Tieteellistä tekstilajia esiteltynä

Koulumme alumni Jyrki Korpua teki väitöskirjan Tolkienin maailmasta. Lue se täällä.



Alla esimerkki tieteellisestä tekstistä. Paljon vilisee passiivia, lauseenvastikkeita ja yläkäsitteitä.

Data Analyses
First, to test the hypothesis that already in kindergarten and first-grade elementary school conduct problems predict the development of depressive symptoms, an autoregressive lagged model (Jo¨reskog, 1970) was
used. In the autoregressive paths, variables were regressed on their immediate prior values, to study stability within processes. Cross-sectional correlations between conduct problems and depressive symptoms were included in the models. To study possible cross-over effects, cross-time paths from conduct problems to depressive symptoms (and vice versa) were allowed for. Significant cross-lagged paths from conduct problems to depressive symptoms would reflect the anticipated prediction of depressive symptoms by prior levels of conduct problems, while controlling for stability within and concurrent links between these domains of psychopathology.

After having studied the direction of effect between conduct problems and depressive symptoms, the hypothesis on the connecting role of poor social preference in this longitudinal link was tested using triple latent growth models. The development of each construct was represented by two latent growth parameters, an intercept (centered at the first assessment) to assess initial level differences and a linear slope, which represents growth over the repeated assessments. Given our focus on developmental links, we focused on slope associations, although links between intercepts were allowed for. A test for the joint significance of the two directional paths between the slopes (MacKinnon, Lockwood, & Williams, 2004) was used to test for the significance of the indirect path (cascade effect) from conduct problems to depressive symptoms via poor social preference. To test for possible sex differences, multiple group models were used. A model in which the path estimates were freely estimated across gender was compared to a model in which the path estimates were constrained to be equal across boys and girls using a chi-square difference test. All models were fitted using Mplus 5.1 (Muthe´n & Muthe´n, 1998–2009). Model fit was determined through the comparative fit index (CFI; values .90; Bollen & Long, 1993), the Tucker–Lewis index (TLI; values .90; Bollen & Long, 1993), and the root mean square error of approximation (RMSEA; values .08; Browne & Cudeck, 1992). As data were nested within classes, standard errors were adjusted using a sandwich estimator (Williams, 2000). Missing data were handled through Full Information Maximum Likelihood

Leave a Reply