PSYC1001 Lecture Notes - Lecture 6: External Validity, Experiment, Quasi
Document Summary
Science and statistics: everything founded on science, science is an attitude continually asks questions, collection of unchanging truths, non-overlapping magisterial. Interpreted stories about a single occurrence in past no scientific value: build bias, change with each retelling. Case studies: all details, whether or not they seem relevant are recorded, objective nature of note in case studies mean alternative explanations are possible later. Iv thing that is manipulated by experiment. True experiment: all iv"s are (cid:272)o(cid:374)trolled a(cid:374)d a(cid:271)le to (cid:271)e ra(cid:374)do(cid:373)ly allo(cid:272)ated, strong causal inference due to random allocation, all iv randomly allocated. If difference in dv found, conclude that differences in iv caused it. Quasi experiment: one key variable of interest can be randomly allocated eg gender, pre-existing groups cause weaknesses in causal inference. If differe(cid:374)(cid:272)e i(cid:374) dv, (cid:272)a(cid:374)"t say (cid:449)ith (cid:272)ertai(cid:374)ty (cid:449)hat (cid:272)aused it: categorisation of this type uses matching or involves random allocation of some variables but cant for others, attempt to control one iv by matching.