The negative symptoms of schizophrenia can determine functional outcome in patients. Despite its clinical significance, no treatment exists to date, as numerous pharmacological and non-pharmacological clinical trials have failed to demonstrate efficacy. Many of these trials evaluated negative symptoms as a single clinical construct. However, consistent evidence in the past two decades has found that negative symptoms constitute at least two independent clinical dimensions, namely deficits in motivation and pleasure (MAP) and in emotional expression (EXP). These dimensions are best evaluated using new assessment tools, such as the Brief Negative Symptom Scale (BNSS). However, older assessment tools, and particularly the Positive and Negative Syndrome Scale (PANSS), remain widely used in past and current research. Here, we sought to predict BNSS MAP and EXP dimensions from the PANSS. Using complementary modelling approaches across three heterogeneous, multi-centre, multi-culture patient samples (n = 1241 patients, 1846 observations), we show that MAP can be estimated (43–60 % variance explained) predominantly using N2 and N4. Moreover, EXP can be estimated predominantly using the two PANSS items N1 and N6 (55–81 % variance explained across models and samples). Additionally, PANSS-derived MAP shows associations with functioning similar to those measured by the BNSS MAP dimension. Together, our results suggest that while EXP can be reliably estimated from PANSS, MAP cannot be consistently estimated from PANSS across samples and cultures. This warrants caution when using the PANSS to estimate MAP and emphasises the need for using the newer assessment tools for negative symptoms.