PSYC217 Chapter 14 - Generalizability Intro - A single study is conducted with a particular sample and in a particular context. - The external validity of that study is the extent to which the results can be generalized beyond that study and sample to other populations and settings. Issue of external validity is complex, raising deep questions about usefulness of the knowledge we create when doing research. Generalizing Results - Replication - Replication: repeating a study to see if one observes the same result, to increase confidence in that result or demonstrate that it is not systematically observable. · Important to overcome some questions of generalization that stem from results of a single study. · Two types of replications to consider: direct and conceptual replications Direct Replication - Direct replication: attempted replication of a study following the same procedures that were used in the original research as closely as possible. " In best practice, individual researchers attempt to directly replicate their work when possible. · Crucial for determining whether an original finding can generalize to other samples drawn from the same population (e.g., undergraduates at an institution) · To prove an initial result was not a Type I error (is an effect, when no effect is present) Researchers may engage in direct replication when starting new research. - · May replicate studies to make sure you understand the procedures and obtain the same results. . Researchers attempt a direct replication and pair it with a novel study to builds on findings replication and extension > " When you can replicate an original research finding using the same or similar procedures, confidence in generality of original findings is increased. - A researcher may be unable to replicate previous findings. " A single failure to replicate does not always mean the original phenomenon does not truly exist, · Failure to replicate share the same difficulties of interpretation as statistically non-significant results -> could mean original results are invalid or replication attempt was flawed. · A well-designed replication with a large sample failing to detect an effect should be considered evidence that the effect may not exist. o In some cases, the effect may only occur under certain, limited circumstances. o May turn out the original researcher misrepresented results or engaged in questionable research practices to produce false evidence for a phenomenon (scientific fraud) - "Mozart effect" > proves importance of replications. " Original study -> university students listened to ten minutes of a Mozart sonata then exhibited better performance on a spatial-reasoning task compared to controls · Replication study > tested effect using different measure of spatial ability o Effect was temporary (ten minutes) o Generalized Mozart sonatas could increase a child's intelligence when played " Next few years -> many failures to replicate Mozart effect o Meta-analysis (3,000 participants) concluded a small effect of listening to Mozart on spatial reasoning compared to not listening to his music o Listening to music was found to be an effect of equal size >