Conducting research using a valid and reliable method to foster data analysis and interpretation
Inspirational framework 'Experiment and Analysis' for PhD candidates
Added valueof theresearch stage
Objective of the inspirational framework
Achievementsand training opportunities
Bestpractice by PhD candidate
Expectations
Objective of the inspirational framework
Objective
- To provide inspiration on how to approach the 'Experiment and Analysis' stage during the PhD trajectory
- To clarify the minimal expectations regarding the experiment and analysis stage in PhD research
- To guide the members of the doctoral advisory committee and project counselor
towards their coaching opportunities for the various research stages
Added value of the research stage 'Experiment and Analyse'
Added value
- To validate hypotheses
- To generate reliable data
- To identify patterns and relationships
- To learn iterative
- To detect and refine error/bias
Expectations
- To develop and master skills and methods associated with the research domain
- To shape various research chapters of the PhD dissertation
- To apply these skills for the discussion section of the PhD dissertation
Expectations
PhD candidates enrolled from AY 2024-2025 onward are required to have at least one publication as (shared) first author in a peer-reviewed international scientific journal in order to fulfill the faculty minimum criteria prior to public defence. Nevertheless, the Orpheus Best Practice Handbook 2024 advices that the final thesis is normally based on the equivalent of approximately three papers or manuscripts (not all of which need to be accepted), although fewer may be accepted if published in highly rated journals and represents the output of at least 2½ years full-time research.
Take a look!
Achievements and training opportunities
Achievements
Observation
Conclusion
Question
Project- management
Analysis
Hypothesis
Experiment
Achievements and training opportunities: Projectmanagement
Project management is the result of the extent to which you can manage your project, your environment (various stakeholders involved in your project) and yourself. Key components of the project involve defining its scope, planning and execution, while maintaining a clear overview of all relevant subcomponents. Strong project management skills are essential to the success of the project.
Achievements
Take a look!
Take a look at the Doctoral School offer to increase your project management skills!
Achievements and training opportunities: Observation - Question - Hypothesis
Achievements
Observations arising from literature, previous experiments and/or pilot studies help generate new research questions. Based on these, a clear hypothesis can be formulated, providing a direction for determining the appropriate experimental design. Don’t forget to design your Data Management Plan (DMP) and pre-register your study protocol before starting with your experiments! Take a look at the frameworks and for PhD candidates, and get inspired!
Literature and Research Questions
Research Design and Management
Achievements and training opportunities: Experiment
Achievements
Sample collection
Data collection
Reproducibility and open data science
Training opportunities (FRTP)
Achievements and training opportunities: Analysis
Data exploration and visualization
Achievements
Validity and reliability analysis
Data analysis
Reproducibility and open data science
Training opportunities (FRTP)
Achievements and training opportunities: Conclusion
Achievements
Once your data analysis has been completed, it is important to interpret the results properly, whether they are expected or unexpected. The results should be linked to the previously stated research question and hypothesis, placed in context, and used to demonstrate how they advance the scientific field. This can, on its turn, help shape new research questions.
Take a look!
Take a look at the Nature Masterclasses platform to train you in interpreting the results of your data analysis and form a proper conclusion!
Best practice by PhD candidate
Best practice
Anke Adriaansen was a doctoral researcher at the Department of Rehabilitation Sciences. During her PhD, she studied the effect of voice therapy on children with vocal fold nodules.
Recruiting participants was challenging because this condition is not life-threatening. To facilitate recruitment, she created a flyer and shared it through schools, (social) media, speech therapists, and the Ghent University Hospital. In line with recent trends in her research field, she also conducted individual data analyses alongside group analyses, which provided deeper insight into her data. Following additional statistics courses helped her to draw accurate conclusions based on her research findings.
Acknowledgements and references
AcknowledgementsSpecial thanks to Anne Magherman and Valery Labarque for the outline of the inspirational frameworks. Many thanks to Team Open Science at the Ghent University - especially Laura Standaert - for providing a solid content-related foundation for the development of this inspirational framework. Sincere gratitude to dr. Anke Adriaansen for sharing her experience on the pitfalls and opportunities during the research phase of experiments and data analysis.ReferencesNature masterclasses Orpheus Best Practice Handbook 2024 KCGG Doctoral School of Ghent University Open Science at the Ghent University
Acknowledgementsand references
Data exploration (data cleaning, descriptives, outliers,...) and visualization (tables and/or figures) involve analyzing and presenting data in a way that helps identify patterns, trends, and insights.Tools: SPSS, R, Matlab, Python,... Take a look at the Doctoral School offer , the Knowledge 2 Connect seminars of the KCGG and the hands-on courses of the Biostatistics Unit to require more skills for data exploration and visualisation!
Findings from this research phase can lead to refinements in research questions, methods, or even entirely new avenues of investigation.
It is important to verify the quality of your data and analysis, in terms of validity and reliability. Validity ensures that you’re measuring what you intend to measure, while reliability ensures that your measurement tool or test is consistent and reproducible. Ideally, you perform validity and reliability analyses a priori through pilot studies. Or use statistical tests (correlation, factor analysis, or regression) to confirm validity/reliability. Take a look at the Inspirational Framework 'Research Design and Management' for PhD candidates and the courses of the Biostatistics Unit !
Data analysis is the process of inspecting, cleaning, transforming, and modelling various types of data:- Qualitative data: descriptive data that captures experiences, perceptions, and emotions (e.g., data from interviews) - Quantitative data: numerical, measurable data that can be analyzed statistically (e.g., data from wearable sensors) - Mixed method data: contains both qualitative and quantitative data. Take a look at the Nature Masterclasses platform with concrete tips to help you through this research stage!
Before collecting the sample of your study, it is essential to define and plan your experimental design.
Depending on the model type (human, animal or cellular), the process of collecting the study sample may differ:
• For human studies, recruitment usually begins with participant outreach through various communication channels (e.g., flyer, social media, etc.), followed by screening respondents based on predefined inclusion/exclusion criteria via interviews or questionnaires. This process should be preceded by a power calculation and ethical approval.
• For non-human studies (animal or cellular), sample collection focuses on defining sample characteristics and selection criteria, ensuring ethical approval, verifying availability and sourcing of the biological material, and confirming the power analysis and feasibility of obtaining the required sample size.Take a look at the Health, Innovation and Research Institute (HIRUZ) and Ghent University Ethics Committee for more information!
Analysis helps uncover trends, correlations, and causal relationships that may not be obvious through theoretical exploration alone.
Through systematic data collection, accurate and reliable analyses of qualitative or quantitative data can be performed.
Projectmanagement Systematic data collection Error detection/mitigationHypothesis testing and subsequent iteration/refinement (if necessary) Pattern identification Data interpretation and conclusion formulating
This research phase can increase insight in your own data and helps identify inconsistencies, biases, or errors.
Explore the more general training opportunities of the Faculty Research Training Program on 'Experiment and analysis' Module 'Biostatistics': Analyze your data appropriately Module 'Data Visualisation': Clear and attractive visuals for complex research data Module 'Artificial Intelligence (AI)': Learn essential AI knowledge and skills for application in (bio)medical research Or subscribe for more specific/specialist courses applicable to your specific PhD project (e.g., omics data analysis for biomolecular research, statistical parametric mapping for neuroimaging/biomechanical research,…).
Take a look at the Doctoral School offer for more information!
Foster reproducibility and open data science by documenting and publishing your entire data analysis workflow, including traceability of your source code from A to Z, version control, documenting data sources, and depositing your (raw) data freely available in appropriate repositories. This is in line with the FAIR Data Principles to make sure your data are findable, accessible, interoperable and reusable. Note that not all data or research projects are suitable for open sharing. There are legitimate reasons to restrict access to data (e.g., personal or patient data, valorisation goals, research security). Take a look at Research Data Management (RDM) and Open Science at the Ghent University and get inspired! Important: Some peer-reviewed journals consider the publication of a data paper, which is in line with the Ghent University's Open Science vision. Check the journal's author guidelines to see the options.
Foster reproducibility and open data science by keeping track of your lab notes and research methods during your experiments, for example by maintaining an Electronic Lab Notebook (ELN). Take a look at Research Data Management (RDM) documentation methods , and the Ghent University RSpace ELN to get inspired!
Explore the more general training opportunities of the Faculty Research Training Program on 'Experiment and analysis' Module 'Module: Citizen Science': Involve citizens/patients in your research Module 'Core Facilities / Biobank': Correctly use the research infrastructure Module 'Artificial Intelligence (AI)': Learn essential AI knowledge and skills for application in (bio)medical research Or expand your skills applicable to your specific PhD project through specific/specialist courses (e.g., 'Advanced MRI data acquisition and management' for neuroimaging research). Take a look at the Doctoral School offer for more information!
Data can be collected by conducting original experiments, either independently or in collaboration with other research groups, or by using existing datasets from your own research group, collaborating groups, or publicly available sources, or even through the application of computational tools (e.g., computer science, statistics, machine learning, and mathematics).In each case, you should check your data on quality, completeness and consistence. When reusing existing datasets, it's also essential to verify the conditions under which the data can be used (i.e., what are you allowed to do with the data, and what is not permitted?). There may be a contract, license, or specific terms and conditions with the party providing the data. Not sure? Ask for legal support via Tech Transfer or the Health, Innovation and Research Institute (HIRUZ) . Examples of data collection methods are surveys (e.g., REDCap), interviews, clinical assessments, wearable devices, public data repositories,...
Analysis of the data from your experiments can either confirm or reject your initial assumptions.
Inspirational framework 'Experiment and analyse'
Onderwijs FGE
Created on July 18, 2024
Start designing with a free template
Discover more than 1500 professional designs like these:
View
Hierarchy of Needs
View
Customer Empathy Map
View
Squares Diagram
View
Customer Journey Map
View
HR Organizational Chart
View
SWOT PRO
View
Branching diagram
Explore all templates
Transcript
Conducting research using a valid and reliable method to foster data analysis and interpretation
Inspirational framework 'Experiment and Analysis' for PhD candidates
Added valueof theresearch stage
Objective of the inspirational framework
Achievementsand training opportunities
Bestpractice by PhD candidate
Expectations
Objective of the inspirational framework
Objective
- To provide inspiration on how to approach the 'Experiment and Analysis' stage during the PhD trajectory
- To clarify the minimal expectations regarding the experiment and analysis stage in PhD research
- To guide the members of the doctoral advisory committee and project counselor
towards their coaching opportunities for the various research stagesAdded value of the research stage 'Experiment and Analyse'
Added value
Expectations
Expectations
PhD candidates enrolled from AY 2024-2025 onward are required to have at least one publication as (shared) first author in a peer-reviewed international scientific journal in order to fulfill the faculty minimum criteria prior to public defence. Nevertheless, the Orpheus Best Practice Handbook 2024 advices that the final thesis is normally based on the equivalent of approximately three papers or manuscripts (not all of which need to be accepted), although fewer may be accepted if published in highly rated journals and represents the output of at least 2½ years full-time research.
Take a look!
Achievements and training opportunities
Achievements
Observation
Conclusion
Question
Project- management
Analysis
Hypothesis
Experiment
Achievements and training opportunities: Projectmanagement
Project management is the result of the extent to which you can manage your project, your environment (various stakeholders involved in your project) and yourself. Key components of the project involve defining its scope, planning and execution, while maintaining a clear overview of all relevant subcomponents. Strong project management skills are essential to the success of the project.
Achievements
Take a look!
Take a look at the Doctoral School offer to increase your project management skills!
Achievements and training opportunities: Observation - Question - Hypothesis
Achievements
Observations arising from literature, previous experiments and/or pilot studies help generate new research questions. Based on these, a clear hypothesis can be formulated, providing a direction for determining the appropriate experimental design. Don’t forget to design your Data Management Plan (DMP) and pre-register your study protocol before starting with your experiments! Take a look at the frameworks and for PhD candidates, and get inspired!
Literature and Research Questions
Research Design and Management
Achievements and training opportunities: Experiment
Achievements
Sample collection
Data collection
Reproducibility and open data science
Training opportunities (FRTP)
Achievements and training opportunities: Analysis
Data exploration and visualization
Achievements
Validity and reliability analysis
Data analysis
Reproducibility and open data science
Training opportunities (FRTP)
Achievements and training opportunities: Conclusion
Achievements
Once your data analysis has been completed, it is important to interpret the results properly, whether they are expected or unexpected. The results should be linked to the previously stated research question and hypothesis, placed in context, and used to demonstrate how they advance the scientific field. This can, on its turn, help shape new research questions.
Take a look!
Take a look at the Nature Masterclasses platform to train you in interpreting the results of your data analysis and form a proper conclusion!
Best practice by PhD candidate
Best practice
Anke Adriaansen was a doctoral researcher at the Department of Rehabilitation Sciences. During her PhD, she studied the effect of voice therapy on children with vocal fold nodules.
Recruiting participants was challenging because this condition is not life-threatening. To facilitate recruitment, she created a flyer and shared it through schools, (social) media, speech therapists, and the Ghent University Hospital. In line with recent trends in her research field, she also conducted individual data analyses alongside group analyses, which provided deeper insight into her data. Following additional statistics courses helped her to draw accurate conclusions based on her research findings.
Acknowledgements and references
AcknowledgementsSpecial thanks to Anne Magherman and Valery Labarque for the outline of the inspirational frameworks. Many thanks to Team Open Science at the Ghent University - especially Laura Standaert - for providing a solid content-related foundation for the development of this inspirational framework. Sincere gratitude to dr. Anke Adriaansen for sharing her experience on the pitfalls and opportunities during the research phase of experiments and data analysis.ReferencesNature masterclasses Orpheus Best Practice Handbook 2024 KCGG Doctoral School of Ghent University Open Science at the Ghent University
Acknowledgementsand references
Data exploration (data cleaning, descriptives, outliers,...) and visualization (tables and/or figures) involve analyzing and presenting data in a way that helps identify patterns, trends, and insights.Tools: SPSS, R, Matlab, Python,... Take a look at the Doctoral School offer , the Knowledge 2 Connect seminars of the KCGG and the hands-on courses of the Biostatistics Unit to require more skills for data exploration and visualisation!
Findings from this research phase can lead to refinements in research questions, methods, or even entirely new avenues of investigation.
It is important to verify the quality of your data and analysis, in terms of validity and reliability. Validity ensures that you’re measuring what you intend to measure, while reliability ensures that your measurement tool or test is consistent and reproducible. Ideally, you perform validity and reliability analyses a priori through pilot studies. Or use statistical tests (correlation, factor analysis, or regression) to confirm validity/reliability. Take a look at the Inspirational Framework 'Research Design and Management' for PhD candidates and the courses of the Biostatistics Unit !
Data analysis is the process of inspecting, cleaning, transforming, and modelling various types of data:- Qualitative data: descriptive data that captures experiences, perceptions, and emotions (e.g., data from interviews) - Quantitative data: numerical, measurable data that can be analyzed statistically (e.g., data from wearable sensors) - Mixed method data: contains both qualitative and quantitative data. Take a look at the Nature Masterclasses platform with concrete tips to help you through this research stage!
Before collecting the sample of your study, it is essential to define and plan your experimental design. Depending on the model type (human, animal or cellular), the process of collecting the study sample may differ: • For human studies, recruitment usually begins with participant outreach through various communication channels (e.g., flyer, social media, etc.), followed by screening respondents based on predefined inclusion/exclusion criteria via interviews or questionnaires. This process should be preceded by a power calculation and ethical approval. • For non-human studies (animal or cellular), sample collection focuses on defining sample characteristics and selection criteria, ensuring ethical approval, verifying availability and sourcing of the biological material, and confirming the power analysis and feasibility of obtaining the required sample size.Take a look at the Health, Innovation and Research Institute (HIRUZ) and Ghent University Ethics Committee for more information!
Analysis helps uncover trends, correlations, and causal relationships that may not be obvious through theoretical exploration alone.
Through systematic data collection, accurate and reliable analyses of qualitative or quantitative data can be performed.
Projectmanagement Systematic data collection Error detection/mitigationHypothesis testing and subsequent iteration/refinement (if necessary) Pattern identification Data interpretation and conclusion formulating
This research phase can increase insight in your own data and helps identify inconsistencies, biases, or errors.
Explore the more general training opportunities of the Faculty Research Training Program on 'Experiment and analysis' Module 'Biostatistics': Analyze your data appropriately Module 'Data Visualisation': Clear and attractive visuals for complex research data Module 'Artificial Intelligence (AI)': Learn essential AI knowledge and skills for application in (bio)medical research Or subscribe for more specific/specialist courses applicable to your specific PhD project (e.g., omics data analysis for biomolecular research, statistical parametric mapping for neuroimaging/biomechanical research,…). Take a look at the Doctoral School offer for more information!
Foster reproducibility and open data science by documenting and publishing your entire data analysis workflow, including traceability of your source code from A to Z, version control, documenting data sources, and depositing your (raw) data freely available in appropriate repositories. This is in line with the FAIR Data Principles to make sure your data are findable, accessible, interoperable and reusable. Note that not all data or research projects are suitable for open sharing. There are legitimate reasons to restrict access to data (e.g., personal or patient data, valorisation goals, research security). Take a look at Research Data Management (RDM) and Open Science at the Ghent University and get inspired! Important: Some peer-reviewed journals consider the publication of a data paper, which is in line with the Ghent University's Open Science vision. Check the journal's author guidelines to see the options.
Foster reproducibility and open data science by keeping track of your lab notes and research methods during your experiments, for example by maintaining an Electronic Lab Notebook (ELN). Take a look at Research Data Management (RDM) documentation methods , and the Ghent University RSpace ELN to get inspired!
Explore the more general training opportunities of the Faculty Research Training Program on 'Experiment and analysis' Module 'Module: Citizen Science': Involve citizens/patients in your research Module 'Core Facilities / Biobank': Correctly use the research infrastructure Module 'Artificial Intelligence (AI)': Learn essential AI knowledge and skills for application in (bio)medical research Or expand your skills applicable to your specific PhD project through specific/specialist courses (e.g., 'Advanced MRI data acquisition and management' for neuroimaging research). Take a look at the Doctoral School offer for more information!
Data can be collected by conducting original experiments, either independently or in collaboration with other research groups, or by using existing datasets from your own research group, collaborating groups, or publicly available sources, or even through the application of computational tools (e.g., computer science, statistics, machine learning, and mathematics).In each case, you should check your data on quality, completeness and consistence. When reusing existing datasets, it's also essential to verify the conditions under which the data can be used (i.e., what are you allowed to do with the data, and what is not permitted?). There may be a contract, license, or specific terms and conditions with the party providing the data. Not sure? Ask for legal support via Tech Transfer or the Health, Innovation and Research Institute (HIRUZ) . Examples of data collection methods are surveys (e.g., REDCap), interviews, clinical assessments, wearable devices, public data repositories,...
Analysis of the data from your experiments can either confirm or reject your initial assumptions.