58 | Chapter 12 a value of “1” in the year preceding the event (the “anticipatory year”). Next to these anticipatory effects, we also took “knock-on” effects into consideration, as life events may become influential only after a certain amount of time has elapsed. Thus, we also assigned a value “1” to the year after a life event took place (the “knock-on year”). We assigned a value of “0” to all other years. This procedure also accounted for a certain degree of uncertainty (i.e., recall bias) in the answering of the retrospective questions. We included several control variables in the analyses. Because research has shown that men (1) and women (0) often choose different organisational forms (Tiessen-Raaphorst et al., 2010), we considered this as the first control variable. Furthermore, the elderly are less apt to participate in sport due to the physical restrictions they experience (Van den Dool et al., 2009). Therefore, we included age as a continuous variable, ranging from 18 up to and including 65. We introduced the continuous variable period to capture a possible upswing in general interest in sport over the study period; in other words, starting a sport may occur increasingly frequently (or less frequently) over the course of time. The variable period extends from 1948 up to and including 2010. Various empirical studies have confirmed that cultural resources can be influential in sport participation. The higher a person’s education, the more likely they are to participate in sport (Farrell & Shields, 2002; Lunn, 2010; Scheerder et al., 2002; Stamm & Lamprecht, 2005). We measured educational level (highest level completed or currently in progress; as a proxy for cultural resources) in three categories: low (1), average (2) and high (3).2 We coded migration background in three categories as follows: native Dutch origin (0), non-native of non-Western origin (1) and non-native of Western origin (2). Non-native respondents of nonWestern origin were from Africa, Latin America and Asia (except Indonesia and Japan) or Turkey. We likewise accounted for the influence of the number of sports a person participated in during a particular person-year, because this could affect the odds of starting an additional new sport. We measured this by subtracting from the total number of sports practised by a respondent in a given year, the number of sports that he or she had started in that year. To determine if sport participation in youth affected the odds of starting a sport, we assigned a value of “1” to respondents who between their eight and twelfth year of age had participated in one or more sports for a period of one year or more. We assigned non-sport participants at that age a value of “0”. Table 2.1 presents an
RkJQdWJsaXNoZXIy MjY0ODMw