School Leadership That Works: From Research to Results (2005)
Marzano, Robert J.; Waters, Timothy; McNulty, Brian
Chapter 1:
Leadership vital to effectiveness of schools: “one great man” theories/ trait theories/ environmental—they emerge as a result of time, place, circumstance
Few studies on relationship of leadership to achievement—but books recommending various practices in school leadership are common
Basic claim: Meta-analysis shows specific leadership behaviors & their effect on achievement
Meta-analysis—synthesizing research quantitatively : Impossible to control all of the error in a study—why researchers assign a probability statement to results-- @ .05 level, the findings could occur 5 times in 100 as a function of error (chance).
Narrative reviews subject to erroneous conclusions
Reviews are biased by the theory to which reviewer subscribes
Basic findings from 69 studies: Computed correlations between leadership behavior and the average academic achievement of students to be R = .25. Example: Assume a new principal is hired to work in a school at the 50th percentile in the average achievement of its students. Also assume principal is at the 50th percentile in leadership abilities. If we increase principal’s ability by one SD (84th percentile), correlation of .25 assumes we can predict the average achievement of school to rise to the 60th percentile.
NOTE: This interpretation assumes a linear relationship between leadership and achievement. Also, the graphs on page 11 depict percentiles as an equal-interval scale, which is impossible.
Chapter 2: Theories and Theorists on Leadership
Transactional: Trading one thing for another: Maintaining the status quo/pay attention to issues that arise, set standards, monitor behavior/set goals, clarify desired outcomes/reward-for-performance
Transformational—popular in education—more focused on change: form mutual relationships/ convert followers into leaders
-The 4 “I’s”: Individual consideration (personal attention to the neglected)/ Intellectual stimulation (old problems in new ways)/ Inspirational motivation (high performance expectations)/ Idealized influence (modeling behavior through exemplary personal achievements, character, behavior)
Total Quality Management:
Comes from Demming (NOTE: Systems thinking)—restructured Japanese manufacturing post WWII
TQM Principles:
-Change Agency: isolate and eliminate structures that work against change
-Teaming—common tasks and goals (interdependence)
-Continuous Improvement—goal setting/incremental (link to first order change here)
-Eradication of Short Term Goals—ie., goals that are based on quotas, short-term
Servant Leadership:—Leader is positioned at center of organization/in contact with all aspects/not interacting with just a few high-level managers/listening, developing people’s skills
Situational Leadership: The leader adapts leadership behavior to the followers’ “maturity”; based on their willingness and ability
-unskilled and unwilling—directive
-unskilled and willing—friendly but concrete and directive
-skilled and unwilling- persuade and engage
-skilled and willing—let them do the work—no interference (link to laissez faire here)
Instructional Leadership:
-resource provider
-instructional resource
-visible presence
Chapter 3: Meta-Analysis
R (correlation) effect sizes were used
Extended leadership styles research
Typical study in the analysis used some type of questionnaire asking teachers about the principal’s leadership behaviors
Average score for teacher responses within each school was then correlated with average achievement of the students (NOTE: No evidence of analysis of how long the principal had been at the school or change over time, but there is an assumption of change with changes in principal abilities on the “21”)
The “n” or unit of analysis was the school—each school had a summary score
The average R of .25 is much higher than that reported in another meta-analysis by Witziers, Basker, & Kruger (2003). They reported an average correlation of .02. Their study used schools from various countries. Studies outside US report low outcomes. Marzano et al. excluded outliers where Witziers et al. did not.
Note: Marzano et al. report using Lipsey & Wilson’s procedures
Marzano et al. also corrected for attenuation—some studies used leadership questionnaires with very low reliabilities. Low reliability of an instrument will underestimate the R—so they adjusted upward.
Moderator Analysis: Figure 3.4 p. 35 is a frequency distribution of the correlations for each study. Results are very discrepant. Moderator analysis was necessary due to low homogeneity:
-Quality of studies (divided into low, medium, and high quality)—studies that were rated the highest produced the largest average correlation—they interpret this to mean the correlation they found was not an artifact of poor research design… However, they go on to say that the “high” studies were more precise in their definition and measurement of leadership behaviors.
NOTE: They report using Hedges & Olkin’s (1985) procedures
NOTE: Test for null hypothesis was non-significant…they say it “approaches significance”. Specific criteria for what makes a study “high”, “medium” or “low” quality: Manner in which sample was identified, appropriateness of measures used for independent and dependent variables, survey response rate, and appropriateness of data analysis---this is all irrelevant, as no “study quality” effect was found.
-Level of the school: no real differences between levels
NOTE: No evidence of a Q test
No significant results for any other moderator tested (there were 8); and a few had lack of sufficient data
Chapter 4: The 21 Responsibilities of the School Leader
-Affirmation—celebrates school accomplishments
-Change Agent—challenge status quo
-Contingent rewards—rewards individual accomplishments
-Culture—foster shared beliefs
-Discipline—protects teachers from interruption of the educational process
-Focus—clear goals
-Input-involves teachers in design and implementation
-Intellectual Stimulation—makes sure staff aware of most current theories
-Involvement in curriculum
-Knowledge of curriculum
-Monitoring and Evaluating
-Optimizer—new and challenging innovations
-Order –in the form of standard operating procedures
-Outreach—advocate for school in the community
-Resources-materials and PD
-Situational Awareness
Relative Effect- A correlation that does not include the value 0 is significant at the .05 level—All of the confidence intervals for correlations of the above behaviors to achievement were significant at the .05 level. All of the correlations were very close in magnitude.
NOTE: Interval is estimate of the “true” correlation between achievement and the various leadership responsibilities. The level of certainty that the average correlation accurately represents the true correlation is reported in the 95% confidence interval. Includes the range of correlations in which one can be 95% sure that the true correlation falls…
NOTE: It does not appear that the “responsibilities” correlations were weighted according to the number of studies (or schools) in which that characteristic was measured.
Chapter 5: Two Types of Change
To address the issue of relatedness, they completed a factor analysis of a scale they developed to to measure principals’ behavior in terms of the 21 responsibilities. They claim to have found two factors, and they situated those factors under First Order Change and Second Order Change.
NOTE: This was an initial validation of their scale. 652 principals responded on a website. This was an exploratory and not a confirmatory factor analysis. The cutoff for factor loadings was very low (R = .15)—they did this because of the number of respondents: “because the statistical significance between a factor loading and an eigenvector depends on the sample size, the criterion for classifying a variable as an element of a factor should be based on the value of the factor loading (correlation) needed to achieve an acceptable error rate.” P. 165
NOTE: There were 91 items on the questionnaire, which will significantly impact the alpha of 95% they report.
First Order change: Incremental—the next most obvious step
Second Order—alters the system in fundamental ways—requires new ways of thinking and acting.
p. 70: “Saying that all 21 responsibilities are related to first order change is another way of saying that all 21 should define the standard operating procedures in a school”.
Leadership for second order change: They found second order change is related to 7 of the responsibilities:
-Knowledge of Curriculum
-Intellectual Stimulation
-Change Agent
Second order change manifests only in context of a specific problem or issue
Three of the 7 responsibilities important to second order change are ranked low in terms of their relative importance to first order change (NOTE: When there is little or no statistical difference between the correlations for each “responsibility”, it is very hard to “rank them”—yet they are in this study)…
The three are : Challenging status quo, Optimizer, and Flexibility
Revealing: Some responsibilities are negatively affected by second order change: Culture, Communication, Order, and Input—these things may seem to deteriorate as a result of the change
Chapter 6: Doing the Right Work
Downfall of schools has to do with poor decisions about what to work on. Richard Elmore—identifying the right work
Two approaches:
Extent to which CSR have research supporting them varies widely: Highly uneven effect on achievement
The more staff development, the lower the effect size. Marzano et al.—interpretation of this is that only when schools adapt the model to their specific needs does it become ptimized
-Models must be adapted if they are to be sustained.
Marzano “What Works in Schools model—outlines 11 factors that could be the focus of school reform. Each of the 11 represents an area of change that can be implemented w/o big funding and all are doable.
Standards movement has created a crisis of coverage—standards documents identify far more content than teachers can cover in the allotted time. Marzano, Kendall, & Gaddy (1999)—study of state level standards documents & determined that teaching the content in these documents would require 71% more instructional time.
Factor 1: Once curriculum is trimmed to essentials, it can be guaranteed.
Factor 2: Challenging goals and effective feedback (eg. Marzano, 2003 standards based report card)
Factor 3: Parent & Community involvement
Factor 4: Safe & Orderly
Factor 5: Collegiality and Professionalism
Factor 6: Instructional Strategies (eg., the “Marzano strategies identified through meta-analysis)
Factor 7: Classroom management
Factor8: Classroom Curriculum Design—decisions teachers make to adopt the content in the textbooks, standards, curriculum guides
-have to consider what students already know about a particular topic (NOTE: Tie-in to Schumm/Pyramid)
Factor 9: Home environment—discussion re: school, provide work support, monitor TV, monitor homework
Factor 10: Learned Intelligence and Background Knowledge
-Crystallized Intelligence---they claim biggest predictor of student achievement
-Direct vocabulary instruction
-Wide reading
Factor 11: Motivation
-Provide students with meaningful work to do—involve them in long-term projects of their own design (NOTE: Relate to Berger here)
-Provide specific feedback on their knowledge gain ( NOTE: Specific relationships to Berger and Moorman here)
Chapter 7: A Plan for Effective School Leadership
1) Develop Strong School Leadership Team---no one can address the “21” well alone—a purposeful community—accomplishes goals that matter to all/development and use of available assets/agreed-upon processes (NOTE: they mention Good to Great—OB book by Jim Collins: “getting the right people on the bus”….)
2) Distribute the responsibility among the team members
3) Select the right work—eg, the what works in schools Marzano questionnaire
4) Identify Order of magnitude required by the Selected Work: One person’s first order change may be another person’s second order change—defined by the way people react t a proposed innovation
5) Match the Management Style to the Order of the Magnitude of the Change Initiative
First order change requires attention to all of the “21”
Second order change : See Notes Chapter 5