1Developmental Research Methods and DesignJanuary, 2007Types of Data• Observation (lab v. natural)• Survey and Interview• Standardized test• Physiological measures• Case study• History recordResearch Methods in Aging1. Static vs. Dynamic Approach2. Within Group vs. Between Group Differences3. 3 factors influencing within group differences4. 3 fundamental effects in life span research5. 4 developmental study designs6. Experimental Study Designs2How to study Aging• Developmentally static approach–goal is to describepeople at a single, specific age period–age-group contrasts in a cross-sectional study–tends to show large age differencesApproach• Developmentally dynamic approach–goal is to understand the processof change–longitudinal designs which show age changesWithin-Group Differences vsBetween- Group Differences• Within-Group - Older adults show more heterogeneity (differences) than young children• Between-Group - Older adults are different from people at other stages of the lifespan3Why do older adults have higher quality friendships?• Do they BECOME higher quality friendships? – If so, why & how?• Were they always high quality friendships? – If so, why & how?3 Fundamental Effects in aging research•AGE EFFECTS• COHORT EFFECTS• TIME OF MEASUREMENT EFFECTSAge Change Effects• reflect differences due to biological, societal, and psychological changes• inherent changes within the person and are not caused by the passage of time, per se (as we age, we get taller; gray hair, wrinkles)4Cohort Effects• Age differences due to the historical period in which one was raisedTime of Measurement Effects• the social, historical, and political influences of the time period in which the data is collected The kinds of questions we ask influence the Research Methods we choose, which influence the type of data gathered and the conclusions that can be drawn.5Cross-sectional designs• A comparison of people varying in some specific variable (age, cohort, gender, race) at a single point in time• demonstrates group-level differencesExample• A cross-sectional design contrasting 20 and 40 yr olds in 1980Birth Cohort1980 --1960 Age 201940 Age 40Time of Measurement 1980Key Points: Cross-sectional• Are "quick"• Are relatively inexpensive• Controls for Age Differences (Time of Measurement), but cannot control for Age ChangeEffects or Cohort Differences6Longitudinal designs• a study in which a single cohort is followed over a period of time• demonstrates age changeExample• Longitudinal design of change from age 20 to 40 for people born in 1960Birth Cohort1980 --1960 Age 20 Age 40Time of Measurement 1980 2000Key Points: Longitudinal• Is expensive in terms of $, time, staff• Out-dated measures, ‘testing effect’• Drop-outs & Death• Controls for Age Changes, but cannot control for Age Differences (Time of Measurement) or Cohort Differences.7Time-lag designs• A comparison of people of the same age from different cohorts tested at different times of measurementExample• Compare 20 yr olds across cohorts and times of measurementBirth Cohort2000 Age 201980 -- Age 201960 Age 20Time of Measurement 1980 2000 2020Key Points: Time-lag• Very Expensive• Controls for Cohort Differences, cannot control for Age Change Effects or Age Differences (Time of Measurement)8Sequential designs – Key Points• Begins to disentangle the 3 confounds• Controls for Age Change Effects, Age Differences (Time of Measurement), and Cohort DifferencesExample• Simultaneous cross-sectional & longitudinal, depending on whether rows, columns, or diagonalsBirth Cohort2000 Age 201980 -- Age 20 Age 401960 Age 20 Age 40 Age 60Time of Measurement 1980 2000 2020So, do older adults have better quality friendships?•A cross-sectional study will show whether older adults are higher quality friendships than young adults. (time effects/age differences)•A longitudinal study will show friendships become higher in quality with age by following a specific cohort of young adults into late life. (age change effects)•A time-lag design will show whether current older adults have better quality friendships than previous cohorts of older adults. (cohort effects)•A sequential design will show whether or not older adults have higher quality friendships than younger adults at one point in time, over time, and across different cohorts (all three)9Experimental Studies• Hard to do experimental studies on aging…cannot randomly assign people an age. Hard to say that getting older, causes…• Often likely that age is a ‘proxy’ variable for something else going on in the body/mindExperimental Studies1. Age-comparative Studies– Match age groups on demographics– Randomly assign different age groups to different independent variable groups– Use a pretest and postest of dependent variable for each age group (before and after the independent variable)– Nonequivalent control group, pretest-posttestOld Group Old GroupYoung GroupYoung GroupPretestPosttestIntervention* Show older group gained more from the intervention than younger group. Not so much differences.Experimental Studies• Single Age Group Experiments– Only study young adult age groups over a period of time.– Different younger adult experimental & control groups• One group of young adults receives a pretest, receives intervention, and receives a posttest• One group of young adults receives a pretest, receives ‘standard care’, and receives a