PSYCX1015 Lecture Notes - Lecture 9: Internal Validity, Repeated Measures Design, Null Hypothesis
Document Summary
Construct validity must be intact for this to matter. Internal validity must be strong in an experiment. The extent to which we can generalize the findings from our study (sample) to the outside world (population) It would be impossible to have perfect/absolute external validity. Instead, we must decide either before or after an experiment how widely we can interpret our data. How relevant our findings are to the real world". How closely our laboratory study mimics real life. To do that accurately, you can only make predictions within the range of values you have studied. The relationship might change in other parts of the range. Internal validity demands control over conditions and variability. Type 1 error: falsely stating that what we saw in the lab will exist outside the lab ( rejecting the null hypothesis) Acceptable level is alpha, 5% (. 05- the p value, )