The Chi-square test stands as one of the most versatile and widely applied statistical methods in psychological research. As one of the most widely used statistical tools in psychology and the social sciences, the chi-square test helps researchers determine whether the patterns they observe in categorical data are genuinely meaningful — or simply the result of chance. This comprehensive guide explores the fundamentals of conducting a Chi-square test, from understanding its theoretical foundations to interpreting results and recognizing its limitations in psychological research contexts.

What Is the Chi-square Test?

The chi-square test — written as χ² and pronounced "kai-square" — is a non-parametric statistical test used to assess whether there is a significant association between two or more categorical variables. The Chi-square test is a statistical test for categorical data — data organised into distinct groups or categories rather than measured on a continuous scale. It compares observed frequencies (what your data actually shows) against expected frequencies (what theory or hypothesis predicts).

Unlike parametric tests like the t-test or ANOVA, the Chi-square test does not rely on assumptions about population distributions, making it especially useful when data don't meet the conditions for parametric analysis. This characteristic makes it particularly valuable in psychological research, where data often involves categorical classifications such as diagnostic categories, treatment responses, or demographic groupings.

Historical Development and Significance

The logic of hypothesis testing was first developed by Karl Pearson (1857–1936), a renaissance scientist working in Victorian London, who in 1900 published his landmark paper in the Philosophical Magazine elaborating the invention of the chi-square distribution and the goodness-of-fit test. Pearson invented the chi-square distribution specifically to serve the analytical needs of biologists, economists, and psychologists — fields where categorical and frequency-based data were common but lacked appropriate statistical tools.

Before Pearson's groundbreaking contribution, statistical methods were primarily designed for continuous data and relied heavily on the assumption of normal distribution. His work transformed how researchers validate theories against real-world data, and the chi-square test remains one of his most enduring legacies in statistical methodology.

Understanding the Purpose of the Chi-square Test

The primary goal of a Chi-square test is to assess whether there is an association between two categorical variables. The chi-square test is particularly useful when working with categorical data — data organized into mutually exclusive groups or classes. For example, a researcher might categorize participants by gender, political affiliation, or diagnosis status.

In psychological research, this might involve investigating relationships such as whether there is an association between gender and preference for a certain type of therapy, whether diagnostic category relates to treatment outcome, or whether personality type correlates with coping strategy preferences. The test helps researchers determine whether observed differences in frequency distributions are statistically significant or likely due to random chance.

Types of Chi-square Tests

There are three common applications of the Chi-squared test: test of independence, goodness-of-fit test and test for homogeneity. Each serves a distinct research purpose and is suited to different types of research questions.

Chi-square Test of Independence

The Chi-squared test of Independence is used to assess whether there is a relationship between two categorical variables. When a researcher has two categorical variables and wants to know whether they are related or independent of each other, the Chi-square test of independence is the right tool. This test assesses whether belonging to a certain group on one variable affects the likelihood of belonging to a particular group on another variable.

For example, researchers might want to investigate whether gender influences the likelihood of having a particular mental health diagnosis, or whether exposure to a specific intervention relates to behavioral outcomes. The data are arranged in a contingency table — a cross-tabulation where rows represent one variable and columns represent another.

Chi-square Goodness-of-Fit Test

The goodness of fit test asks whether the distribution of a single categorical variable matches a hypothesised distribution. In other words: does your observed data "fit" what you expected? Another application is the Chi-squared goodness-of-fit test, which is used to evaluate if the observed frequencies for a categorical variable match the expected frequencies based on a known distribution. For instance, if a market researcher wants to find out whether the distribution of video streaming service subscriptions—such as Netflix, Amazon, and Disney—in a specific city aligns with the national distribution, this test would be appropriate. By comparing the observed data from the city with the expected frequencies derived from national statistics, the researcher can determine if the distribution significantly differs.

In psychological research, this might be used to test whether the distribution of personality types in a clinical sample matches the distribution found in the general population, or whether responses to a survey question follow a uniform distribution.

Chi-square Test for Homogeneity

In the end, the Chi-squared test for Homogeneity is used to check if different populations share the same distribution for a categorical variable. As an example, a survey might be conducted to explore if subscription rates for streaming services vary across different age groups. This test is particularly useful when comparing multiple groups to determine whether they exhibit similar patterns of categorical responses.

Key Concepts: Observed and Expected Frequencies

Understanding the distinction between observed and expected frequencies is fundamental to grasping how the Chi-square test works.

Observed Frequencies

Observed frequencies are simply the actual counts recorded in each category during a study. For instance, if a researcher surveys 100 people about their preferred therapy approach and 60 choose cognitive behavioral therapy while 40 choose person-centered therapy, those raw numbers — 60 and 40 — are the observed frequencies. These represent the empirical data collected directly from the research sample.

Expected Frequencies

Expected frequencies represent what the data would look like if there were no relationship between the variables being studied — in other words, if the null hypothesis were true. The expected count is the frequency that would be expected in a cell, on average, if the variables are independent, and is calculated as the product of the row and column totals divided by the total number of observations.

The Chi-square statistic measures the discrepancy between these observed and expected frequencies. Larger discrepancies suggest that the variables may be related, while smaller discrepancies suggest independence.

Steps to Conduct a Chi-square Test

Conducting a Chi-square test involves a systematic process that ensures accurate and meaningful results. Here is a detailed breakdown of each step:

Step 1: Formulate Hypotheses

Every Chi-square test begins with clearly defined hypotheses. The null hypothesis (H₀) typically states that there is no association between the variables being studied—that they are independent of each other. The alternative hypothesis (H₁) proposes that an association does exist between the variables.

For example, if studying the relationship between gender and therapy preference:

  • Null Hypothesis (H₀): There is no association between gender and therapy preference; the variables are independent.
  • Alternative Hypothesis (H₁): There is an association between gender and therapy preference; the variables are not independent.

Step 2: Collect Data

Gather categorical data from your sample through appropriate research methods such as surveys, observations, or experimental procedures. The data should be organized in a contingency table (also called a cross-tabulation table) that displays the frequency counts for each combination of categories.

A contingency table typically shows one variable along the rows and another along the columns, with cells containing the frequency counts for each combination. Marginal totals (row and column totals) are also included to facilitate calculations.

Step 3: Calculate Expected Frequencies

Determine what the frequencies would be in each cell if the null hypothesis were true—that is, if there were no association between the variables. For a contingency table, the expected frequency for each cell is calculated using the formula:

Expected Frequency = (Row Total × Column Total) / Grand Total

This calculation must be performed for every cell in the contingency table. The expected frequencies represent the theoretical distribution that would occur if the two variables were completely independent.

Step 4: Compute the Chi-square Statistic

The Chi-square statistic quantifies the overall discrepancy between observed and expected frequencies. It is calculated using the formula:

χ² = Σ[(Observed − Expected)² / Expected]

This formula involves:

  • Subtracting the expected frequency from the observed frequency for each cell
  • Squaring this difference to eliminate negative values
  • Dividing by the expected frequency to standardize the contribution
  • Summing these values across all cells

The resulting Chi-square value indicates the magnitude of difference between what was observed and what was expected under the null hypothesis.

Step 5: Determine Degrees of Freedom

Degrees of freedom (df) represent the number of values that are free to vary when calculating a statistic. For a contingency table in a Chi-square test of independence, degrees of freedom are calculated as:

df = (number of rows − 1) × (number of columns − 1)

For example, in a 2×3 table (2 rows and 3 columns), the degrees of freedom would be (2−1) × (3−1) = 1 × 2 = 2. Degrees of freedom are essential for determining the critical value from the Chi-square distribution table.

Step 6: Compare to Critical Value or Calculate P-value

The final step involves determining whether the calculated Chi-square statistic is large enough to reject the null hypothesis. This can be done in two ways:

Method 1: Critical Value Approach
Use a Chi-square distribution table to find the critical value corresponding to your chosen significance level (typically α = 0.05) and the calculated degrees of freedom. If your calculated Chi-square statistic exceeds the critical value, you reject the null hypothesis.

Method 2: P-value Approach
Calculate the p-value associated with your Chi-square statistic and degrees of freedom. If the p-value is less than your chosen significance level (e.g., 0.05), you reject the null hypothesis. Most statistical software packages automatically provide p-values, making this the more commonly used approach in contemporary research.

Interpreting the Results

If the calculated Chi-square value exceeds the critical value (or if the p-value is less than the significance level), you reject the null hypothesis, suggesting there is a significant association between the variables. This means the observed frequencies differ from what would be expected by chance alone to a degree that is statistically meaningful.

Conversely, if the Chi-square value is less than the critical value (or the p-value is greater than the significance level), you fail to reject the null hypothesis, indicating no evidence of a relationship. This does not prove the variables are independent, but rather that the data do not provide sufficient evidence to conclude they are associated.

Beyond Statistical Significance: Effect Size

While the Chi-square test tells you whether an association exists, it does not indicate the strength or practical importance of that association. Cramer's V is a widely used effect size measure for Chi-square tests. Like Pearson's r, its benchmarks are: small = .10, medium = .30, and large = .50. Always report Cramer's V alongside your χ² result to give readers a full picture of your findings.

Effect size measures provide context for interpreting the practical significance of your findings. A statistically significant result with a small effect size may have limited practical importance, while a large effect size indicates a strong and potentially meaningful association.

Critical Assumptions of the Chi-square Test

Like all statistical tests, the Chi-square test relies on certain assumptions that must be met for the results to be valid and interpretable. Violating these assumptions can lead to misleading conclusions.

Assumption 1: Categorical Variables

Both variables must be categorical (nominal or ordinal) rather than continuous. Continuous data – such as scores on a psychological scale, reaction times, or blood pressure readings – does not meet this assumption. Because the chi-square test relies on frequency counts for nominal data, continuous variables violate its underlying framework.

However, if continuous data is meaningfully collapsed into discrete categories (for example, grouping ages into "18-25," "26-40," and "41+"), the chi-square test may then be applied – but this must be done thoughtfully, with categories defined before data collection rather than after seeing results.

Assumption 2: Independence of Observations

Independence is arguably the most critical assumption of the chi-square test. Each data point must genuinely stand on its own – the value of one observation should in no way influence, predict, or depend on another.

If the same subjects are measured at multiple time points – such as pre-test and post-test comparisons – the chi-square test is not appropriate, because those observations are linked. Similarly, if data is collected from naturally related individuals (such as siblings, couples, or members of the same household), the assumption of independence is likely violated. In these cases, alternatives like McNemar's test (for paired categorical data) are more appropriate.

When observations are dependent, the effective sample size is smaller than the recorded count, which inflates the chi-square statistic and raises the risk of incorrectly rejecting the null hypothesis – a Type I error.

Assumption 3: Adequate Expected Frequencies

One of the most important practical assumptions concerns the size of expected frequencies in each cell of the contingency table. When expected cell counts fall below 5, the chi-square test becomes unreliable, often producing inflated test statistics and an increased risk of false positives.

A standard (and conservative) rule of thumb (due to Cochran) is to avoid using the chi-square test for tables with expected cell frequencies less than 1, or when more than 20% of the table cells have expected cell frequencies less than 5. No cell should have an expected frequency below 1.

When this assumption is violated, researchers have a few options. The most common solution is to combine adjacent categories to increase cell counts – for example, merging two small response groups into a single broader category. If combining categories is not conceptually justified, Fisher's Exact Test is the recommended alternative for small expected frequencies, particularly in 2×2 tables.

Assumption 4: Mutually Exclusive Categories

Each observation must belong to one and only one category for each variable. Categories must be mutually exclusive, meaning an individual cannot be classified into multiple categories simultaneously. This ensures that frequency counts accurately represent distinct groups without overlap or duplication.

Assumption 5: Random Sampling

As with parametric tests, the non-parametric tests, including the χ2 assume the data were obtained through random selection. However, it is not uncommon to find inferential statistics used when data are from convenience samples rather than random samples. (To have confidence in the results when the random sampling assumption is violated, several replication studies should be performed with essentially the same result obtained).

Corrections and Alternatives for Assumption Violations

Yates' Continuity Correction

Frank Yates suggested a correction for continuity that adjusts the formula for Pearson's chi-squared test by subtracting 0.5 from the absolute difference between each observed value and its expected value in a 2 × 2 contingency table. This reduces the chi-squared value obtained and thus increases its p-value. This correction is particularly relevant for 2×2 tables with small sample sizes, though its necessity has been debated in recent statistical literature.

Fisher's Exact Test

The second is the Fisher's exact test, which is a bit more precise than the Chi-square, but it is used only for 2 × 2 Tables. For example, if the only options in the case study were pneumonia versus no pneumonia, the table would have 2 rows and 2 columns and the correct test would be the Fisher's exact. Fisher's Exact Test calculates the exact probability of obtaining the observed distribution of data, making it more appropriate when expected frequencies are small.

McNemar's Test

When dealing with paired or matched categorical data—such as before-and-after measurements on the same subjects—McNemar's Test is the appropriate alternative. This test accounts for the dependency between observations and is specifically designed for analyzing changes in categorical responses within the same individuals over time.

Practical Applications in Psychological Research

The Chi-square test finds extensive application across various domains of psychological research. Understanding these applications helps researchers recognize when this test is most appropriate.

Clinical Psychology

In clinical settings, researchers might use Chi-square tests to examine relationships between diagnostic categories and treatment outcomes, to assess whether certain demographic characteristics are associated with specific mental health conditions, or to evaluate whether different therapeutic interventions produce different rates of symptom remission.

Social Psychology

By employing the chi-square test, the study examines the associations between respondents' demographic characteristics and their attitudes towards inequality. Social psychologists frequently use Chi-square tests to analyze survey data, examining relationships between demographic variables and attitudes, beliefs, or behaviors. This might include studying associations between social identity categories and political preferences, or between group membership and prosocial behavior.

Developmental Psychology

Developmental researchers might employ Chi-square tests to investigate whether developmental milestones are associated with particular environmental factors, or whether age groups differ in their categorical responses to experimental manipulations. For example, examining whether attachment style categories differ across different parenting style categories.

Educational Psychology

Educational researchers can apply Chi-square tests to analyze relationships between teaching methods and student performance categories, or to examine whether learning style preferences are associated with academic achievement levels. This helps in understanding the effectiveness of different educational interventions across diverse student populations.

Reporting Chi-square Results in Research

Proper reporting of Chi-square results is essential for transparency and replicability in psychological research. A complete report should include:

  • The Chi-square statistic value: Report the calculated χ² value
  • Degrees of freedom: Include the df in parentheses
  • Sample size: Report the total N
  • P-value: State the exact p-value or indicate p < .001 for very small values
  • Effect size: Include Cramer's V or phi coefficient
  • Descriptive statistics: Provide the contingency table or describe the pattern of results

Example: "A Chi-square test of independence revealed a significant association between gender and therapy preference, χ²(2, N = 200) = 12.45, p = .002, Cramer's V = .25. Examination of the contingency table indicated that female participants were more likely to prefer cognitive-behavioral therapy, while male participants showed a preference for solution-focused therapy."

Important Considerations and Limitations

Sample Size Considerations

Adequate sample size is essential for reliable Chi-square test results. The test is sensitive to sample size. In very large samples, even trivial differences between observed and expected frequencies may produce statistically significant results, while in very small samples, the test may lack power to detect meaningful associations.

With very large samples, researchers should pay particular attention to effect sizes to determine whether statistically significant results are also practically meaningful. With small samples, researchers may need to combine categories or use alternative tests like Fisher's Exact Test.

Association Does Not Imply Causation

A fundamental limitation of the Chi-square test is that it only assesses association, not causation. Finding a significant relationship between two categorical variables does not mean that one variable causes changes in the other. The association could be due to a third variable, reverse causation, or other confounding factors.

To establish causal relationships, researchers need to employ experimental designs with random assignment, control groups, and manipulation of independent variables. The Chi-square test can be used within such designs to analyze categorical outcomes, but the test itself does not establish causality.

Directionality and Specificity

The Chi-square test indicates whether an overall association exists but does not specify the nature or direction of that association. When a significant result is obtained, researchers must examine the contingency table and calculate standardized residuals or conduct post-hoc analyses to understand which specific cells contribute most to the overall association.

Loss of Information Through Categorization

When continuous variables are converted into categories to use the Chi-square test, information is lost. For example, converting age from a continuous variable into categories like "young," "middle-aged," and "older" discards the precise age information. Whenever possible, researchers should use statistical tests appropriate for the original measurement level of their data rather than artificially categorizing continuous variables.

Advanced Topics: The 2×2 Contingency Table

The 2 × 2 contingency table — also called a fourfold contingency table — is the simplest and most frequently encountered form of the independence test in psychology research. It involves two categorical variables each with exactly two levels, producing four cells (A, B, C, D). The rows represent two classifications of one variable (e.g., outcome positive / outcome negative) and the columns represent two classifications of another variable (e.g., treatment group / control group).

The 2×2 table is particularly common in clinical trials, diagnostic accuracy studies, and experimental research with binary outcomes. Special considerations apply to 2×2 tables, including the option to use Yates' continuity correction and the availability of Fisher's Exact Test as an alternative when expected frequencies are small.

Software Implementation

Modern statistical software packages make conducting Chi-square tests straightforward. Popular options include:

  • SPSS: Provides comprehensive Chi-square analysis through the Crosstabs procedure, including automatic calculation of expected frequencies, residuals, and effect sizes
  • R: Offers the chisq.test() function with options for continuity correction and simulation-based p-values
  • Python: The scipy.stats module includes chi2_contingency() for Chi-square tests
  • Excel: Includes the CHISQ.TEST function for basic Chi-square calculations
  • SAS: Provides Chi-square analysis through PROC FREQ

While software automates calculations, researchers must still understand the underlying logic, verify assumptions, and interpret results appropriately. Software cannot determine whether the Chi-square test is appropriate for a given research question or whether assumptions have been met.

Common Mistakes to Avoid

Using Percentages Instead of Frequencies

The data in the cells should be frequencies, or counts of cases rather than percentages or some other transformation of the data. The Chi-square test requires raw frequency counts, not proportions or percentages. Converting frequencies to percentages before analysis will produce incorrect results.

Ignoring Expected Frequency Requirements

Proceeding with a Chi-square test when expected frequencies are too small is a common error that can lead to unreliable results. Always check expected frequencies before interpreting results, and use appropriate alternatives when this assumption is violated.

Applying the Test to Dependent Data

Using the standard Chi-square test for repeated measures or matched pairs violates the independence assumption. Researchers must use specialized tests like McNemar's Test for such designs.

Failing to Report Effect Sizes

Reporting only the Chi-square statistic and p-value without including an effect size measure provides an incomplete picture of the results. Effect sizes are essential for understanding the practical significance and magnitude of associations.

Post-hoc Category Collapsing

Care should be taken when cell categories are combined (collapsed together) to fix problems of small expected cell frequencies. Collapsing can destroy evidence of non-independence, so a failure to reject the null hypothesis for the collapsed table does not rule out the possibility of non-independence in the original table. Categories should be defined a priori based on theoretical considerations, not adjusted after seeing the data to achieve desired results.

Enhancing Chi-square Analysis: Post-hoc Procedures

When a Chi-square test reveals a significant overall association in a table with more than two rows or columns, researchers often want to know which specific cells or combinations contribute most to the significant result. Several post-hoc procedures can provide this information:

Standardized Residuals

Standardized residuals indicate how much each cell's observed frequency deviates from its expected frequency, expressed in standard deviation units. Cells with standardized residuals greater than ±2 or ±3 are considered to contribute substantially to the overall Chi-square statistic and warrant closer examination.

Adjusted Residuals

Adjusted residuals account for the varying sample sizes across cells and can be interpreted similarly to z-scores. They provide a more refined assessment of which cells differ significantly from expected values.

Partitioning Chi-square

For larger contingency tables, researchers can partition the overall Chi-square into components by conducting separate Chi-square tests on subtables. This approach requires careful consideration of multiple comparison issues and appropriate adjustment of significance levels.

Integrating Chi-square Tests into Research Design

Effective use of Chi-square tests requires thoughtful integration into the overall research design. Researchers should consider several factors during the planning stage:

Power Analysis

Conducting a priori power analysis helps determine the sample size needed to detect an association of a given magnitude with adequate statistical power. This prevents underpowered studies that may fail to detect meaningful associations and overpowered studies that detect trivial associations.

Category Definition

Careful consideration should be given to how categories are defined. Categories should be:

  • Theoretically meaningful and relevant to the research question
  • Mutually exclusive and exhaustive
  • Defined before data collection begins
  • Sufficiently broad to ensure adequate cell frequencies
  • Not so broad that important distinctions are lost

Multiple Testing Considerations

When conducting multiple Chi-square tests within a single study, researchers must address the increased risk of Type I errors. Bonferroni correction or other multiple comparison procedures should be applied to maintain the overall error rate at the desired level.

Real-World Example: A Complete Analysis

To illustrate the complete process, consider a hypothetical study examining the relationship between stress management technique preference and personality type in a sample of 240 university students.

Research Question: Is there an association between personality type (introvert vs. extrovert) and preferred stress management technique (meditation, exercise, or social support)?

Hypotheses:

  • H₀: Personality type and stress management preference are independent
  • H₁: Personality type and stress management preference are associated

Data Collection: Students complete a personality assessment and indicate their preferred stress management technique. Results are organized in a 2×3 contingency table.

Assumption Checking:

  • Both variables are categorical ✓
  • Observations are independent (each student counted once) ✓
  • Categories are mutually exclusive ✓
  • Expected frequencies all exceed 5 ✓
  • Random sampling from target population ✓

Analysis: Calculate expected frequencies for each cell, compute the Chi-square statistic, determine degrees of freedom (df = 2), and obtain the p-value.

Results: Suppose χ²(2, N = 240) = 15.82, p < .001, Cramer's V = .26. The significant result indicates an association between personality type and stress management preference, with a small-to-medium effect size.

Interpretation: Examination of standardized residuals reveals that introverts show a stronger preference for meditation, while extroverts prefer social support as a stress management technique. Exercise shows no clear pattern by personality type.

Future Directions and Advanced Applications

As statistical methodology continues to evolve, several advanced extensions of the Chi-square test have emerged for specialized applications:

Log-linear Models

Log-linear models extend Chi-square analysis to examine relationships among three or more categorical variables simultaneously. These models can assess main effects, interactions, and complex patterns of association in multidimensional contingency tables.

Logistic Regression

When one categorical variable can be considered a dependent variable and others as predictors, logistic regression provides a more flexible framework than Chi-square tests. This approach allows for the inclusion of both categorical and continuous predictors and provides odds ratios for interpretation.

Correspondence Analysis

Correspondence analysis provides a graphical representation of associations in contingency tables, allowing researchers to visualize patterns of relationship between categories. This technique is particularly useful for large, complex tables where patterns may not be immediately apparent from numerical results alone.

Practical Tips for Researchers

Based on extensive application in psychological research, here are practical recommendations for using Chi-square tests effectively:

  • Plan ahead: Determine during the design phase whether Chi-square will be appropriate for your research question and data type
  • Check assumptions systematically: Create a checklist of assumptions and verify each one before interpreting results
  • Report comprehensively: Include all relevant statistics, effect sizes, and descriptive information in your results section
  • Visualize your data: Create bar charts or mosaic plots to help readers understand the pattern of associations
  • Consider alternatives: Be aware of alternative tests that may be more appropriate for specific situations
  • Interpret cautiously: Remember that significant associations do not imply causation and require theoretical interpretation
  • Replicate findings: Given the sensitivity to sample size and other factors, replication strengthens confidence in results

Conclusion

The Chi-square test remains an indispensable tool in the psychological researcher's statistical toolkit. Its versatility in analyzing categorical data, combined with its relatively straightforward interpretation and minimal distributional assumptions, makes it particularly valuable for addressing research questions involving nominal or ordinal variables.

Understanding how to conduct a Chi-square test properly—from formulating hypotheses through checking assumptions, calculating statistics, and interpreting results—allows psychologists to analyze categorical data effectively. This leads to better insights into human behavior, mental processes, and the complex relationships between psychological variables.

However, researchers must remain mindful of the test's limitations and assumptions. The Chi-square test assesses association, not causation. It requires adequate sample sizes and expected frequencies. It is sensitive to violations of independence. And it provides information about whether an association exists, but not necessarily about the strength or practical importance of that association without accompanying effect size measures.

By combining proper application of Chi-square tests with thoughtful research design, appropriate alternative tests when needed, comprehensive reporting practices, and careful interpretation, psychological researchers can leverage this powerful statistical method to advance understanding of categorical relationships in human psychology. As research questions become increasingly complex and datasets larger, the fundamental principles underlying the Chi-square test continue to provide a solid foundation for analyzing categorical data in psychological science.

For researchers seeking to deepen their understanding of categorical data analysis, exploring resources on statistical methods in psychology, consulting comprehensive statistical textbooks, and engaging with the broader statistical community through professional organizations can provide valuable additional insights and keep practitioners current with evolving best practices in this essential area of psychological research methodology.