For educational leaders, data-driven decisions are often focused on comparisons, such as comparing methods of instruction, methods of delivery, instructional focus, teacher performance, professional staff, or even schools. When two groups are compared, the statistic that is most useful in data-driven decisions is called a students’ sample t-test. The purpose of a students’ samples t-test, is to determine if there is a statistically significant difference between the mean scores of two groups.
A paired samples t-test, a variation of a students’ samples t-test, is when you have two sets of scores on the same individuals or when the individuals are tested twice (pre-test/post-test). Ultimately, the two groups of data are really the same individual with two scores. The OnTarget STAAR Impact Report – Year to Year by Student Group, utilizes the paired samples t-test which can be of great benefit to educational leaders who want to determine impact, efficacy, and persistence over time.
Some examples of when to utilize this report might include the following:
• When assessing the ability of a new instructional method to raise standardized test scores
• When measuring the impact of an innovative program on students’ attitudes
• When weighing the cost benefit of a novel approach on remediation
• When tracking community support from last year to this year • When monitoring the effect of professional development training on knowledge acquisition
The STAAR Impact Report – Year to Year by Student Group report answers the question, “Was the performance of these students statistically significantly different over their performance from the prior year?” And if so, “What was the effect or the impact of that difference?” In other words, did implementing some program have an impact on the same students’ performance over the prior year?”
The STAAR Impact Group to State report answer the question, “Was the performance of my students statistically significantly different than the rest of the population in the state?” And if so, “What was the effect or the impact of that difference?” In other words, did implementing some program have an impact on overall student performance, when compared to everyone else in the state who took that same test.
The t-test is a very useful tool in educational settings. Often we are comparing two methods, two programs, two groups of students, or two groups of stakeholders to make a decision. The t-test is a simple and straightforward statistic that allows us to get beyond a frequency count and is a tool that educational leaders can use to make decisions confidently and then take the appropriate action based on the results.
NOTE: Impact or effect does not automatically mean or suggest causation. Other confounding or contributing variables or factors could alter the effect or the impact. It is up to the person reviewing the report to know and understand program implementation and identify confounding variables to best ascertain the level of the impact or effect.