Das Bild zeigt rechts ein Trichterglas mit blauer Flüssigkeit und Glasstab zum Umrühren. Daneben steht ein Reagenzglasständer mit Reagenzgläsern, die ebenfalls blaue Flüssigkeit enthalten. Im Hintergrund ist ein Forscher zu sehen, der in der rechten Hand eine Pipette hält.

GendAge

You are here:

Abstract

Sex and gender are distinct constructs that require specific measurements. As intercultural differences may be an important factor in capturing the construct "gender", we aimed to re-construct a gender score comparable to the index published by L Pilote (APS 2015) in a retrospective analysis of baseline data of Berlin Aging Study (BASE-II). 

We used data obtained in 2009-2014 of 1870 participants aged 60 years or above. From the psychosocial and gender-related variables we selected 33 variables for reconstruction of a gender score. Principal component analysis (PCA) was performed to reduce the no. of variables. Logistic regression models (LRA) were calculated, and non-significant variables were removed one by one in a descending order of their p-value. Coefficient estimates of these variables associated with sex were used to calculate gender score on a scale of 0 to 100.

At first a correlation analysis of 33 selected variables was conducted and in each significantly correlated pair one of the two variables were randomly removed. In total, seven variables were removed in this step. In the second step of PCA 26 variables were used. Five variables with too few cases were removed and remaining 21 variables loaded onto nine components. One variable, loading equally high on two components was also removed. The 20 identified variables were used to perform the LRA. After calculating 4 models 7 most significant variables for sex as an outcome variable were extracted (PSS composite, Affection, Big Five-Neuroticism, UCLA composite, Hours per day on housework, Family status, and Big Five-Extraversion.

The coefficients of the final LRA were used to calculate the gender score (0-100). Gender score distribution was significantly different between women and men. The procedure enabled us to construct a gender score in a retrospective. In future, we will examine the validity of the score by linking differences in gender with psychosocial, medical, and cognitive data and by comparison with a prospectively assessed gender score.