The paper presents an attempt to work out a scale measuring religious attitudes. The object of the attitude investigated is the problem of the supernatural as understood by Christian religion. The scale was constructed with the view to measure one characteristic of religious attitude — intensity.In the process of construction were isolated first a number of statements having a distinctly discriminating value as regards positive or negative response to religion. That work was carried out by 70 competent judges. Only those (...) statements were accepted as differentiating which were ascribed extreme ranks (1—3 and 6—7) by the judges in an eleven-point continuum (Thurstone type). 30 statements were thus obtained: 17 positive and 13 negative towards religion. In order that the constructed scale might provide a valid measure of intensity a seven-point principle (Likert’s type) was adopted: I definitely agree, I agree, etc.The reliability of the scale wąs tested by the calculation of the coefficient of stability, the method applied being that of double investigation of the same population with a week’s interval between. The two sets of data were correlated by means of Pearson’s coofficient r for primary data. The index obtained was 0.979 with a standard deviation of 0.005 (for N = 70).The validity of the scale was tested by means of external criteria. Two methods were applied: a ’’scale” of religious practices, and a hierarchy of values. The "scale” of religious practices included: frequency of practices (prayers, mass attendance, communion), their motivation and meaning for the individual investigated. 219 individuals were tested with the religious scale and the ’’scale” of practice. The data were then correlated, yielding a coefficient (Pearson’s r) at the level of 0.669 with 0.031 standard deviation.The second method — grading of values — included the following items: one’s own life, property, fame, religious convictions and values, social good, man’s (woman’s) love, possibility of attaining higher learning, the good of one’s own family. The correlation between the two sets of data — religious scale and hierarchy of values — was measured by Spearman’s method of sequence correlation. The coefficient was 0.84 with a standard deviation of 0.02 (for N = 255).The population investigated consisted of graduates or individuals approaching the end of undergraduate studies. (shrink)