Divorce: • How does it affect children? Methods Matter Direct experience matters, but people have different experiences • Often we don’t just need to know what’s best for us, but what’s best for other people too What if we want to give advice, or set policy? • Lots of funding for finding better relationships especially in low-income homes What tools have we got for sorting out different versions of the truth about intimate relationships? • How can we figure out how to choose between all the contradictory advice about relationships? • Relationship researchers have developed a specific set of tools: • What do we measure? Relationship science measures constructs Construct: • An abstract idea (love, commitment, satisfaction) • Not directly measurable, constructed by humans • Not imaginary, but an abstract idea (exists, but difficult to measure) Operationalization: • The process of translating an abstract idea into concrete terms so it can be studied • We cannot measure constructs, but we can measure operationalizations of a construct How can we measure Love? Strategy 1: Self-reports • Very common • Operationalization: “the love scale” – numerical scores (one possible form of operationalization) • There are many ways to operationalize a construct Some things you just have to ask about Strategy 2: Systematic Observation • What should we observe? o Verbal communication  People in love: more “we/us” statements, people less in love: “me/I” o Non-verbal communication o Written communicationo Biological responses  Measure active brain areas  Parts of brains that light up are the same that light up for drug addicts (strong sensors that make us addicted)  Brains of older people in love looks the same as younger ones in love (50 yrs old vs. 5 weeks old couple – same brain activity) • Pros and Cons of Self-Reports: Pros: • Easy to do • Cheap, quick • Sometimes gets at exactly what we want to know Cons: • Don’t always tell the truth (violent behavior) • Don’t always know the truth Study: What behaviors are listed good relationships? Asked both partners the same questions for behaviors during the last 24 hours • Unhappy relationships: reported lots of negative behaviors and not a lot of positive behaviors • Spouses disagreed about what kind of behaviors they engaged in in the last 24 hours • Interpretation of lack of agreement about events in last 24 hours: o People don’t know (they are not lying, they simply don’t know/remember) o Little ability to remember mundane behavior • Assume certain behaviors more/less common according to whether they felt they are in a good or bad relationship o Happy people reported pos behaviors – sentimate override  When we don’t know the answer to a specific question, we use our global sentimate to override (= sentimate override) • Pros and Cons of Observation Pros: • Very relevant (if you choose well) • Can watch verbal + non-verbal communication • But can’t watch everything (infidelity, violence) • Great detail• Ability to define things Cons: • Very expensive? • Cant do it on the internet in general, need to pay people to watch + observe • Reactivity • Does watching a behavior change the behavior? • People act differently when being watched • Difficult to ‘fake’ a conversation: o Some couples don’t know how awful they are (even though they are being taped) • What we see in the lab is a conservative estimate of what people really do (act more worse at home) • The meaning problem • “you are such a jerk” can be intended as malicious but also cute + teasing • Can be very ambiguous – observer cant always tell what a behavior means • “sorry” is not always genuine (tone of voice plays an important role, as well as other non-verbal cues) ⇒ The more precise we are, the easier it is to observe and measure something Designing Research Cross-sectional research: • Helps with description (good if we want to describe something) • Looks at something once Longitudinal research: • Helps with predictions • Don’t just describe but helps with seeing how something changes or making predictions • Basically a bunch of cross-sectional studies of the same people over time Cross-sectional vs. Longitudinal • The case of the disappearing curve o How does marital satisfaction change over time? • First: cross-sectional studies (many samples at one point in time) o How long have you been married? How happy are you? o ⇒ An inverted curve: start out happy, then fall, then rise again with number of years married = “the U-curve” o Children put stress on relationship in middle years (down part of U-curve) o But the U-curve measured different couples, not over time → thus, lower part of U-curve could be those that just got divorced• ⇒ Longitudinal study is the only study that can find the causes o Marital satisfaction declines over time (downward linear, no more U-curve) o ⇒ Different results according to how research is conducted! • Pick a research design appropriate for question Who are we Studying? Intimate relationships may differ for different groups of people • American – Chinese (we study more Americans – Americans resolve conflicts differently. Chinese express more negativity. Can’t generalize to other countries). • Rich – poor (SES) • Young – old • Dating – married…etc • Every culture wants good partners • Bias of samples who participate in studies o Ex) volunteer effect engages extremes (very happy or very unhappy couples participate in studies) • Satisfaction or resources? → which matters more in staying with partner? o ⇒ Satisfaction matters more Studying individuals or couples? Is this Ethical? The Power of Mere Questions • Just asking a question can humiliate a person (sexual questions) • Be sensitive to audience • Successively advance to question: how often to you yell? Throw stuff? Hit partner? The Effects of Being Studied • People questioned about their relationship become more of what their relationship is • If you are asked if you are happy you are reminded that you are happy – better • Good couples become better, bad couples become worse Costs vs. Benefits • Is it worth asking a sensitive question? Ethical Obligations • Be sensitive, protect anonymity +