Results Interpretation Assistant
Enter a result from your statistics software, and the assistant explains what it means, whether it is statistically significant, how large the effect is, and what it says about your hypotheses, objectives and research question. It never calculates statistics; it explains the ones you have.
Your project
Enter what you already know, so each result can be linked to your hypotheses and objectives.
One per line.
One variable per line. Put its indicators after a colon, separated by semicolons, such as “wellbeing: life satisfaction; positive mood”.
A plan from the Sample Size Calculator, using Cochran's formula at 95% confidence.
Your result
Choose the analysis, then enter the numbers exactly as your software reported them.
Results explained
Every kind of result the assistant interprets: what it tells you, what to enter, and the mistakes to avoid. References will be added after academic review.
Descriptive statistics
Summarises who took part and what each variable looks like before any test.
Learn more about interpreting Descriptive statistics
What to enter
- Sample size (N), required
- Mean (M), required
- Standard deviation (SD), required
- Median (Mdn)
- Lowest value observed (min)
- Highest value observed (max)
Common mistakes to avoid
- Reporting a mean without its standard deviation.
- Using the mean for skewed data, where the median describes the typical value better.
- Generalising from descriptive statistics to the population without a test or confidence interval.
References awaiting academic review.
Frequency
Counts how many participants gave each answer or fall in each category.
Learn more about interpreting Frequency
What to enter
- Count in the category (n), required
- Total (N), required
Common mistakes to avoid
- Leaving out the total the count is out of.
- Comparing counts between groups of different sizes instead of percentages.
References awaiting academic review.
Percentage
Shows each category's share of the sample, so groups of different sizes can be compared.
Learn more about interpreting Percentage
What to enter
- Percentage (%), required
- Total it is a percentage of (N)
Common mistakes to avoid
- Reporting a percentage without the number it is of.
- Giving more decimal places than a small sample justifies.
- Letting percentages of overlapping categories add up to more than 100% without saying why.
References awaiting academic review.
Mean
Gives the average value of a numeric variable.
Learn more about interpreting Mean
What to enter
- Mean (M), required
- Standard deviation (SD)
- Sample size (N)
- Lowest possible score (min)
- Highest possible score (max)
Common mistakes to avoid
- Reporting a mean without its standard deviation.
- Averaging a single ordinal item, such as one rating, without justification.
- Reading a mean near the middle of a scale as “neutral” when answers may be split between the ends.
References awaiting academic review.
Median
Gives the middle value, which extreme values don't distort.
Learn more about interpreting Median
What to enter
- Median (Mdn), required
- Sample size (N)
- Lowest possible score (min)
- Highest possible score (max)
Common mistakes to avoid
- Reporting the median without the range or interquartile range.
- Using the median for a numeric variable that is symmetric, where the mean would use more of the data.
References awaiting academic review.
Standard deviation
Shows how spread out a numeric variable's values are around the mean.
Learn more about interpreting Standard deviation
What to enter
- Standard deviation (SD), required
- Mean (M)
- Lowest possible score (min)
- Highest possible score (max)
Common mistakes to avoid
- Reporting the standard deviation without the mean.
- Confusing the standard deviation (spread of scores) with the standard error (precision of the mean).
References awaiting academic review.
Cronbach's alpha
Estimates the internal consistency of a set of items that form one scale.
Learn more about interpreting Cronbach's alpha
What to enter
- Cronbach's alpha (α), required
- Number of items (k)
- Sample size (N)
Common mistakes to avoid
- Treating a high alpha as proof that the items measure one thing; it doesn't test that.
- Calculating alpha before reverse-scoring reverse-worded items.
- Reporting one alpha for items that measure several different constructs.
- Treating .70 as a fixed pass mark rather than a convention.
References awaiting academic review.
Correlation
Measures how strongly two variables move together.
Learn more about interpreting Correlation
What to enter
- Correlation coefficient (r), required
- p-value (p), required
- Sample size (N)
Common mistakes to avoid
- Saying the correlation shows that one variable causes the other.
- Reporting a non-significant correlation as proof that the variables are unrelated.
- Judging strength from the p-value instead of the coefficient.
- Leaving out the direction (positive or negative) or the sample size.
References awaiting academic review.
Pearson correlation
Measures the strength and direction of a straight-line relationship between two numeric variables.
Learn more about interpreting Pearson correlation
What to enter
- Pearson's r (r), required
- p-value (p), required
- Sample size (N)
Common mistakes to avoid
- Saying the correlation shows that one variable causes the other.
- Reporting a non-significant correlation as proof that the variables are unrelated.
- Judging strength from the p-value instead of the coefficient.
- Leaving out the direction (positive or negative) or the sample size.
- Using Pearson for a curved relationship or with extreme outliers.
References awaiting academic review.
Spearman's rank correlation
Measures how consistently two variables rise or fall together, using ranks.
Learn more about interpreting Spearman's rank correlation
What to enter
- Spearman's rho (rₛ), required
- p-value (p), required
- Sample size (N)
Common mistakes to avoid
- Saying the correlation shows that one variable causes the other.
- Reporting a non-significant correlation as proof that the variables are unrelated.
- Judging strength from the p-value instead of the coefficient.
- Leaving out the direction (positive or negative) or the sample size.
- Describing Spearman's rho as measuring a straight-line relationship; it measures a consistently rising or falling one.
References awaiting academic review.
Simple linear regression
Predicts a numeric outcome from one variable.
Learn more about interpreting Simple linear regression
What to enter
- Unstandardised coefficient (B), required
- Standardised coefficient (β)
- R squared (R²)
- p-value (p), required
- Sample size (N)
Common mistakes to avoid
- Reading a coefficient as a causal effect when the design doesn't support it.
- Comparing unstandardised coefficients between predictors measured in different units.
- Reporting R² without saying whether the model is significant, or the reverse.
- Ignoring assumptions such as linearity and normally distributed residuals.
References awaiting academic review.
Multiple regression
Predicts a numeric outcome from several variables at once, each adjusted for the others.
Learn more about interpreting Multiple regression
What to enter
- R squared (R²), required
- Adjusted R squared (adj. R²)
- F statistic (F)
- p-value for the whole model (p), required
- Number of predictors (k)
- One predictor's standardised coefficient (β)
- That predictor's p-value (p)
- Sample size (N)
Common mistakes to avoid
- Reading a coefficient as a causal effect when the design doesn't support it.
- Comparing unstandardised coefficients between predictors measured in different units.
- Reporting R² without saying whether the model is significant, or the reverse.
- Ignoring assumptions such as linearity and normally distributed residuals.
- Ignoring multicollinearity: predictors that overlap strongly make individual coefficients unstable.
- Treating a significant model as meaning every predictor is significant.
References awaiting academic review.
Hierarchical regression
Adds predictors in stages, showing what each set adds beyond the ones before, such as control variables first.
Learn more about interpreting Hierarchical regression
What to enter
- Change in R squared (ΔR²), required
- p-value for the change (p), required
- F for the change (ΔF)
- Final R squared (R²)
Common mistakes to avoid
- Reading a coefficient as a causal effect when the design doesn't support it.
- Comparing unstandardised coefficients between predictors measured in different units.
- Reporting R² without saying whether the model is significant, or the reverse.
- Ignoring assumptions such as linearity and normally distributed residuals.
- Reporting only the final model and not the change in R² at each step.
- Choosing the order of entry after seeing the results.
References awaiting academic review.
Logistic regression
Predicts the chance of a two-category outcome, such as yes or no.
Learn more about interpreting Logistic regression
What to enter
- Odds ratio (OR), required
- Lower 95% confidence limit (CI lower)
- Upper 95% confidence limit (CI upper)
- p-value (p), required
- Nagelkerke R squared (R²N)
Common mistakes to avoid
- Reading an odds ratio as a risk ratio or a probability.
- Reporting an odds ratio without its confidence interval.
- Using R²-style measures as if they meant the same as in linear regression.
References awaiting academic review.
One-sample t-test
Compares the mean of a numeric outcome in one sample with a known or hypothesised value, such as a target or a published average.
Learn more about interpreting One-sample t-test
What to enter
- t statistic (t), required
- Degrees of freedom (df), required
- p-value (p), required
- Sample mean (M)
- Comparison value (μ₀)
- Standardized mean difference (d)
- Sample size (N)
Common mistakes to avoid
- Choosing the comparison value after looking at the data.
- Using a value taken from the same sample instead of a value specified in advance.
- Reporting t and p without the sample mean, comparison value and effect size.
References awaiting academic review.
Independent-samples t-test
Compares the means of a numeric outcome between two separate groups.
Learn more about interpreting Independent-samples t-test
What to enter
- t statistic (t), required
- Degrees of freedom (df), required
- p-value (p), required
- Mean of the first group (M₁)
- Mean of the second group (M₂)
- Cohen's d (d)
Common mistakes to avoid
- Reporting p without an effect size such as Cohen's d.
- Not saying which group, or which time, had the higher mean.
- Treating a non-significant result as proof that the groups are the same.
- Using it for paired data, such as the same people measured twice.
References awaiting academic review.
Paired-samples t-test
Compares the means of a numeric outcome measured twice on the same participants, such as before and after.
Learn more about interpreting Paired-samples t-test
What to enter
- t statistic (t), required
- Degrees of freedom (df), required
- p-value (p), required
- Mean at the first time or condition (M₁)
- Mean at the second time or condition (M₂)
- Cohen's d (d)
Common mistakes to avoid
- Reporting p without an effect size such as Cohen's d.
- Not saying which group, or which time, had the higher mean.
- Treating a non-significant result as proof that the groups are the same.
- Using it for two separate groups of people.
References awaiting academic review.
One-way ANOVA
Compares the means of a numeric outcome across three or more independent groups.
Learn more about interpreting One-way ANOVA
What to enter
- F statistic (F), required
- Between-groups degrees of freedom (df₁), required
- Within-groups degrees of freedom (df₂), required
- p-value (p), required
- Eta squared (η²)
Common mistakes to avoid
- Stopping at a significant F without post hoc tests to show which groups differ.
- Running many t-tests instead, which inflates the chance of a false positive.
- Reporting F without an effect size such as eta squared.
References awaiting academic review.
Two-way ANOVA
Compares means across two grouping factors at once, including whether they interact.
Learn more about interpreting Two-way ANOVA
What to enter
- F for the interaction (F), required
- p-value for the interaction (p), required
- Partial eta squared for the interaction (ηp²)
- F for the first factor (F)
- p-value for the first factor (p)
- F for the second factor (F)
- p-value for the second factor (p)
Common mistakes to avoid
- Interpreting main effects as if there were no interaction when the interaction is significant.
- Leaving out the interaction test.
- Reporting F without an effect size.
References awaiting academic review.
ANCOVA
Compares group means on a numeric outcome after adjusting for a numeric covariate, such as a control variable or a pretest score.
Learn more about interpreting ANCOVA
What to enter
- F statistic for the adjusted group effect (F), required
- Numerator degrees of freedom (df₁), required
- Denominator degrees of freedom (df₂), required
- p-value (p), required
- Partial eta squared (ηp²)
Common mistakes to avoid
- Reporting an adjusted group effect without describing the covariate.
- Ignoring whether the covariate-outcome relationship is similar across groups.
- Treating statistical adjustment as if it made non-equivalent groups randomized.
References awaiting academic review.
Chi-square test of independence
Tests whether two categorical variables are associated.
Learn more about interpreting Chi-square test of independence
What to enter
- Chi-square statistic (χ²), required
- Degrees of freedom (df), required
- p-value (p), required
- Sample size (N)
- Cramér's V (V)
- Smaller of rows − 1 and columns − 1 (k)
- Smallest expected count (E)
Common mistakes to avoid
- Using it when expected counts are below 5.
- Reading the size of chi-square as the strength of the association; use Cramér's V.
- Using counts from participants who appear in more than one cell.
References awaiting academic review.
Chi-square goodness-of-fit test
Tests whether the counts in one categorical variable's categories match expected proportions, such as equal shares or known population shares.
Learn more about interpreting Chi-square goodness-of-fit test
What to enter
- Chi-square statistic (χ²), required
- Degrees of freedom (df), required
- p-value (p), required
- Sample size (N)
- Cohen's w (w)
Common mistakes to avoid
- Confusing one variable's observed counts against expected proportions with a test of association between two variables.
- Choosing expected proportions after inspecting the observed counts.
- Reporting chi-square without the expected proportions and their source.
References awaiting academic review.
Factor analysis
Finds or confirms groups of items that measure the same underlying construct.
Learn more about interpreting Factor analysis
What to enter
- KMO measure (KMO), required
- Bartlett's test p-value (p), required
- Number of factors kept (factors)
- Total variance explained (%) (%)
- Lowest main loading (λ)
Common mistakes to avoid
- Keeping factors only because their eigenvalue is above 1, without checking the scree plot or parallel analysis.
- Ignoring items that load on several factors.
- Treating an exploratory result as confirmed without testing it in new data.
References awaiting academic review.
Structural equation modelling
Tests a whole model of relationships at once, often between constructs measured by several items each.
Learn more about interpreting Structural equation modelling
What to enter
- Comparative fit index (CFI), required
- Root mean square error of approximation (RMSEA), required
- Standardised root mean square residual (SRMR)
- Tucker–Lewis index (TLI)
- p-value of the model chi-square test (p)
Common mistakes to avoid
- Relying on the chi-square test alone, which is almost always significant with large samples.
- Treating good fit as proof that the model is correct; other models may fit as well.
- Changing the model to improve fit without a theoretical reason.
References awaiting academic review.
PLS-SEM
Estimates a structural model by maximising the variance explained in the outcomes.
Learn more about interpreting PLS-SEM
What to enter
- Path coefficient (β), required
- p-value (p), required
- R squared of the outcome (R²)
- Average variance extracted (AVE)
- Composite reliability (CR)
- Heterotrait–monotrait ratio (HTMT)
Common mistakes to avoid
- Reporting CB-SEM fit indices as if they applied to PLS-SEM.
- Skipping the measurement model (reliability and validity) before interpreting paths.
- Using p-values from methods other than bootstrapping.
References awaiting academic review.
Next: References
The sources you cite, kept together as you work.
How the assistant works
- The assistant reads your project draft: research question, objectives, hypotheses, variables, design, sampling, sample size and the analysis plan the Data Analysis Recommender makes from them. Enter them in the first step; later versions will bring them across from the other research tools.
- Enter the numbers your statistics software reported. The assistant checks they are possible, then explains them; it doesn't recalculate anything.
- Significance is judged against the level you choose. Effect sizes are labelled by the conventions researchers most often cite, which are rules of thumb.
- Each result is linked to the hypothesis it tests, the objectives and research question it serves, and its place in your analysis plan. A result supports or doesn't support a hypothesis; it never proves one.
What this tool can't do
- The assistant doesn't calculate statistics or read data. It explains numbers you enter from your statistics software, and can only be as right as those numbers.
- Effect size labels and cut-offs, such as Cohen's conventions and fit-index thresholds, are rules of thumb; what counts as large depends on your field.
- Significance is judged as p below your chosen level. It doesn't account for multiple testing: if you run many tests, adjust the level or the p-values first.
- Links to hypotheses, objectives and the research question are made by matching variable names; check each one.
- Wording is a starting point for your results chapter, not text to submit unchanged. Check it against your institution's style guide.
Awaiting review
- Statistical review: every interpretation, threshold and warning, by a statistician.
- Reporting style review: the academic sentences against APA and other common styles.
- References: sources for the effect size conventions and fit-index cut-offs, after academic review.
- Worked examples: published results checked against the interpretations.
Privacy
Everything stays in your browser. Nothing is sent anywhere or saved, and it is cleared when you leave the page. If you start a project in your research workspace, your work is also saved in this browser, and only this browser, until you delete the project.