EDF 802 Dr. Jeffrey Oescher Formative Exercise 3 Answers Revised 19 March 2013 Requisite Information 1. Statistical Hypotheses H0: μ1 - μ2 = 0 H1: μ1 - μ2 ≠ 0 2. Alpha Level The researcher has indicated some concern about committing a Type I error. To protect against this, an alpha level of .01 was selected. 3. Test Statistic and Sampling Distribution Assuming the null hypothesis is true, a sampling distribution of F with 1 and 189 degrees of freedom was generated. The actual test statistic is an F-statistic. This information was taken from the printout once the analysis was complete. You can calculate the degrees of freedom using the formulas in the outline or text, but they are contained in the ANOVA table between the Sums of Squares and Means Squares columns. Results 1. Statistical Information Describing the Sample and Relevant Variables Very little information is given about the sample beyond the fact that teachers in two types of schools were surveyed. Both types of schools were high poverty; one type was high performing (HPHP) while the other was low performing (HPLP). Eighty-seven and 104 teachers responded from HPHP and HPLP schools respectively. Table 2 presents the descriptive statistics for the total score for each type of school. Table 2A Descriptive Statistics for Organizational Health by Type of School Type of School HPHP HPLP Total N 87 104 191 Mean 3.05 2.69 2.85 SD 0.47 0.45 0.49 In general, teachers in high poverty, high performing schools perceived their schools as healthy organizations (see Table 2 in Formative Exercise 3). Teacher’s perceptions varied moderately in these schools. In contrast, teachers in high poverty, low performing schools perceived their schools as healthy in some respects and unhealthy in others. Teacher’s perceptions in these schools varied moderately also. An alternative perspective for examining this data is presented below. This information is offered only to provide some insight into the analysis and reporting of data. Please note the MOST important aspect of reporting this information is the interpretation of the scores. Please take note of the manner by which the six scores within each group were discussed. Table 2 presents the descriptive statistics for each subscale and the total scale for each type of school. Table 2A Descriptive Statistics for Organizational Health by Type of School Type of School HPHP HPLP Subscale CL II RI TA AE Total CL II RI TA AE Total N 87 87 87 87 87 87 104 104 104 104 104 104 Mean 3.11 3.03 3.07 3.21 2.65 3.05 2.77 2.95 2.59 2.83 2.11 2.69 SD 0.76 0.57 0.49 0.61 0.46 0.47 0.82 0.52 0.59 0.63 0.58 0.45 An examination of the data in Table 2 indicates teachers in HPHP schools perceive their schools as healthy organizations. Their total scores as well as four of the five subscale scores fell into this higher category of healthiness. The exception was the AE subscale (i.e., Academic Emphasis) where teachers perceived some healthy as well as unhealthy characteristics. Teacher’s perceptions tended to vary greatly or moderately on the subscales and moderately on the total scale. In contrast to these findings, teachers in HPLP schools perceive their schools as healthy in some respects and unhealthy in others. Scores across all subscales and the total scale were consistent in this regard. Teacher’s perceptions varied greatly to moderately on all subscales and moderately on the total scale. 2. Inferential Results Table 3 presents the results of the ANVOA. The computed F-statistic was statistically significant (F1,189 = 28.90, p = .000). The following information is the typical format for reporting the results of an ANOVA. The F-statistic for the analysis must be reported, all assumptions must be discussed, observed effect size mentioned, and the observed level of power identified. The assumption of normality was not tested as the ANOVA procedure is robust with respect to its violation. The assumption of the independence of observations is assumed to be true. The assumption of homogeneity of variance was tested using Levene’s statistic. The result (F1,189 = 0.00, p = .984) was non-significant indicating equivalent variances across the two groups. The level of power was 1.00, indicating very little if any chance of a Type II error. The effect size as measured by partial eta squared ( p ) was .13. This is considered large using the criteria established 2 by Huck (2010). Table 3 ANOVA Results Source SS Type of School 6.14 df 1 MS F Sig p2 Power 6.14 28.90 .000 .13 1.00 Error Total 40.17 46.32 189 190 .21 Table 3 in my file is split between two pages. This is unacceptable in APA. Sometimes it is the result of using a particular printer; other times the issue is with the margins. Either way, you cannot split a table unless it is longer than a single page. In Word, you can highlight the text you want to keep together, use the PARAGRAPH menu, select LINE AND PAGE BREAKS, and click on KEEP LINES TOGETHER. This will keep all of the highlighted text together on the same page. You DO NOT want to adjust you margins. 3. Statistical Conclusions The null hypothesis was rejected. There is a statistically significant difference in the organizational health of HPHP and HPLP schools. HPHP schools are more healthy organizations than are HPLP schools. Note: The statistical conclusions are only about the acceptance or rejection of the null hypothesis. This decision should lead directly from the statistical results and be consistent with them. There is no need to mention any statistical information in this section. Conclusions Relating the Statistical Results to the Research Questions 1. The underlying conceptual framework for this study suggested that organizational health represented a holistic view of the characteristics of a school that transcends culture and climate. Although both culture and climate are known to be related to student achievement, they represent rather narrow constructs in comparison to organizational health. That is, culture and climate are important, but not sufficient, to define the construct of organizational health. The results of this study validate - at a descriptive level - the hypothesis that organizational health is related to student achievement. However, additional research is needed to explore the relationships between leadership and organizational health in general as well as the specific leadership characteristics and behaviors that contribute to those factors of organizational health that are most influential on student achievement. This section is purely conjecture without some deeper understanding of the research that was undertaken and the conceptual framework behind it. An acceptable response need only refer to some conceptual interpretation of the results. Inferential Logic 1. The null and alternative hypotheses were stated. A conservative alpha level of .01 was established to protect against a Type I error. The null hypothesis was assumed to be true, allowing a sampling distribution of F with 1 and 189 degrees of freedom to be generated. An observed F-statistic was computed (F1,189 = 28.90) and mapped into the underlying sampling distribution. It was determined to be atypical of the F-statistics in this distribution (p = .000). Therefore the null hypothesis was rejected, and the alternative hypothesis accepted.