Statistical significance tests are routinely used in research reports on language teaching and learning. They are considered an essential characteristic of all good quantitative research. Therefore, as part of their strive for rigor and objective approaches to research, many journals also require such tests for publication of studies that involve numerical data. All the seven articles in this issue of Language Teaching Research have used statistical significance tests. In this editorial, I will briefly discuss significance testing and how it has been used in the studies. Statistical significance tests are mathematical techniques used to establish the likelihood of specific results observed in the sample data given the assumption that the data were drawn from the population. They are used to reject what is called the null hypothesis, which proposes that any relationships or effects seen in the sample do not exist in the population and have occurred as a result of sampling error. The population is the whole group of people of interest, and a sample is a subset of the population. If the null hypothesis is maintained, the researcher concludes that the observed relationship in the sample has occurred by chance. Otherwise, it is concluded that the relationship or effect is systematic and therefore exists in the population. Since it is not possible to determine the truth of a null hypothesis in absolute terms, hypothesis testing results are based on probabilities and do not provide definitive proofs. These probabilities are presented in the form of a numerical value known as the p-value. Researchers use a threshold level called the significance level (α, alpha) to reject or accept the null hypothesis, which is often set at.05 in social sciences. If the observed p-value is lower than or equal to this cut-off level, the null hypothesis is rejected and the results are considered to be statistically significant. There are different types of statistical significance tests used in different situations, depending on the research question and the kind of data. Some of the most common ones are t-tests, analysis of variance (ANOVA), correlation, regression, and chi-square tests.

中文翻译：

---

语言教学研究中的统计显着性检验

统计显着性检验通常用于语言教学和学习的研究报告中。它们被认为是所有优秀定量研究的基本特征。因此，作为努力追求严谨和客观的研究方法的一部分，许多期刊也要求在发表涉及数值数据的研究时进行此类测试。本期《语言教学研究》的七篇文章均使用了统计显着性检验。在这篇社论中，我将简要讨论显着性检验及其在研究中的使用方式。统计显着性检验是一种数学技术，用于确定样本数据中观察到的特定结果的可能性，假设数据来自总体。它们被用来拒绝所谓的零假设，这表明样本中看到的任何关系或影响在总体中都不存在，而是由于抽样错误而发生的。总体是感兴趣的整个群体，样本是总体的一个子集。如果维持原假设，研究人员会得出结论，样本中观察到的关系是偶然发生的。否则，得出的结论是这种关系或影响是系统性的，因此存在于总体中。由于不可能绝对确定零假设的真实性，因此假设检验结果基于概率并且不提供明确的证据。这些概率以称为 p 值的数值形式表示。研究人员使用称为显着性水平 (α, alpha) 拒绝或接受原假设，在社会科学中通常设置为 .05。如果观察到的 p 值低于或等于此截止水平，则拒绝原假设，结果被认为具有统计显着性。根据研究问题和数据类型，在不同情况下使用不同类型的统计显着性检验。一些最常见的是 t 检验、方差分析 (ANOVA)、相关性、回归和卡方检验。取决于研究问题和数据类型。一些最常见的是 t 检验、方差分析 (ANOVA)、相关性、回归和卡方检验。取决于研究问题和数据类型。一些最常见的是 t 检验、方差分析 (ANOVA)、相关性、回归和卡方检验。