Objective  To determine the value of patient‐reported symptoms in diagnosing endometriosis. Design  A national case–control study. Setting  Data from the UK General Practice Research Database for years 1992–2001. Sample  A total of 5540 women aged 15–55 years, diagnosed with endometriosis, each matched to four controls without endometriosis. Methods  Data were analysed to determine whether specific symptoms were highly indicative of endometriosis. Odds ratios for these symptoms were derived by conditional logistic regression analysis. Main outcome measures  Symptoms associated with endometriosis. Results  The prevalence of diagnosed endometriosis was 1.5%. A greater proportion of women with endometriosis had abdominopelvic pain, dysmenorrhoea or menorrhagia (73%) compared with controls (20%). Compared with controls, women with endometriosis had increased risks of abdominopelvic pain (OR 5.2 95% CI: 4.7–5.7), dysmenorrhoea (OR 8.1 95% CI: 7.2–9.3), menorrhagia (OR 4.0 95% CI: 3.5–4.5), subfertility (OR 8.2 95% CI: 6.9–9.9), dyspareunia and/or postcoital bleeding (OR 6.8 95% CI: 5.7–8.2), and ovarian cysts (OR 7.3 95% CI: 5.7–9.4), and of being diagnosed with irritable bowel syndrome (IBS) (OR 1.6 95% CI: 1.3–1.8) or pelvic inflammatory disease (OR 3.0 95% CI: 2.5–3.6). Women with endometriosis were also found to consult the doctor more frequently than the controls and were twice as likely to have time off work. Conclusions  Specific symptoms and frequent medical consultation are associated with endometriosis and appear useful in the diagnosis. Endometriosis may coexist with or be misdiagnosed as pelvic inflammatory disease or IBS.