Estimate the errors involved in Exercise 47, parts (a) and (b). How large should be in each case to guarantee an error of less than 0.00001?
EXPERIMENTS ● Non-experimental method = do not involve manipulation of an independent research ○ Descriptive research = describes how people think, feel, and behave ● Correlational research = assess whether two measured variables are related to one another ○ Ex. happiness is positively associated with work productivity ○ Provides clues for carrying out experiments that examine possible causal effects ● The researcher designs participants to conditions and manipulates at least one independent variable ● Vary as least one independent variable to assess its effects on participants’ responses Independent Variable ● Thought to influence or cause the behavior ● Has levels ○ Conditions, groups, or categories within the IV ○ Must have at least 2 levels ● Can be discrete or continuous ● Discrete = categorical ○ Qualitative ● Continuous = numerical scale ○ Quantitative differences ● Environmental manipulations = changing the physical or social environment ○ Ex. noise (loud vs quiet) ○ Ex. are people’s memories affected by false info during discussion with another person ● Instructional manipulations = changing the instruction given to one group over another ● Invasive manipulations = creating physical changes in the participants ○ Physical stimulation ○ Drugs Dependent Variables ● Variables the researcher measures ● What has been influenced by the IV ● Typically involve either observations of actual behavior or measures of physiological reactions ● Self-reports (surveys) = closed or open ended behavior ○ Reaction time ○ Frequency of behavior ○ Test scores ● Neurological measures ○ Functional magnetic resonance imaging ● Physiological measures ○ Galvanic skin response ○ Cortisol ○ Facial electromyography ● The milkshake experiment = people’s belief in the kind of shake they consumed affected their hunger levels, even though the same shake was consumed by both groups ● Between-subjects design = each experimental condition has different people ○ Randomized ● Within-subjects design = a single group of people serves in all conditions of the experiment ○ Repeated measures Assigning Participants ● You want to ensure that participants do not differ from one another on various dimensions ○ If groups differ, it’s possible that observed differences in conditions are due to different groups rather than the IV ● Simple random assignment = the groups should end up being equal in most characteristics ● Matched random assignment = matching people on a characteristic that is relevant to the outcome and then assigning one person to one group and another to the other ○ Makes sure both groups have participants who are matched on an important trait ○ Usually requires a pretest ● Control all extraneous variables that may influence participants responses to determine causality ● Extraneous variables = other factors separate from the IV that may affect the behavior of DV and could serve as an alternative explanation for the changes in the behavior ○ You want to minimize or control these Internal Validity ● The degree to which differences in performance can be attributed to an effect of an independent variable ● Involves the relation between the IV and DV ● Your goal is to control for all extraneous factors that might compromise this relation ● An experiment is internally valid when it eliminates all potential sources of extraneous variables ● Nothing can differ between conditions other than the IV Confounding Variable ● An extraneous variable that the researcher unintentionally varies along with the independent variable ● Can seriously undermine the internal validity of the study