Skip to main content

Results & Analysis

Durvey's Results & Analysis suite provides 7+ sophisticated analysis types to extract insights from your Delphi study. From consensus detection to stakeholder comparisons, sentiment correlation to scenario modeling—all with publication-ready visualizations and exportable data.

Analysis Types Overview​

1. Standard Analysis​

What it shows: Consensus metrics, box plots, and distributional statistics for each thesis.

Key metrics:

  • Interquartile Range (IQR)
  • Standard Deviation (SD)
  • Median and mean ratings
  • Consensus thresholds
  • Box plot visualizations

When to use: Start here for every Delphi study—foundational consensus analysis.

2. Stakeholder Analysis​

What it shows: Compare responses across expert groups (academic vs. industry, domestic vs. international, etc.).

Key features:

  • Group-by-group breakdowns
  • Mann-Whitney U tests for statistical significance
  • Pairwise comparisons between groups
  • Sample size validation

When to use: When you have tagged participants by stakeholder type and want to understand perspective differences.

3. Dissent Analysis​

What it shows: Hidden disagreement beneath apparent consensus—outliers, bimodality, and desirability bias.

Detection methods:

  • Outlier identification (IQR and Z-score methods)
  • Bimodality coefficients
  • Desirability bias correlation tests

When to use: When IQR looks acceptable but you suspect hidden disagreement or social desirability effects.

4. Sentiment Analysis​

What it shows: Relationship between overall topic sentiment and specific thesis ratings.

Key insights:

  • Sentiment-thesis correlations
  • Optimism/pessimism bias patterns
  • Longitudinal sentiment tracking

When to use: When you collected pro/contra comments or overall sentiment ratings alongside thesis evaluations.

5. Scenario Analysis​

What it shows: Correlation patterns and future scenario clusters based on thesis relationships.

Analysis components:

  • Correlation matrices between theses
  • Fuzzy C-means clustering
  • Cross-impact network diagrams
  • Scenario narrative generation

When to use: When you have multiple related theses and want to identify coherent future scenarios or trend bundles.

6. Comments Analysis​

What it shows: Qualitative feedback organized by thesis, with sentiment tagging and filtering.

Features:

  • Theme coding and pattern detection
  • Pro/contra comment segregation
  • Quote extraction for publications
  • Integration with quantitative ratings

When to use: When you enabled participant comments and want to understand the "why" behind ratings.

7. Export & Raw Data​

What it shows: Complete data exports in multiple formats for external analysis.

Export options:

  • CSV/Excel raw data
  • SPSS/R/Python-ready formats
  • Publication-quality charts (PNG/SVG)
  • Comprehensive analysis reports

When to use: When you need data for custom analyses, statistical software, or publication supplements.

Accessing Analysis​

Requirements​

Analysis features unlock when:

  • Project is in Pre-Phase, Main Phase, or Completed status
  • At least one participant has submitted responses
  • Trial projects have limited analysis access
  1. Open your project
  2. Click Results & Analysis in sidebar
  3. Choose analysis type from tabs
  4. Select thesis (for thesis-specific analyses)
  5. Configure options (filters, groups, etc.)
  6. View results

Real-Time Updates​

Analyses refresh automatically as new responses arrive:

  • Metrics recalculate instantly
  • Charts update live
  • No manual refresh needed

Manual refresh: Click refresh button to force update.

Visualization Options​

Interactive Charts​

All visualizations support:

  • Hover tooltips (detailed values)
  • Zoom and pan
  • Legend toggling
  • Responsive design (mobile-friendly)

Chart Types​

Box Plots (Standard Analysis):

  • Quartile distributions
  • Outlier markers
  • Median indicators
  • Whisker ranges

Grouped Bar Charts (Stakeholder Analysis):

  • Side-by-side group comparisons
  • Confidence intervals
  • Color-coded by stakeholder

Correlation Matrices (Scenario Analysis):

  • Heatmaps showing thesis relationships
  • Positive/negative correlation indicators
  • Significance markers

Network Diagrams (Scenario Analysis):

  • Node-link visualizations
  • Correlation strength as edge weight
  • Cluster groupings

Exporting Charts​

Download publication-ready visualizations:

Formats:

  • PNG: High-resolution (300 DPI) for papers
  • SVG: Vector format for editing
  • PDF: Embedded in documents

Actions:

  1. Hover over chart
  2. Click export icon
  3. Choose format
  4. Download

Filtering and Customization​

Global Filters​

Apply filters across all analysis types:

By Participant Group:

  • Filter to specific tag groups
  • Compare filtered vs. full panel
  • Isolate expert subsets

By Round (delayed Delphi):

  • View round-by-round evolution
  • Track consensus formation
  • Compare round 1 vs. round 2

By Date Range:

  • Analyze early vs. late responses
  • Track temporal patterns
  • Identify submission waves

Analysis-Specific Options​

Standard Analysis:

  • Adjust consensus thresholds (IQR <1.5, SD concentration >75%)
  • Show/hide outliers
  • Choose quartile calculation method

Stakeholder Analysis:

  • Select groups to compare
  • Set significance level (p<0.05 default)
  • Enable/disable pairwise tests

Scenario Analysis:

  • Choose number of clusters (2-5)
  • Set correlation strength threshold
  • Include/exclude low-variance theses

Statistical Rigor​

Built-In Tests​

Durvey automatically applies appropriate statistical methods:

Non-parametric tests (Mann-Whitney U, Kruskal-Wallis):

  • Used for ordinal Likert data
  • No normality assumptions required
  • Robust to outliers

Effect sizes (Cohen's d, r):

  • Quantifies magnitude of differences
  • Complements significance tests
  • Reports in results tables

Confidence intervals:

  • 95% CIs for medians
  • Bootstrap methods
  • Displayed in visualizations

Sample Size Warnings​

Analysis warns when:

  • Group sample sizes <5 (unreliable comparisons)
  • Total responses <10 (insufficient for consensus)
  • Missing data >20% (potential bias)

Recommendations: System suggests when to:

  • Recruit more participants
  • Merge small groups
  • Extend data collection

Exporting Results​

Quick Exports​

Each analysis type has dedicated export button:

Standard Analysis Export:

  • Thesis-by-thesis summary statistics
  • Consensus status table
  • Box plot images

Stakeholder Analysis Export:

  • Group comparison tables
  • Mann-Whitney test results
  • Pairwise significance matrix

Comments Export:

  • Full comment text with metadata
  • Sentiment tags
  • Associated quantitative ratings

Comprehensive Export​

Batch Download:

  1. Click Export All Results
  2. Choose format (Excel, CSV, ZIP)
  3. Select analyses to include
  4. Download complete package

What's Included:

  • Raw participant data
  • All analysis tables
  • All charts (PNG/SVG)
  • Metadata and study configuration

Publication-Ready Outputs​

Tables​

Formatted for journals:

  • APA-style table formatting
  • Significance indicators (*, **, ***)
  • Professional typography
  • Copy-paste to Word/LaTeX

Figures​

Meets journal requirements:

  • 300 DPI resolution
  • Clear axis labels
  • Professional color schemes
  • Accessibility-compliant (colorblind-safe)

Reporting Standards​

Includes required information:

  • Sample sizes for each group
  • Test statistics and p-values
  • Effect sizes
  • Confidence intervals
  • Missing data rates

Advanced Features​

Longitudinal Analysis (Multi-Round)​

For delayed Delphi with multiple rounds:

  • Track consensus convergence
  • Identify stable vs. shifting opinions
  • Visualize opinion change over rounds
  • Calculate convergence rates

Custom Analysis Parameters​

Adjustable settings:

  • Consensus thresholds (customize IQR cutoffs)
  • Significance levels (0.05, 0.01, 0.001)
  • Clustering algorithms (fuzzy vs. k-means)
  • Correlation methods (Pearson, Spearman)

API Access (Pro Plan)​

Programmatic access to analysis results:

  • REST API endpoints for all analysis types
  • JSON-formatted data export
  • Automated report generation
  • Integration with external tools

Trial Project Limitations​

Trial projects have restricted analysis access:

  • Standard Analysis only
  • No stakeholder comparisons
  • No scenario analysis
  • Export limited to CSV

Upgrade to unlock: All 7+ analysis types, advanced visualizations, and full export options.

Best Practices​

Starting Your Analysis​

  1. Run Standard Analysis first (baseline consensus)
  2. Check Dissent Analysis if IQR seems acceptable but you suspect hidden disagreement
  3. Use Stakeholder Analysis if you have tagged groups
  4. Explore Scenario Analysis for multi-thesis futures
  5. Dive into Comments Analysis for qualitative context

Interpreting Results​

  1. Start with descriptive statistics (median, IQR)
  2. Check consensus thresholds
  3. Look for patterns across theses
  4. Investigate unexpected outliers
  5. Cross-reference quantitative and qualitative findings

Reporting Findings​

  1. Use box plots for consensus visualization
  2. Report both significance and effect sizes
  3. Include sample sizes in all tables
  4. Quote representative comments
  5. Acknowledge limitations

Troubleshooting​

No Data Showing​

Causes:

  • No participants have completed survey
  • Filters too restrictive
  • Analysis not available for project phase

Solutions:

  • Check participant completion status
  • Clear filters
  • Verify project is in active phase

Charts Not Loading​

Causes:

  • Insufficient data for visualization
  • Browser compatibility
  • Slow internet connection

Solutions:

  • Ensure at least 5 responses
  • Try different browser
  • Refresh page

Statistical Tests Unavailable​

Causes:

  • Sample sizes too small (<5 per group)
  • Only one group selected
  • Trial project limitation

Solutions:

  • Recruit more participants
  • Select multiple groups
  • Upgrade plan

Learn More​