Negative binomial regression was ap

Negative binomial regression was applied to the total number of serious offenses participants reported engaging in between Waves 3 and 4 because the dependent variable was a count measure, and the distribution was overdispersed (α = 5.05, p < .001). Although indirect effects and bootstrapping cannot be computed with negative binomial regression, coefficients for the two legs of the cross-lagged target and comparison effects were calculated. The results indicated that while both legs of the target effect (Peer-1 → PCT-2 and PCT-2 → Offending) were significant, only the PCT-1 → Peer-2 leg of the comparison effect was significant.

結果 (繁體中文) 1:

負二項回歸應用於嚴重犯罪參與者的總數報告波3和4之間接合在由於因變量是一個計數度量，並且分佈overdispersed（α= 5.05，P <0.001）。儘管間接效應和自舉不能與負二項回歸來計算，分別計算的交叉滯後目標和對比效果的兩條腿係數。結果表明，雖然目標效果（對1→PCT-2和PCT-2→違犯的）兩個腿是顯著，只的比較效應的PCT-1→同行2腿是顯著。

結果 (繁體中文) 2:

負二項回歸應用於報告在波3和4之間參與的嚴重犯罪參與者的總數，因為因變數是計數度量值，並且分佈過度分散（= = 5.05，p = .001）。雖然不能用負二項回歸計算間接效應和引導，但計算了交叉滯後目標兩條腿的係數和比較效應。結果表明，雖然目標效應的兩條腿（Peer-1 = PCT-2和PCT-2 = 犯罪）顯著，但只有PCT-1= Peer-2段的比較效果顯著。

結果 (繁體中文) 3:

由於因變數是一個計數名額，且分佈被過度分散（α=5.05，P<0.001），囙此對參與者報告在第3波和第4波之間參與的嚴重犯罪的總數應用負二項回歸。雖然不能用負二項回歸計算間接效應和自舉效應，但計算了交叉滯後目標兩條腿的係數和比較效應。結果表明，兩條腿的目標效應（peer-1→pct-2和pct-2→冒犯）均顯著，但只有pct-1→peer-2腿的比較效應顯著。