a below.

Tab a The results of item analysis