3003PSY Week 11 Mini-Transcript (2) PDF

SPEAKER 0 Welcome back to the mini Lecter Syriza three by three ps Y I'm Dr Natalie Luxton and in his mini SPEAKER 1 lecture, we look at the last longitudinal designs, the cross lighting design, and then we'll look at some of the assumptions. SPEAKER 0 Certainly there's always assumptions in the b...

SPEAKER 0 Welcome back to the mini Lecter Syriza three by three ps Y I'm Dr Natalie Luxton and in his mini SPEAKER 1 lecture, we look at the last longitudinal designs, the cross lighting design, and then we'll look at some of the assumptions. SPEAKER 0 Certainly there's always assumptions in the breeze. SPEAKER 1 Many lectures We spoke about the simplex design, which tests but stability and change of the time. The Longitudinal Correlation, which attempts to test the temple president. SPEAKER 0 And the residual longitudinal regression, which allows the test of both stability and design and a large part temporal presidents. One of the limitations of the residual regression, though, is that we can only test for by very only test SPEAKER 1 for uni directional effects. In the residual regression, only the dependent variable is measured at bo time points, and the ideas only measured at one time point. Also in the residual design I need, the dependent variable SPEAKER 0 is test is the criteria, and the idea is not tested as a potential outcome, And this announcements the cross SPEAKER 1 lake announces there is more than one dependent variable. We are simply mimicking the I've twice and adding this to the model you cross legged models. The two measures serve as both the independent independent variable. SPEAKER 0 To avoid confusion, we'll just refer to variable one invariable to rather than IV's and deles cross like models. Effectively combined to residual is long internal regressions into the SPEAKER 1 one analysis. SPEAKER 0 As a result, the stability of the constructors against assisted Lee removed alive each variable to predict the change in the other variable. The big advantage here is that this approach allows an investigation of bi directional fix. SPEAKER 1 Just note that we will end up having six parameters in a two way across Lake Model Two Way simply SPEAKER 0 refers to two time points. These families include the correlations between the variables a time SPEAKER 1 one and the correlation between the variables of time, too. SPEAKER 0 It also includes the testes stability for both variables, most importantly, to cross their path. Do these model seemed familiar. Now the top model is the test of variable to predicting changing variable one, and the bottom model is the test of variable one predicting changing variable too. So these air are six parameter okay, but she was an example to teach that little part in this study. 1000 6 adolescents were followed over two years, the researchers SPEAKER 1 were interested in the law maternal relationship between binge drinking and sharing alcohol content on social networking sites. This model investigates the direction of the relationship between sharing alcohol content and been streaky. SPEAKER 0 If one of the terrible predicts change in the other over time, then we could say that we have a unit drink to beauty, directional relations and therefore we can assert Temple President. If both variables predict the change in each other, there is evidence of a bidirectional relationship. In other words, temperatures cannot be inferred. Without Temple president, it's impossible to make a conclusion regarding cause ality. SPEAKER 1 This is what they found. SPEAKER 0 Have any included the parameters related to the stability of SPEAKER 1 the constructs in the crossed legs? SPEAKER 0 We're not a particular interest to the research question. We can see that there are significant water. Regressive associations recall that these air the associations across time for each variable and captures the stability of the construct here on that have been streaking at timeline is correlated with been shrinking two years later. Similarly, the amount of sharing about call references a tight SPEAKER 1 one is correlated time, too. SPEAKER 0 There is unique variance that is residual in each variable a time, too, after controlling for schools that Taiwan, looking SPEAKER 1 at the path and binge drinking at time, want to sharing of alcohol references a tired too. SPEAKER 0 There is significant explanation off the change in sherry alcohol references by area binge drinking. SPEAKER 1 Likewise, sharing that alcohol references a tight one is significantly predicting change in binge drinking time, too, as both cross SPEAKER 0 later. Significant then we could say that we have bidirectional effects, SPEAKER 1 but it's been shrinking and share. Have alcohol. SPEAKER 0 References in social media in early adolescence are predictive of sharing alcohol references and been shrinking two years later. The great advantage of this type of design is that it estimates stability and change. It predicts. Change accounts for so cross sectional associations as well as both uni directional and bidirectional relationships. This table summarises the different estimates across the four designs. As you can tell, the cross leg is the best of all before the best of all fourth, but it's the most expensive to run and requires additional software beyond the SS. They're also more advanced approaches that are well beyond the scope of this course. Finally, as always, their assumptions of longitudinal analysis, the firsts in the individual stability. SPEAKER 1 This techniques examined stability and change in the sample over SPEAKER 0 time. SPEAKER 1 Thus, it's assume that there is no systematic differences in the stability and change between us. SPEAKER 0 Thiss assumptions violate when one group of participants and just faster or slower relative to the other participants. For example, if I was interested in tracking the stability and change of cumulative wealth from the age of 50 18 to 50 those who have more money at the age of 18, perhaps they've been giving a small light of a million dollars from their parents. I can increase the wealth faster because they hold an advantage. SPEAKER 1 The rich get richer. SPEAKER 0 Observation makes it difficult to examine stability and change and wealthy using these techniques. If so, they're more advanced approaches a quiet. The second assumption is consistent measurement. SPEAKER 1 It is assumed that the measurement is saying when using repeated measures. SPEAKER 0 In practical terms, this means that the items administered to need to be exactly the same across time periods. In perceptual terms, participants need to read interpret the questions SPEAKER 1 in exactly the same way across time, often in longitudinal SPEAKER 0 research, researchers will examine the validity and reliability of their stealth after an initial time point and then adapt the scout have better psychometric properties. In this instance, I need the items that have been employed across all time periods can be used. SPEAKER 1 This is why it's important to pilot measures prior to SPEAKER 0 a long, costly data collection process. They need to be synchronicity. SPEAKER 1 This means that it's a soon assumed that the administration off question is because at the same interval between time periods for all participants, ideally, if researchers are examining developmental SPEAKER 0 outcomes and feel like adolescents across the school, year's would be good if the measures were administered exactly 12 months apart. For all students, however, developmental psychologists are at the mercy of school principals about when and where they're allowed onto school grounds to complete the research. Credible as such. For Steve, it was like to be variation in the number of days between the measurements. No student is going to be exactly 360 days apart, and they'll also be variations between the participants. SPEAKER 1 The fourth assumption is timeframe. In order to find a long tune or fix the left, the time between measurements needs to be considered the SPEAKER 0 time frame, maybe to talk too short for a variable to impact another, and so you won't find. In effect. SPEAKER 1 The time frame may also be too long, and the impact of one variable another may have dissipated in general, the greater the temple distance between measurements. SPEAKER 0 The week of the relationships This point is important because SPEAKER 1 most longitudinal states only conducted here. SPEAKER 0 Lee follow up yearly. Finally, we need to consider other potential variables. It's important to ensure that important variables are not emitted from the study. This idea is related to a cause of process or pathway. This is actually covered in honours. For this reason, it's easy to say that a variable time want plants is or impact something rather saying that it causes if the goal of the long tunnel research SPEAKER 1 is too unfair, cause ality living at a third variable SPEAKER 0 makes a substantial impact on your interpretation. SPEAKER 1 For example, originally established finding is that time special social SPEAKER 0 networking sites is related to lower levels of well being. But the storey is not that simple. For example, what time that was spent on social networking sites is related to mostly stances, which is which, in turn is related to lower levels of well being. Ignoring that variable can have impacts on the conclusions of your research. SPEAKER 1 To summarise across the design is the gold standard of panel designs. This design overcomes the limitations in the previous designs by testing for uni directional and bi directional relationships to waive cross legged design. Has six parameters cross sectional associations between the variables? A time winning time, too? The association's over time for each variable, which represents stability and change, and the cross leg past reached variable a time. One predicting the other a time too belonged to my design has several assumptions, including into individual stability, consistent measurement and synchronicity. Other considerations. Longitudinal designs include the length of the time frame and potential other third variables.