AP Psych Unit 1 - Research Methods

hindsight bias
Click the card to flip 👆
1 / 35
Terms in this set (35)
used to cut down on bias

2 parts: a precise statement of the experimenter's procedures and concepts and something that is measured numerically

a good OD should be detailed enough to enable other scientists to replicate the experiment if they wanted -> this way other scientists can support or debunk hypothesis
naturalistic observationwatching a person or animal behave in its normal surroundings think of a person squatting in a jungle taking notes on monkeys the idea is to interfere or be as little intrusive as possible so the subject behaves normally, then record only what is observed purpose: gather info strengths: handle large numbers of ppl and data weakness: doesnt show causationcorrelationwhen 2 things are related or they go together, they are correlatedcorrelation coefficientmeasure how closely two things go together and is seen either numerically or in scatterplots numerically: scale or one to negative one with pos 1 representing most correlated and neg 1 being most inversely correlated and 0 is not correlation at all scatterplots: bottom left to top right line = positively correlated top left to bottom right = negatively correlated or inverselycorrelation and causationjust bec 2 things correlated does not mean one causes the other, cant prove things from correlation -> could be another factor involved only an experiment shows causation because it isolates one variable to be tested ex: a persons shoe size and vocabulary correlated highly but it is not causationillusionary correlationsppl see correlations that are not there when 2 things happen randomly we link them together however we always remember the times things work and forget the times they do not (when we wrongly conclude the correlate)experimentationthe gold-standard research method is the experiment. all other methods have their pluses but experiments alone show cause and effect purpose: to find cause and effect strengths: shows cause and effect weaknesses: costly, ethical factors may make experiments more impracticalparts of a good experimentrandom selection, random assignment, double-blind procedure, placebo effect, independent, dependent, and confounding variablerandom selectionThe participants come from a large population and are randomly selected to be involved.random assignmentonce chosen to be in the study, participants are randomly assigned to either the control or experimental groupdouble-blind proceduretechnique in which neither the persons involved for those conducting the experiment know in what group to participate is involved cuts down biasplacebo effectppl are often given a placebo in double-blind experiments; tho fake, they believe its real and have real positive benefitsindependent variablewhat the experimenter manipulates. the only thing diff between the experimental and control groupsdependent variablewhat the independent variable supposedly affects -> what is measuredconfounding variablesother factors that might make the experiment go wrong. ones that might effect the DV. we want an experiment where only the IV effects the DV or else we wont know what caused the effect. if the factor occurs the results are confounded random selection, assignment, and good procedures filter out confounding variablescentral tendencyrefers to the center of a bunch of numbers. there are 3 usual measurements of central tendency: mean (average), mode (most frequent number), and median (middle number)measures of variationrange, standard deviation, normal/bell curverangeDistance between highest and lowest scores in a set of data.standard deviationmeasurement of how much the numbers vary from the mean, if all numbers are close the SD will be low and if there is a wide range it will be highnormal/bell curvea graph that often occurs in nature with things like height and intelligent scores on testsvaliditya test or bit of research measures what it's supposed to measure ex: when taking a math test it covers algebra and not sciencereliabilitythe test yields the same results over and overstatistical significancethat the observed difference is probably not due to chance. it is measured by the p value and goes by a 5% rule. the numbers are statistically significant in there is less than a 5% chance they were caused by chancepsychology appliedresearch in a laboratory can test principles that can be applied to everyday life culture comes into play during research; a study in one culture may not be able to be applied to anotheranimal researchscientists study animals bec: their biological systems are similar to humans, life-cycles are short so its faster to study life spans, and bec of ethic to experimenting on humans but controversial unoffical ranking of animals: humans, primates and household pets, mammals, birds, fish, reptiles, insectshuman researchEthical considerations must be made the APA suggest: informed consent and debriefing