Recommending and School Behaviour- perception analysis of school, based on a comprehensive attitudes to school survey
- Simon A
- Dec 5, 2023
- 3 min read
Updated: Feb 10
This survey (see attached results) was created and analysed initially to practice some regression skills learnt during an MBA course. The idea of using regression and different statistical techniques adds perspective and a greater understanding to the survey. Recently our organisation used an external provider, who is widely used in the industry. They did not use regression? This survey was completed earlier in the year but realised there was additional data that was not considered. This was the postal code data and the duration of the survey. The postal code helped in calculating travel distances and my question was whether this influenced perceptions? Also, when we rush through a survey, generally the answers are inaccurate and the question was whether this has an impact? Additionally the analysis looked at a new time period - does this change perspectives? And do different year levels?
Note: The survey was undertaken within several High School classes. It was not systematically completed by a set group or number per year or time period. It has a sample of over 100 respondents, of which some have repeated the survey a year later to see if there were changes in perceptions.
Methodology: Our analysis employed factor and regression analyses to unearth key factors influencing student behaviour and their propensity to endorse their school. We specifically included two additional variables - the time taken to complete the survey and travel time to school - to enhance our understanding of the diverse aspects of student experiences.
Year 7-10 vs Year 11-12 Analysis:
KMO Measure:
Year 7-10: Showed moderate consistency in students' responses.
Year 11-12: Higher KMO, indicating a more structured pattern in responses, especially in the combined analysis (0.6).
Factor Analysis:
Year 7-10: Key factors included Accessibility, Peer Friendship, and School Experience, with varying factor loadings indicating their influence on student perceptions.
Year 11-12: Factors like School Experience, Teaching Experience, and Student Behaviour became more prominent, especially in the combined data.
Regression Analysis:
Year 7-10: Significant predictors of Recommendation included factors like School Experience and Peer Friendship.
Year 11-12: School Experience emerged as a consistent predictor across all analyses, while Other Engagement was a significant predictor for Student Behaviour.
Analysis Across Different Periods:
Before 30/11/2023 vs On/After 30/11/2023:
KMO values showed a decrease over time, suggesting a shift in the consistency of responses.
Factor Analysis: Notable changes in factor loadings over time, with an increased emphasis on School Experience and Teaching Experience in the later period.
Regression Analysis: Varied influences on Recommendation and Student Behaviour over time, with a significant role of factors like Other Engagement and School Experience.
Combined Analysis (All Data):
KMO Measure: 0.79 for the combined data, indicating good suitability for factor analysis.
Factor Analysis: A comprehensive set of factors, such as School Experience, Teaching Experience, and Personal Effort, influenced student perceptions.
Regression Analysis:
Regression Analysis for 'Recommendation'
This model explored what factors predict a student's likelihood to recommend their school. Key findings include:
School_Experience (Positive): A strong positive predictor, indicating that positive experiences at school significantly increase the likelihood of students recommending their school.
Other_Engagement (Positive): Suggests that engagement in activities (other than academics) positively influences students' likelihood to recommend the school.
Personal_Effort (Negative): Indicates that higher personal effort is associated with a lower likelihood of recommending the school, which could reflect a disconnect between personal academic efforts and overall school satisfaction.
Teacher_Friendly: While this was significant in predicting class behaviour perceptions, it is not a significant predictor for recommending the school.
Regression Analysis for 'Student_Behaviour'
This model aimed to predict students' perceptions of their class behaviour based on various factors. Significant predictors included:
Teacher_Friendly (Positive): Indicates that students who perceive their teachers as friendly are likely to have a more positive view of their classmates' behaviour.
Personal_Effort (Negative): Suggests that students who put more effort into their studies tend to view their classmates' behaviour less favourably.
Other_Engagement (Positive): Implies that students engaged in activities beyond academics tend to have a more positive perception of class behaviour.
Conclusions:
The survey answered several questions. The time travelled did not affect students perceptions and neither did the duration to complete the survey. However different time periods during the year did change perceptions and the perceptions differed by year level.
R code and overall analysis:
Visual Data Insights:
