critica de articulo, Apuntes de Ecología

¡Descarga critica de articulo y más Apuntes en PDF de Ecología solo en Docsity!

Running head: ARTICLE CRITIQUE Article Critique CSULB October 28, 2009 Article Critique Lee, O., Maerten-Rivera, J., Penfield, R.D., LeRoy, K. & Secada, W.G. (2008). Science

achievement of English language learners in urban elementary schools: Results of a first year professional development intervention. Journal of Research in Science Teaching, 45, 31-52. doi:10.1002/tea. Summary Introduction The No Child Left Behind Act of 2001 expects all students to achieve proficient levels of knowledge in core subject areas. Teachers of English language learners (ELL) face the added challenge of providing meaningful and accessible curricula while integrating English language and literacy development. This research study addresses ELL students’ low science achievement in the context of national standards and accountability in the 2006-2007 school year. Several studies have examined the influence of professional development interventions on students’ science achievement. Research suggests that hands-on and inquiry-based science lessons develop literacy as well as content knowledge. Research also indicates that students’ science achievement is positively correlated with the amount of teacher professional development. This study builds upon existing research by using a quasi-experimental design to assess students’ science achievement after the first-year implementation of a professional development intervention that focused on science achievement, literacy, and math skills. Specifically, the study addresses three research questions: (1) whether treatment group students show gains in science achievement, (2) whether gaps in science achievement change for ELL and low-literacy (retained) students in the treatment group, and (3) whether treatment group students perform differently compared with non-treatment group students on a statewide mathematics test, particularly on the measurement strand that is emphasized in the intervention. Participants

student responses on the pretest and posttest with an inter-rater agreement of 90%. Internal consistency reliability estimates were .60 for the pretest and .71 for the posttest. The reliability estimate for the pretest fell below the range generally considered acceptable, while the reliability estimates for the posttest were within acceptable range. The statewide mathematics test assessed five strands, including measurement, the strand emphasized in the intervention. The test consisted of multiple-choice, extended response, and short answer questions. Data Analysis To examine the effectiveness of the intervention, the treatment group was tested for science achievement. The original sample size was 1,134 students. However, only 818 students received both the pretest and posttest. Descriptive statistics provided a general idea of student gains from pretest to posttest. To analyze gaps in science achievement due to language and literacy levels, a hierarchical linear modeling (HLM) analysis was used to provide two levels of models for analysis, one of which is the mean-as-outcome model, using gender, ethnicity, exceptional student education (ESE), English to speakers of other languages (ESOL), and retention (low literacy) as independent variables. To compare the treatment and comparison groups, results from the statewide mathematics test were analyzed. Results were obtained from the school district database for 942 students in the treatment group and 966 students in the comparison group. Descriptive statistics provided the overall picture, and an HLM analysis examined whether differences between the groups were attributable to the treatment. Results and Conclusions

Descriptive statistics for the treatment group show a mean pretest score of 7.40 (SD = 3.36) and a mean posttest score of 14.34 (SD = 4.30). Mean scores for ESOL were lower than for non-ESOL at both pretest and posttest. Similarly, mean scores for retained were lower than for never retained at both pretest and posttest. Inferential statistics are presented, including variable, coefficient, standard error, t- statistic, degrees of freedom, and p-value. They specify an alpha of p < .01. For the science test, the overall mean gain score equaled 8.65 (p = .000). They use the p-value to conclude that the overall gain differed significantly. When analyzing the subgroups, they conclude that overall gains for ESOL and retention were not statistically significant (p = .329 for ESOL and p =. for retention). For the math test, results for the treatment group differed significantly from zero (p = .003). The coefficient differed significantly for ESOL (p = .004). The study did not find a significant difference between retained and never retained students. They identify three findings: (1) the treatment group showed a significant increase in science achievement, (2) there was no significant difference between ESOL and non-ESOL students, and there was no significant difference between retained and never retained students, and (3) the treatment group showed significantly higher scores on the math test than the comparison group, particularly in the measurement strand emphasized in the intervention. The researchers conclude that the intervention focusing on science inquiry was effective, particularly on the measurement strand of the math test. They offer implications of their study—the need for continued longitudinal research and more efforts to improve science learning simultaneously with English language and literacy development. Critique

Participants Treatment and non-treatment students were similar in terms of ethnicity, gender, and socioeconomic status with slight variations in the number of Hispanic and African-American students. The percentage of retained students was similar for both groups. The demographics of the teachers implementing the intervention are described. However, the demographics of the non-treatment teachers were not included. Although the method of selecting this sample was clearly described and the large sample size reduces potential error, ultimately, participation in the research was voluntary and therefore non-random. Although not explicitly stated as such, it appears that the authors used a convenience sampling method. Therefore, non-random sampling is a limitation of the study. For example, volunteering schools may have been more willing to implement and support curriculum change, which may have lead to improvement that cannot be attributed solely to the intervention. Instruments The contents of both instruments used in this study are described in detail. There was no data available from past uses of the first instrument, a science test. However, the test was developed using similar questions or actual questions from standardized tests used in science assessment. Internal consistency reliability estimates were acceptable for the posttest but were outside of the acceptable range for the pretest. No mention is made of validity. The second instrument, a statewide mathematics assessment, had no data given regarding its reliability or validity. Procedure

The design is appropriate to answer the questions of the study. First, the pretest and posttest administered to the treatment group assessed the effectiveness of the intervention. Given the context of national standards and high-stakes testing, it was appropriate to use a test to measure students’ science achievement. Second, the treatment and comparison groups were given the statewide mathematics test once towards the end of the school year. In a quasi- experimental design, it is appropriate to test students once to determine differences between the treatment and non-treatment groups. One limitation of the study is that pretesting the treatment group may have affected the internal validity of the results—that is, the improvement in scores may be due to student familiarity with the instrument. Data Analysis There are several strengths in this section. Sample size was adequate, even when broken down into subcategories of gender, ethnicity, ESE, ESOL and retention. Descriptive analysis seemed like a proper procedure for comparison of science pretest and posttest for the treatment group. The HLM analysis was an appropriate program to use since it is a more advanced form of single linear regression. The HLM allowed researchers to analyze data on multiple levels, such as the five independent variables, whereas single linear analysis would not have allowed them to do that. The only issue with using HLM analysis is that it does not appear to be a well-known program, so confirming the research would be difficult to do and that may raise suspicion. Another issue is that some of the data was lost due to mortality—in the treatment group, two teachers failed to administer posttests to their students. However, the mortality rate (28%) is not unusual for field studies of this nature. Results and Conclusions