{"id":1015,"date":"2026-10-13T21:16:00","date_gmt":"2026-10-13T21:16:00","guid":{"rendered":"https:\/\/freestatisticaltools.com\/blog\/?p=1015"},"modified":"2026-10-13T21:16:00","modified_gmt":"2026-10-13T21:16:00","slug":"logistic-regression-odds-ratios","status":"publish","type":"post","link":"https:\/\/freestatisticaltools.com\/blog\/logistic-regression-odds-ratios\/","title":{"rendered":"Logistic Regression Explained: How to Interpret Odds Ratios"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\"><strong>Binary logistic regression<\/strong> predicts a yes\/no outcome: did the patient develop diabetes, did the customer churn, did the student pass? It is the go-to method when your dependent variable has two categories, and its results are usually reported as <strong>odds ratios<\/strong>, which many readers find confusing.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This guide explains odds and odds ratios in plain language, walks through a real-style output, and shows how to report it.<\/p>\n\n\n<div class=\"fst-tool\"><span class=\"ico\"><svg viewBox=\"0 0 24 24\" fill=\"none\" stroke=\"currentColor\" stroke-width=\"1.8\" stroke-linecap=\"round\" stroke-linejoin=\"round\" aria-hidden=\"true\"><path d=\"M3 21 21 3\"\/><circle cx=\"6\" cy=\"16\" r=\"1.5\"\/><circle cx=\"11\" cy=\"12\" r=\"1.5\"\/><circle cx=\"16\" cy=\"9\" r=\"1.5\"\/><\/svg><\/span><span><span class=\"t-eyebrow\">Free tool \u00b7 runs in your browser<\/span><a class=\"t-name\" href=\"https:\/\/freestatisticaltools.com\/logistic-regression-calculator\/\" style=\"text-decoration:none;color:inherit\">Binary Logistic Regression Calculator<\/a><span class=\"t-desc\">Free logistic regression calculator: odds ratios with 95% CI, Wald tests, Nagelkerke R\u00b2, Hosmer-Lemeshow test and classification table.<\/span><\/span><a class=\"btn accent\" href=\"https:\/\/freestatisticaltools.com\/logistic-regression-calculator\/\" data-tool=\"logistic-regression-calculator\">Run a logistic regression \u2192<\/a><\/div>\n\n\n\n<h2 class=\"wp-block-heading\">Why not ordinary regression for a yes\/no outcome?<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">If you code the outcome 0\/1 and run <a href=\"\/blog\/multiple-regression-interpretation\/\">multiple regression<\/a>, predicted values can fall below 0 or above 1, residuals cannot be normal, and their variance depends on the predicted value. Logistic regression solves this by modelling the <strong>log-odds<\/strong> of the outcome, which can take any value, and converting it back to a probability between 0 and 1.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Probability, odds and log-odds<\/h2>\n\n\n\n<figure class=\"wp-block-table\"><table><thead><tr><th>Probability (p)<\/th><th>Odds = p \/ (1 \u2212 p)<\/th><th>Log-odds = ln(odds)<\/th><\/tr><\/thead><tbody><tr><td>.20<\/td><td>0.25 (1 to 4)<\/td><td>\u22121.39<\/td><\/tr><tr><td>.50<\/td><td>1.00 (1 to 1)<\/td><td>0.00<\/td><\/tr><tr><td>.80<\/td><td>4.00 (4 to 1)<\/td><td>1.39<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">The model is linear on the log-odds scale:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><code>ln(p \/ (1 \u2212 p)) = b\u2080 + b\u2081X\u2081 + b\u2082X\u2082 + \u2026<\/code><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Each coefficient B is the change in <strong>log-odds<\/strong> for a one-unit increase in the predictor. Exponentiating it, Exp(B), gives the <strong>odds ratio (OR)<\/strong>.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">How to interpret an odds ratio<\/h2>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>OR = 1:<\/strong> the predictor makes no difference to the odds.<\/li>\n\n<li><strong>OR &gt; 1:<\/strong> higher values of the predictor increase the odds. OR = 1.12 means the odds are <strong>12% higher<\/strong> for each one-unit increase.<\/li>\n\n<li><strong>OR &lt; 1:<\/strong> higher values decrease the odds. OR = 0.55 means the odds are <strong>45% lower<\/strong> (1 \u2212 0.55 = 0.45).<\/li>\n<\/ul>\n\n\n\n<div class=\"fst-note warning\"><b>Odds are not probabilities<\/b><p>An odds ratio of 2 does <strong>not<\/strong> mean &quot;twice as likely&quot;. It means twice the <em>odds<\/em>. When the outcome is rare (below about 10%), odds ratios and risk ratios are similar; when it is common, the odds ratio exaggerates the risk ratio. Say &quot;odds&quot;, not &quot;likelihood&quot; or &quot;risk&quot;, in your write-up.<\/p><\/div>\n\n\n\n<h2 class=\"wp-block-heading\">Worked example<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">A public health researcher studies 500 adults and predicts type 2 diabetes (1 = yes, 0 = no) from <strong>age<\/strong> (years), <strong>BMI<\/strong> (kg\/m\u00b2) and <strong>physical activity<\/strong> (1 = meets guidelines, 0 = does not).<\/p>\n\n\n\n<figure class=\"wp-block-table\"><table><thead><tr><th>Predictor<\/th><th>B<\/th><th>SE<\/th><th>Wald<\/th><th>p<\/th><th>OR<\/th><th>95% CI for OR<\/th><\/tr><\/thead><tbody><tr><td>Age<\/td><td>0.039<\/td><td>0.010<\/td><td>15.21<\/td><td>&lt; .001<\/td><td>1.04<\/td><td>[1.02, 1.06]<\/td><\/tr><tr><td>BMI<\/td><td>0.113<\/td><td>0.025<\/td><td>20.43<\/td><td>&lt; .001<\/td><td>1.12<\/td><td>[1.07, 1.18]<\/td><\/tr><tr><td>Physically active<\/td><td>\u22120.598<\/td><td>0.220<\/td><td>7.39<\/td><td>.007<\/td><td>0.55<\/td><td>[0.36, 0.85]<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">Interpretation, holding the other predictors constant:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Age:<\/strong> each additional year increases the odds of diabetes by about 4%. Over 10 years, that compounds to 1.04\u00b9\u2070 \u2248 1.48, so 48% higher odds.<\/li>\n\n<li><strong>BMI:<\/strong> each additional BMI point increases the odds by 12%.<\/li>\n\n<li><strong>Physical activity:<\/strong> people who meet activity guidelines have 45% lower odds of diabetes than those who do not (OR = 0.55).<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">The 95% confidence intervals for the ORs do not include 1, so all three effects are significant.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Model fit statistics<\/h2>\n\n\n\n<figure class=\"wp-block-table\"><table><thead><tr><th>Statistic<\/th><th>Value<\/th><th>What it tells you<\/th><\/tr><\/thead><tbody><tr><td>Omnibus (likelihood ratio) test<\/td><td>\u03c7\u00b2(3) = 68.4, p &lt; .001<\/td><td>The model predicts better than an intercept-only model<\/td><\/tr><tr><td>Nagelkerke R\u00b2<\/td><td>.21<\/td><td>A pseudo-R\u00b2; a rough guide to explanatory power, not &quot;% of variance explained&quot;<\/td><\/tr><tr><td>Hosmer-Lemeshow test<\/td><td>\u03c7\u00b2(8) = 6.42, p = .60<\/td><td>A non-significant result suggests predicted probabilities match observed rates<\/td><\/tr><tr><td>Classification accuracy<\/td><td>78% (at a 0.5 cut-off)<\/td><td>Share of cases predicted correctly<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">Two warnings. First, pseudo-R\u00b2 values are typically much lower than linear-regression R\u00b2 values, so do not compare them directly. Second, classification accuracy depends on the cut-off and on how common the outcome is; if 80% of people are &quot;no&quot;, predicting &quot;no&quot; for everyone already gives 80%. For a fuller picture of discrimination, look at the <a href=\"\/roc-curve-auc-calculator\/\">ROC curve and AUC<\/a>.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Assumptions<\/h2>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>Binary outcome<\/strong> with independent observations.<\/li>\n\n<li><strong>Linearity of the logit:<\/strong> continuous predictors relate linearly to the log-odds. Check with the Box-Tidwell test or by adding a squared term.<\/li>\n\n<li><strong>No severe multicollinearity<\/strong> among predictors.<\/li>\n\n<li><strong>Enough events:<\/strong> a common rule is at least 10 events (cases in the smaller outcome category) per predictor. With 3 predictors you want at least 30 people with the outcome, and preferably more.<\/li>\n\n<li><strong>No complete separation:<\/strong> if a predictor perfectly predicts the outcome, coefficients and SEs explode.<\/li>\n<\/ol>\n\n\n\n<h2 class=\"wp-block-heading\">How to report logistic regression in APA 7<\/h2>\n\n\n\n<div class=\"fst-report\"><b>How to report it (APA 7)<button type=\"button\" data-copy-report>Copy<\/button><\/b><div><p>A binary logistic regression was performed to examine the effects of age, BMI and physical activity on the likelihood of type 2 diabetes. The model was statistically significant, \u03c7\u00b2(3) = 68.4, <i>p<\/i> &lt; .001, Nagelkerke <i>R<\/i>\u00b2 = .21, and the Hosmer-Lemeshow test indicated adequate fit, \u03c7\u00b2(8) = 6.42, <i>p<\/i> = .60. Older age (<i>OR<\/i> = 1.04, 95% CI [1.02, 1.06]) and higher BMI (<i>OR<\/i> = 1.12, 95% CI [1.07, 1.18]) were associated with higher odds of diabetes, whereas meeting physical activity guidelines was associated with lower odds (<i>OR<\/i> = 0.55, 95% CI [0.36, 0.85]).<\/p><\/div><\/div>\n\n\n\n<p class=\"wp-block-paragraph\">Present the coefficients table (B, SE, Wald, p, OR and 95% CI) as an APA table.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Related methods<\/h2>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Outcome with <strong>three or more unordered categories<\/strong>: multinomial logistic regression.<\/li>\n\n<li><strong>Ordered<\/strong> outcome (low\/medium\/high): ordinal logistic regression.<\/li>\n\n<li>You need <strong>marginal effects<\/strong> (changes in probability rather than odds): see <a href=\"\/probit-logit-marginal-effects\/\">probit and logit with average marginal effects<\/a>.<\/li>\n\n<li>Comparing two groups on a yes\/no outcome without other predictors: a <a href=\"\/blog\/chi-square-test-of-independence\/\">chi-square test<\/a> is enough.<\/li>\n<\/ul>\n\n\n\n<h2 class=\"wp-block-heading\">Frequently asked questions<\/h2>\n\n\n\n<div class=\"fst-faq\"><details><summary>How do you interpret an odds ratio less than 1?<\/summary><div class=\"a\"><p>An odds ratio below 1 means the odds of the outcome decrease as the predictor increases. Subtract it from 1 to get the percentage decrease: an OR of 0.55 means 45% lower odds.<\/p><\/div><\/details><details><summary>What is the difference between odds ratio and relative risk?<\/summary><div class=\"a\"><p>The odds ratio compares odds (p \/ (1 \u2212 p)); the relative risk compares probabilities. They are close when the outcome is rare, but the odds ratio is further from 1 than the relative risk when the outcome is common.<\/p><\/div><\/details><details><summary>What is a good Nagelkerke R squared?<\/summary><div class=\"a\"><p>There is no universal cut-off. Pseudo-R\u00b2 values in logistic regression are usually lower than R\u00b2 in linear regression; values of .20 to .40 often indicate a reasonably useful model.<\/p><\/div><\/details><details><summary>What does a significant Hosmer-Lemeshow test mean?<\/summary><div class=\"a\"><p>It suggests that the model&#39;s predicted probabilities do not match the observed outcome rates well. Note that with very large samples the test can flag trivial differences.<\/p><\/div><\/details><\/div>\n\n\n<div class=\"fst-report\"><b>Cite the tool (APA 7)<button type=\"button\" data-copy-report>Copy<\/button><\/b><div>Xavier, B. (2026). <i>Free Statistical Tools<\/i> [Computer software]. Zenodo. https:\/\/doi.org\/10.5281\/zenodo.23372149 (web app: https:\/\/freestatisticaltools.com\/logistic-regression-calculator\/)<\/div><small>Name the tool in your methods: \u201cThe analysis was carried out with the Binary Logistic Regression Calculator in Free Statistical Tools.\u201d<\/small><\/div>\n","protected":false},"excerpt":{"rendered":"<p>A plain-language guide to binary logistic regression, covering odds vs probability, what an odds ratio of 1.12 or 0.55 really means, model fit statistics and how to report the results.<\/p>\n","protected":false},"author":1,"featured_media":1016,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_fst_seo_title":"Logistic Regression: How to Interpret Odds Ratios (Examples)","_fst_meta_desc":"Understand binary logistic regression, from odds vs probability and log-odds coefficients to odds ratios, Nagelkerke R\u00b2, Hosmer-Lemeshow and APA reporting.","_fst_primary_tool":"logistic-regression-calculator","_fst_canonical":"","_fst_noindex":false,"_fst_no_toc":false,"_fst_no_autolink":false,"footnotes":""},"categories":[7],"tags":[44,42,43,38],"class_list":["post-1015","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-correlation-regression","tag-health-research","tag-logistic-regression","tag-odds-ratio","tag-regression"],"_links":{"self":[{"href":"https:\/\/freestatisticaltools.com\/blog\/wp-json\/wp\/v2\/posts\/1015","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/freestatisticaltools.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/freestatisticaltools.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/freestatisticaltools.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/freestatisticaltools.com\/blog\/wp-json\/wp\/v2\/comments?post=1015"}],"version-history":[{"count":0,"href":"https:\/\/freestatisticaltools.com\/blog\/wp-json\/wp\/v2\/posts\/1015\/revisions"}],"wp:attachment":[{"href":"https:\/\/freestatisticaltools.com\/blog\/wp-json\/wp\/v2\/media?parent=1015"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/freestatisticaltools.com\/blog\/wp-json\/wp\/v2\/categories?post=1015"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/freestatisticaltools.com\/blog\/wp-json\/wp\/v2\/tags?post=1015"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}