Mixed methods research combines the strengths of both qualitative and quantitative approaches to provide deeper insights than either approach alone. Dedoose, as a powerful mixed methods tool, offers several quantitative features that complement and integrate with its qualitative analysis capabilities. This blog explores when, why, and how to use these quantitative features effectively in your research.

Before diving into complex multivariate analyses, examining the distributions of individual continuous variables is a critical first step. This fundamental practice serves several important purposes:

1. Sample Representation Assessment: Distribution analysis helps you understand how your research sample compares to the broader population. Significant deviations may limit generalizability or indicate sampling bias.

2. Outlier Identification: Distributions reveal unusual cases that may disproportionately influence your results. These outliers might represent measurement errors, rare cases of interest, or opportunities for deeper qualitative exploration.

3. Natural Grouping Discovery: Multimodal distributions can reveal natural clusters within your data, suggesting meaningful subgroups that warrant separate analysis or comparison.

4. Assumption Verification: Many statistical tests (like t-tests and ANOVA) assume normally distributed data. Early distribution analysis helps determine if your data meets these assumptions or if alternative approaches or interpretations are needed.

Key Psychometric Properties to Examine

When analyzing variable distributions in Dedoose, pay attention to:

• Central Tendency: Mean, median, and mode provide different perspectives on what's ‘typical’ in your data

• Dispersion: Standard deviation and range reveal the spread of your data

• Shape: Skewness (asymmetry) and kurtosis (peakedness) describe the distribution's form

• Visualization: Bar charts provide immediate visual insights into distribution characteristics

Multivariate Analysis: Comparing Groups and Discovering Relationships

Once you understand your individual variables, Dedoose offers several tools for examining relationships between variables:

T-Tests: Comparing Two Groups

1. When to use: When comparing means of a continuous variable between two distinct groups.

2. Example application: Comparing therapy outcomes between an experimental and a control group or examining if income levels differ between participants with different educational backgrounds.

3. How to interpret: The p-value indicates whether differences are statistically significant.

Analysis of Variance (ANOVA): Comparing Multiple Groups

1. When to use: When comparing means across three or more groups on a continuous variable.

2. Example application: Comparing health outcomes across different age categories, treatment modalities, or demographic groups.

3. How to interpret: A significant F-statistic indicates at least one group differs from others. Post-hoc tests, also available in Dedoose, can then identify which specific pairs of groups differ.

Chi-Square: Analyzing Categorical Relationships

1. When to use: When examining relationships between categorical variables.

2. Example application: Testing if gender is associated with program participation rates, or if treatment type relates to categorical outcome measures.

3. How to interpret: A significant result indicates that the observed frequency pattern differs from what would be expected by chance alone.

Bivariate Correlation: Measuring Variable Relationships

1. When to use: When assessing the strength and direction of relationships between continuous variables.

2. Example application: Examining how participant age relates to intervention outcomes, or how different quantitative measures correlate with each other.

3. How to interpret: Correlation coefficients range from -1 to +1, with values closer to these extremes indicating stronger relationships.

‍

The true power of Dedoose emerges when quantitative findings inform qualitative analysis and vice versa. Below are some features that highlight the integrative nature of qualitative and quantiative data in Dedoose.

Strategic Integration Approaches

1. Code Weighting: Dedoose's code weight system allows you to assign quantitative dimensions to qualitative codes, enabling numeric analysis of otherwise qualitative data.

2. Subgroup Analysis: Quantitative findings can identify meaningful subgroups whose qualitative data can then be analyzed separately to reveal group-specific themes.

3. Mixed Displays: Create visualizations that integrate both numeric patterns and illustrative qualitative excerpts.

4. Sequential Exploration: Use quantitative patterns to guide qualitative inquiry into particularly interesting cases or themes.

Practical Workflow for Mixed Methods Analysis in Dedoose

1. Begin with distribution analysis of all continuous variables to understand their characteristics.

2. Create categorical variables based on natural breakpoints or meaningful thresholds in continuous data.

3. Conduct appropriate statistical tests to identify significant patterns (t-tests, ANOVA, chi-square, correlation).

4. Use quantitative findings to inform qualitative coding and excerpt exploration.

5. Apply code weights to integrate quantitative dimensions into qualitative analysis.

6. Iterate between approaches, allowing each methodology to inform and enhance the other.

7. Utilize dynamic data visualizations, by clicking the bars, cell, or segment aspects of Dedoose data visualizations automatically extracts and presents the underlying qualitative data for examination.  

By thoughtfully applying these quantitative features while maintaining connection to your qualitative data, Dedoose enables researchers to develop richer, more nuanced understandings of complex phenomena that neither approach could achieve alone.

‍