CHD4630 Exam Two o Key elements of research design 1 unit of analysis 04 19 2015 o level of social life a RQ is focused on o individuals groups and towns o specifically what we are interested in researching and drawing conclusions about o ex Are server liability laws based on local records related t the frequency of accidents caused by drunk driving assessed by local town police records The unit of analysis here towns o Ex Does efficacy defined as neighbors willingness to help each other affect neighborhood crime rates aggregated self reports Unit of observation cases in which measure are obtained data from individuals but describe neighborhoods o Causal conclusions from study should reflect unit of analysis in study Ecological fallacy Drawing incorrect conclusions about individual level processes from group level data Reductionist fallacy Incorrect conclusions about group level processes from individual data ex students who socialize more have lower grades so schools with more social engagement will have poorer student performance o reductionist fallacy 2 cross sectional and longitudinal designs o cross sectional all data collected at home point in time difficult to identify the time order of effects time order and cross sectional designs the independent variable is fixed at some point prior to the variation in the dependent variable ex Level of education and job performance we believe that respondents can give us reliable reports of what happened to them or what they thought at some earlier point in time our measures are based on records that contain information on cases in earlier periods ex CPS records and HS achievement we know that the value of the dependent variable was similar for all cases prior to the treatment ex English training program to improve English with recent immigrants nobody could speak English prior to the program data are collected at two or more points in time identification of time order of effects is easier 1 repeated cross sectional design trend study data collected at two or more points in time from different samples of the same population appropriate when we want to know how a population as a whole has changed over time not concerned with individuals ex RQ Has racial tolerance changed among Americans over a period of 10 years 2 Fixed sample design panel study Data collected from the same individuals the panel at two or more points in time 3 major difficulties o Very expensive o longitudinal o Attrition when people say they are over it and don t want to participate in your study anymore which is common in longitudinal studies study loses generalizability o Participant fatigue responses become much less reliable 3 Event based design cohort study data are collected at two or more points in time from individuals in a cohort share a common starting point o ex Birth seniority school can be a type of repeated cross sectional different sample or a panel design same sample 3 Quantitative vs Qualitative o Quantitative Causal Explanation Nomothetic causal explanation Belief that variation in an independent variable will be followed by variation in the dependent variable when all other things are equal The value of cases on the DV differs from what their value would have been in the absence of variation in the IV Ex IV is experience of abuse that variation predicts higher risk of using alcohol or drugs in emerging adulthood nomothetic causal explanation o 3 criteria for nomothetic causal explanations 1 empirical association Counterfactual IV Situation as it would have been in the absence of the idiographic causal explanation ex Adolescents going into emerging adulthood will have some type of risk to use alcohol or drugs counterfactual individual sequence of events Thoughts or actions that resulted in a particular outcome for a particular individual or that led to a particular event individualist or historic the notion of probabilistic relationship an average effect does not apply as it does with nomothetic explanations an empirical or observed association between the IV and DV true experiments we can determine whether and association exists between the IV and DV b c there are two or more groups that differed in terms of their value on the IV ex Physically maltreated groups and non maltreated group non experimental research whether values of cases that differ on the IV tend to differ in terms of the DV ex Cases that may range on maltreatment types and occurrences 2 appropriate time order the variation in the DV occurred after the variation in the IV our research design determines our ability to determine time order maternal depression time 1 child behavior problems time 1 3 nonspuriousness maternal depression time 1 child behavior problems time 2 true experiment researcher determine time order controlled environment to strengthen understanding of and likelihood of drawing causally valid conclusions identify causally mechanisms mediators specify context in which causal effects occurs moderators we say that a relationship between two variables is not spurious when it is not due to the variation in a third variable correlation does not prove causation an association between two variables might be caused by something other than an effect of the presumed independent variable on the dependent variable ex Storks deliver babies in Holland However the of storks and birth rates are higher in rural districts extraneous variable than urban districts extraneous variable to reduce risk of spuriousness o randomization part of a true experiment o statistical controls Chapter 7 8 Experiment and Survey Methods o Experiments Internal vs external validity o Internal validity The accuracy in concluding casual effect causal effect or casual validity Threats to internal validity History Testing uncontrolled outside influences on participants during an experiment ex news worthy event participants are expose to sensitization or learn something due to pretest that effects posttest develop bias from the pre test to the measure that effects the results of the post test instrumentation maturation biases due to testing procedures participants develop or change during the experiment as part of an ongoing process separate from the treatment regression the tendency for extreme scores high or low to move toward more typical performance when tested ex Tooth decay someone already has a tooth that is decayed to begin with overtime and the end of the study it will gravitate to just being worse ex